ournal An. Ceran. Soc, 82 [1]111-16(1999) Radial variations in modulus and hardness in SCS-6 Silicon Carbide Fibers Adrian B Mann, t, Mehdi Balooch, s John H. Kinney, and Timothy P. Weihs"t Department of Materials Science and Engineering, The Johns Hopkins University, Baltimore, Maryland Department of CI ry and Materials Science, Lawrence Livermore National Labo Livermore, California The mechanical properties of SCS-6 SiC fibers were mea- To date. most mechanical characterizations of scs fibers sured as a function of fiber radius using nanoindentation have focused on the averaged properties of the fibers measured techniques. Hardness and Youngs modulus were charac- over their full 142 um diameters. The manufacturer reported a terized for the material in all of the major regions of these tensile strength of 3. 95 GPa, Youngs modulus of 400 GPa, and fibers: the carbon core the graphitic core coating, the inner a density of 3.045 kgm for SCS-6 fibers. DiCarlo and wiI- SiC sheath, and the outer SiC sheath. The carbon core of liams measured a Youngs modulus of 390 GPa, and Zywicz the fibers was determined to be uniform in properties but et al. predicted a value of 378 GPa using a micromechanical extremely compliant. Youngs modulus of 28 GPa and a hardness of 4.2 GPa were measured. The graphitic core Strengths of as-deposited and thermally processed SCS-6 fibers oating was found to exhibit considerable anelasticity and have also been measured by bhatt and Hull in an axial orien to have both a low modulus (21 GPa) and a low hardness tation and by Eldridge et al. in a transverse direction. While (1.7 GPa). The inner sheath of the fiber, which contained a these average properties can be used to predict composite be- ed a sharp increase in stiffness and havior, they are limited in accuracy because they do not ac- hardness from the inner core. Modulus and hardness in- count for radial variations in behavior within the fibers Con creased by an order of magnitude over just 1 or 2 um when sider, for example, the difference in the behavior of the inner transversing radially away from the core into the SiC. This carbon core and the outer SiC sheaths. Di Carlo reported that hange in properties was pronounced and clearly defined. the inner carbon core of SCS fibers has a Young's modulus of The outer sheath, which contained a uniform chemistr only 41 GPa* while the SiC in the outer sections of the fibers and microstructure, was consistently stiff and hard when is approximately 10 times stiffer. The inner core bears little of transversing radially. The average modulus and hardness the applied load, yet it fractures in tension before the Sic for the full fiber was 333 GPa. The values reported for sheaths based on X-ray tomographic microscopy(XTM)re- Young s modulus and hardness clearly showed that the sults. The core fails at one half of the composite fiber's mechanical properties of SCS SiC fibers exhibit dramatic ultimate strength and its failure may affect the initiation of hanges across their diameters. cracking within the SiC. The differences in properties across individual fibers should be included in numerical models for . Introduction accurate predictions of the composite fiber performance Some researchers have begun to investigate the manner in ILICON CARBIDE fibers are commonly used to reinforce met which mechanical properties vary across SCS fibers, Sathish et als and ceramics in high-performance composite materials al. 14, 15 attempted to characterize elastic properties indirectly One class of commercial SiC fibers that has found broad com- using scanning acoustic microscopy. They measured the acous- Textron's family of SCS fibers. These fibers have a composite inner and outer Sic sheaths. They also predicted lower and structure because they are fabricated by depositing Sic onto upper bounds for this signature using the Hashin-Shtrikm carbon filaments. Many groups have investigated the micro- theory and Auger spectroscopy data from Ning and Pirouz& structure and chemistry of these fibers and they have found grain size, texture, and Si C ratios vary significantly with radial the inner sheath, jumped -20%at the border with the outer distance from the center of the fibers 2-13 These radial varia- sheath. and then remained constant across the outer sheath. tions are expected to produce similar variations in the me- hanical properties of the SCS fibers which must be quantified ond-order elastic constants, the measured signature suggest to accurately predict the performance of the fibers in bulk the elastic properties of SCS fibers do vary across the Sic sheaths. However. the measured signature falls well outside the lower and upper bounds that Sathish et al. predicted. Subse- uent results by the same group show a significant variation in modulus with radial distance and an average Youngs modu- D. B. Marshalk--contributing editor lus of 335-365 GPa for the whole fiber. Young's modulus increases from 40(+4)GPa in the core to 413(*40) GPa in the outer SiC This paper extends these initial studies using nanoindenta ced In tion experiments that characterize both Youngs modulus (E and the hardness(H) of SCS-6 fibers as a function of fiber radius. This technique has the advantage of characterizing me- chanical properties on a very small length scale ting local variations in properties to be examine author. Email: abmann a jhunix. hcf. jhuedu of submicrometer deep indentations were made on cross- ermore National Laboratory sectional faces of SCS-6 fibers. The resulting force-dis-
Radial Variations in Modulus and Hardness in SCS-6 Silicon Carbide Fibers Adrian B. Mann,†,‡ Mehdi Balooch,§ John H. Kinney,§ and Timothy P. Weihs*,† Department of Materials Science and Engineering, The Johns Hopkins University, Baltimore, Maryland 21218 Department of Chemistry and Materials Science, Lawrence Livermore National Laboratory, Livermore, California 94551 The mechanical properties of SCS-6 SiC fibers were measured as a function of fiber radius using nanoindentation techniques. Hardness and Young’s modulus were characterized for the material in all of the major regions of these fibers: the carbon core, the graphitic core coating, the inner SiC sheath, and the outer SiC sheath. The carbon core of the fibers was determined to be uniform in properties but extremely compliant. Young’s modulus of 28 GPa and a hardness of 4.2 GPa were measured. The graphitic core coating was found to exhibit considerable anelasticity and to have both a low modulus (21 GPa) and a low hardness (1.7 GPa). The inner sheath of the fiber, which contained a varying chemistry, showed a sharp increase in stiffness and hardness from the inner core. Modulus and hardness increased by an order of magnitude over just 1 or 2 µm when transversing radially away from the core into the SiC. This change in properties was pronounced and clearly defined. The outer sheath, which contained a uniform chemistry and microstructure, was consistently stiff and hard when transversing radially. The average modulus and hardness for the full fiber was 333 GPa. The values reported for Young’s modulus and hardness clearly showed that the mechanical properties of SCS SiC fibers exhibit dramatic changes across their diameters. I. Introduction SILICON CARBIDE fibers are commonly used to reinforce metals and ceramics in high-performance composite materials. One class of commercial SiC fibers that has found broad commercial applications and has spawned many research projects is Textron’s family of SCS fibers.1 These fibers have a composite structure because they are fabricated by depositing SiC onto carbon filaments.1 Many groups have investigated the microstructure and chemistry of these fibers and they have found that grain size, texture, and Si:C ratios vary significantly with radial distance from the center of the fibers.2–13 These radial variations are expected to produce similar variations in the mechanical properties of the SCS fibers which must be quantified to accurately predict the performance of the fibers in bulk composites. To date, most mechanical characterizations of SCS fibers have focused on the averaged properties of the fibers measured over their full 142 mm diameters. The manufacturer reported a tensile strength of 3.95 GPa, Young’s modulus of 400 GPa, and a density of 3.045 kg/m3 for SCS-6 fibers.1 DiCarlo and Williams3 measured a Young’s modulus of 390 GPa, and Zywicz et al.11 predicted a value of 378 GPa using a micromechanical numerical model and elastic constants for graphite and SiC. Strengths of as-deposited and thermally processed SCS-6 fibers have also been measured by Bhatt and Hull9 in an axial orientation and by Eldridge et al.13 in a transverse direction. While these average properties can be used to predict composite behavior, they are limited in accuracy because they do not account for radial variations in behavior within the fibers. Consider, for example, the difference in the behavior of the inner carbon core and the outer SiC sheaths. DiCarlo reported that the inner carbon core of SCS fibers has a Young’s modulus of only 41 GPa4 while the SiC in the outer sections of the fibers is approximately 10 times stiffer. The inner core bears little of the applied load, yet it fractures in tension before the SiC sheaths based on X-ray tomographic microscopy (XTM) results.11 The core fails at one half of the composite fiber’s ultimate strength and its failure may affect the initiation of cracking within the SiC.11 The differences in properties across individual fibers should be included in numerical models for accurate predictions of the composite fiber performance. Some researchers have begun to investigate the manner in which mechanical properties vary across SCS fibers. Sathish et al.14,15 attempted to characterize elastic properties indirectly using scanning acoustic microscopy. They measured the acoustic signature of SCS fibers while moving radially across the inner and outer SiC sheaths. They also predicted lower and upper bounds for this signature using the Hashin–Shtrikman theory14 and Auger spectroscopy data from Ning and Pirouz.8 Their measured signature decreased gradually (∼15%) across the inner sheath, jumped ∼20% at the border with the outer sheath, and then remained constant across the outer sheath. Since acoustic signatures depend strongly on a material’s second-order elastic constants, the measured signature suggests the elastic properties of SCS fibers do vary across the SiC sheaths. However, the measured signature falls well outside the lower and upper bounds that Sathish et al. predicted. Subsequent results by the same group15 show a significant variation in modulus with radial distance and an average Young’s modulus of 335–365 GPa for the whole fiber. Young’s modulus increases from 40 (±4) GPa in the core to 413 (±40) GPa in the outer SiC. This paper extends these initial studies using nanoindentation experiments that characterize both Young’s modulus (E) and the hardness (H) of SCS-6 fibers as a function of fiber radius. This technique has the advantage of characterizing mechanical properties on a very small length scale, hence permitting local variations in properties to be examined. Large arrays of submicrometer deep indentations were made on crosssectional faces of SCS-6 fibers. The resulting force–disD. B. Marshall—contributing editor Manuscript No. 191574. Received August 28, 1996; approved April 24, 1998. Partially funded by the U.S. Department of Energy’s Advanced Industrial Materials Program under U.S. DOE Contract No. 2-7405-eng-48. Supported at The Johns Hopkins University by the NSF MRSEC on Nanostructured Materials (Award No. 9632526) and the ARL/Advanced Materials Characterization Program (Award No. 019620047). *Member, American Ceramic Society. † The Johns Hopkins University. ‡ Corresponding author. Email: abmann@jhunix.hcf.jhu.edu. § Lawrence Livermore National Laboratory. J. Am. Ceram. Soc., 82 [1] 111–16 (1999) Journal 111
112 Journal of the American Ceramic Society-Mann et al. Vol. 82. No placement data were analyzed to produce a spatial mapping of texture, and chemistry, and is actually divided into three sepa- modulus and hardness from the center of the carbon core to the rate regions by Ning and Pirouz. 8 The B-Sic crystals are dge of the fiber. The results are compared with measured equiaxed near the graphitic carbon coating, but change rapidly average fiber properties and with the elastic constants for SiC to a textured columnar structure within the first 0.2 um of the and C inner sheath. The radial growth direction favors the(1l1)ori- entation for the SiC crystals. The width of the grains increases IL. Fabrication, Microstructure, and Chemistry of almost 10 times from 12 nm near the carbon core to 100 nm at SCS-6 Fibers the outer edge of the inner sheath. o Ning and Pirouz report similar increases with radiu They also note that the length to SCS-6 fibers are fabricated through a series of chemical width ratios of the columnar SiC grains remain constant near apor depositions(CVD). The process starts by depositing 1.5 10 throughout this section. The microstructure of the carbon um of graphitic carbon on monofilaments of carbon that mea- carbon forms randomly oriented nanometer-sized grains next to nents is designed to smooth the carbon core and to increase its the graphitic carbon coating and then 3 nm thick graphite layers electrical conductivity since it is heated resistively during de around the B-SiC grains. The chemistry within the inner position. In the next step, a 50 um thick SiC layer is deposited sheath varies strongly as well. Sattler et al. o used parallel This layer is the main structural element in the fiber. The layer electron energy loss steo of carbon to silicon atoms decreases roscopy(PEELS)and X-ray micros- has a strong variation in chemistry and structure in the first 15 copy to show that the rat um and a fairly uniform chemistry and microstructure in the smoothly from 2.0 at the edge of the graphitic carbon layer to final 35 um. The first 15 um is commonly called the inn approximately 1. 2 at the boundary between the inner and outer sheath, and the next 35 um the outer sheath. In the final de sheath(Fig. 2). Ning and Pirouz reported a smaller, stepped and it is designed to heal the fiber surface and thereby improve fibers is marked by an abrupt increase in the SiC columnar fracture strength grain size from 100 nm to approximately 250 nm. The distinct The circular cross section of ScS-6 fibers can be divided border is generated by variations in deposition conditions, and into a number of sections. as shown in Fig. 1. The innermost it is clearly visible under an optical microscope. The SiC crys- section is the carbon core. It measures 33 um in diameter and tals in the outer sheath are highly faulted, and they remain onsists of turbostratic carbon blocks that range in size from I strongly textured with(I 11)orientations parallel to the radial to 50 nm. 6-8 The carbon blocks or grains are arranged in a direction. Ning and Pirouz found that the Sic grain size gradu- random orientation so the strength and stiffness of the core are ally decreased across the outer sheath back to 100 nm at its expected to be isotropic. The inner core is surrounded by a 1.5 exterior. A similar value was reported by Sattler et al.10The um thick carbon coating that contains grains of graphite rang- C/Si atomic ratio within the outer sheath starts at 1.2 and ther ing in size from 25 to 50 nm.6-10 The graphitic grains are quickly levels off to the stoichiometric ratio of 1.0 for SiC(Fig extured with(0002) planes aligned normal to the radial direc 2). The final layer of the SCS-6 fibers is a 3.0 um coating that tion.Sattler et al. 10 note that cracks lie within the coating. consists of a carbon matrix with SiC particles parallel to the(0002) plane Moving radially away from the core, the next main section of the SCS-6 fibers is termed the inner sheath. It is approxi- Ill. Experimental Procedure mately 15 um thick and consists of B-SiC crystals imbedded in Large arrays of indentations were made on mechanically a carbon matrix. This section has large variations in grain size polished cross sections of ScS-6 fibers using a nanoindenter XPle and a three-sided Berkovich diamond tip. The polishing was successively finer SiC abrasive paper through 1000 grit, and then in the final stages successively smaller colloidal diamond suspensions(through 0. 1 um)on a Textron SCS-6 SiC Fibers Prulecuive couting Chemistry profile of scs-6 Fiber Outer Si Peak (Ning and Firuz s) 29月 Carbon curlin Fiber Radius (um) Fig. 1. Schematic, cross-sectional view of an SCS-6 SiC fiber from Fig. 2. Reports of chemical variations across SCS-6 SiC fibers
placement data were analyzed to produce a spatial mapping of modulus and hardness from the center of the carbon core to the edge of the fiber. The results are compared with measured average fiber properties and with the elastic constants for SiC and C. II. Fabrication, Microstructure, and Chemistry of SCS-6 Fibers SCS-6 fibers are fabricated through a series of chemical vapor depositions (CVD). The process starts by depositing 1.5 mm of graphitic carbon on monofilaments of carbon that measure 33 mm in diameter. This first coating of the carbon filaments is designed to smooth the carbon core and to increase its electrical conductivity since it is heated resistively during deposition. In the next step, a 50 mm thick SiC layer is deposited. This layer is the main structural element in the fiber. The layer has a strong variation in chemistry and structure in the first 15 mm and a fairly uniform chemistry and microstructure in the final 35 mm. The first 15 mm is commonly called the inner sheath, and the next 35 mm the outer sheath. In the final deposition step, a 3.0 mm thick protective layer is added to the fiber. This layer consists of SiC particles in a carbon matrix, and it is designed to heal the fiber surface and thereby improve fracture strengths. The circular cross section of SCS-6 fibers can be divided into a number of sections, as shown in Fig. 1. The innermost section is the carbon core. It measures 33 mm in diameter and consists of turbostratic carbon blocks that range in size from 1 to 50 nm.6–8 The carbon blocks or grains are arranged in a random orientation so the strength and stiffness of the core are expected to be isotropic. The inner core is surrounded by a 1.5 mm thick carbon coating that contains grains of graphite ranging in size from 25 to 50 nm.6–10 The graphitic grains are textured with (0002) planes aligned normal to the radial direction. Sattler et al.10 note that cracks lie within the coating, parallel to the (0002) planes. Moving radially away from the core, the next main section of the SCS-6 fibers is termed the inner sheath. It is approximately 15 mm thick and consists of b-SiC crystals imbedded in a carbon matrix. This section has large variations in grain size, texture, and chemistry, and is actually divided into three separate regions by Ning and Pirouz.8 The b-SiC crystals are equiaxed near the graphitic carbon coating, but change rapidly to a textured columnar structure within the first 0.2 mm of the inner sheath. The radial growth direction favors the 〈111〉 orientation for the SiC crystals. The width of the grains increases almost 10 times from 12 nm near the carbon core to 100 nm at the outer edge of the inner sheath.10 Ning and Pirouz report similar increases with radius. They also note that the length to width ratios of the columnar SiC grains remain constant near 10 throughout this section. The microstructure of the carbon matrix also changes dramatically within the inner sheath. The carbon forms randomly oriented nanometer-sized grains next to the graphitic carbon coating and then 3 nm thick graphite layers around the b-SiC grains.7 The chemistry within the inner sheath varies strongly as well. Sattler et al.10 used parallel electron energy loss spectroscopy (PEELS) and X-ray microscopy to show that the ratio of carbon to silicon atoms decreases smoothly from 2.0 at the edge of the graphitic carbon layer to approximately 1.2 at the boundary between the inner and outer sheath (Fig. 2). Ning and Pirouz8 reported a smaller, stepped variation in the carbon to silicon ratio over the same area. The transition from the inner to the outer sheath of the SCS-6 fibers is marked by an abrupt increase in the SiC columnar grain size from 100 nm to approximately 250 nm. The distinct border is generated by variations in deposition conditions, and it is clearly visible under an optical microscope. The SiC crystals in the outer sheath are highly faulted, and they remain strongly textured with 〈111〉 orientations parallel to the radial direction. Ning and Pirouz8 found that the SiC grain size gradually decreased across the outer sheath back to 100 nm at its exterior. A similar value was reported by Sattler et al.10 The C/Si atomic ratio within the outer sheath starts at 1.2 and then quickly levels off to the stoichiometric ratio of 1.0 for SiC (Fig. 2). The final layer of the SCS-6 fibers is a 3.0 mm coating that consists of a carbon matrix with SiC particles. III. Experimental Procedure Large arrays of indentations were made on mechanically polished cross sections of SCS-6 fibers using a Nanoindenter XP16 and a three-sided Berkovich diamond tip. The polishing was performed using successively finer SiC abrasive paper through 1000 grit, and then in the final stages successively smaller colloidal diamond suspensions (through 0.1 mm) on a Fig. 1. Schematic, cross-sectional view of an SCS-6 SiC fiber from Textron. Fig. 2. Reports of chemical variations across SCS-6 SiC fibers. 112 Journal of the American Ceramic Society—Mann et al. Vol. 82, No. 1
January 1999 Radial variations in Modulus and Hardness in SCS-6 Silicon Carbide Fibers at brass wheel. This final stage was found to be the best way curves for the inner and outer sheaths ha steeper slopes to ensure an even surface across both the hard Sic outer sheath at the maximum load than that of the core and the soft carbon core. The indentations within each array The material in the inner and outer sheaths deforms elasti- were typically 10 um apart. However, when small arrays were cally and inelastically during indentation as shown by the used to examine the variation in properties across the edge of force-displacement curves in Fig. 3. The first unloading seg- the core, the indentation spacing was reduced to 2 or 3 um. ment is offset from the initial loading segment, and there is a Various indentation procedures were used during the testing; in anent indentation depth when all force is removed. The each case the indentations were performed by increasing the deformation on subsequent reloading and unloading, though, is d at a fixed rate up to a peak load in the range 3 to 40 m recoverable and therefore elastic as seen by the overlapping The indentations were unloaded at a fixed rate, and repeated curves. The indentations which are in the SiC but within a few loading/unloading cycles were utilized. Additionally, a hold micrometers of the carbon core exhibit greater permanent de egment was inserted into the final unload at 20% of the pea formation than those farther from the core. This suggests that load to enable a correction to be made for thermal drift. The the weakest part of the SiC sheath lies in very close proximity peated cycling of force was used to investigate the presence to the core of anelastic deformation The indentation curves for each fiber were analyzed To determine hardness and Youngs modulus from the force the method of oliver and Pharr 7 assuming a poisson rat tio of lacement data, the procedure outlined by oliver and Pharr 0.25, and then averaged so that variations in modulus and was followed. 7 Power law curves were fitted to the first 90% hardness could be plotted as a function of the radial distance of the unloading curves, and contact stiffnesses and contact from the fiber's center. The averaged results(based on over depths were obtained by differentiating and extrapolating these 300 individual data points) are shown in Fig. 4. The inner curves. Care was taken to remove the machine compliance that carbon core shows constant hardness (4.2 GPa)and modulus is associated with the tip shaft and the sample mounting. This (28 GPa) across its full width. The Sic also has approximatel was calibrated by performing a large array of indentations on a constant hardness(34 GPa) and modulus(360 GPa), though standard(fused silica) sample. The machine compliance was close to the core both the hardness and the modulus of the inner ermined to be 1.62 x 10-7 m/N. The data from the calibra SiC sheath fall dramatically. This drop in modulus and hard tion run were also utilized to evaluate the shape of the diamond ness begins gradually, as shown by Fig. 4, but close to the inner ip using the method outlined by Oliver and Pharr. The con- core this drop becomes very rapid In this region the averaged tact depths and the calibrated diamond tip shape were used to data are somewhat misleading as they show a gradual drop in calculate the contact area at peak load for each indentation. modulus and hardness. Individual indentations do not show Hardness was calculated by dividing the maximum applied such a gradual transition. Instead the modulus and hardness for ad by the contact area under load, and Youngs modulus was alculated using the relationship between contact area and con tact stiffness for a parabolic tip Pyrolytic SiC in IV. Results ore Outer siC Matri叉 e-displacement data from three indentations into ar 4联 CS-6 SiC fiber are shown in Fig. 3. The indentations were made in three separate positions across the fiber's face: the carbon core, the edge of the inner sheath(close to the core), and the outer sheath. The loading curves and the unloading curves are all parabolic in shape. The indentation into the carbon core is predominantly elastic in nature. The loading and unloading curves show little separation, and when force is cycled, the path of the first loading curve is repeated. This suggests there is little inelastic or nonrecoverable deformation in the carbon that under the diamond indenter. The shallow slope for this in- mentation shows clearly that the carbo is very compliant compared to the inner and outer sheaths. The indentation IlI it Fraction of Radius Fig 4. Avera Fig 3. Load-displacement data for indentations into three different fiber's radius. Ea regions of an SCS-6 SiC fib
flat brass wheel. This final stage was found to be the best way to ensure an even surface across both the hard SiC outer sheath and the soft carbon core. The indentations within each array were typically 10 mm apart. However, when small arrays were used to examine the variation in properties across the edge of the core, the indentation spacing was reduced to 2 or 3 mm. Various indentation procedures were used during the testing; in each case the indentations were performed by increasing the load at a fixed rate up to a peak load in the range 3 to 40 mN. The indentations were unloaded at a fixed rate, and repeated loading/unloading cycles were utilized. Additionally, a hold segment was inserted into the final unload at 20% of the peak load to enable a correction to be made for thermal drift. The repeated cycling of force was used to investigate the presence of anelastic deformation. To determine hardness and Young’s modulus from the force displacement data, the procedure outlined by Oliver and Pharr was followed.17 Power law curves were fitted to the first 90% of the unloading curves, and contact stiffnesses and contact depths were obtained by differentiating and extrapolating these curves. Care was taken to remove the machine compliance that is associated with the tip shaft and the sample mounting. This was calibrated by performing a large array of indentations on a standard (fused silica) sample. The machine compliance was determined to be 1.62 × 10−7 m/N. The data from the calibration run were also utilized to evaluate the shape of the diamond tip using the method outlined by Oliver and Pharr.17 The contact depths and the calibrated diamond tip shape were used to calculate the contact area at peak load for each indentation. Hardness was calculated by dividing the maximum applied load by the contact area under load, and Young’s modulus was calculated using the relationship between contact area and contact stiffness for a parabolic tip.17 IV. Results Force–displacement data from three indentations into an SCS-6 SiC fiber are shown in Fig. 3. The indentations were made in three separate positions across the fiber’s face: the carbon core, the edge of the inner sheath (close to the core), and the outer sheath. The loading curves and the unloading curves are all parabolic in shape. The indentation into the carbon core is predominantly elastic in nature. The loading and unloading curves show little separation, and when force is cycled, the path of the first loading curve is repeated. This suggests there is little inelastic or nonrecoverable deformation in the carbon that is under the diamond indenter. The shallow slope for this indentation shows clearly that the carbon core is very compliant compared to the inner and outer sheaths. The indentation curves for the inner and outer sheaths have much steeper slopes at the maximum load than that of the core. The material in the inner and outer sheaths deforms elastically and inelastically during indentation as shown by the force–displacement curves in Fig. 3. The first unloading segment is offset from the initial loading segment, and there is a permanent indentation depth when all force is removed. The deformation on subsequent reloading and unloading, though, is recoverable and therefore elastic as seen by the overlapping curves. The indentations which are in the SiC but within a few micrometers of the carbon core exhibit greater permanent deformation than those farther from the core. This suggests that the weakest part of the SiC sheath lies in very close proximity to the core. The indentation curves for each fiber were analyzed using the method of Oliver and Pharr17 assuming a Poisson ratio of 0.25, and then averaged so that variations in modulus and hardness could be plotted as a function of the radial distance from the fiber’s center. The averaged results (based on over 300 individual data points) are shown in Fig. 4. The inner carbon core shows constant hardness (4.2 GPa) and modulus (28 GPa) across its full width. The SiC also has approximately constant hardness (34 GPa) and modulus (360 GPa), though close to the core both the hardness and the modulus of the inner SiC sheath fall dramatically. This drop in modulus and hardness begins gradually, as shown by Fig. 4, but close to the inner core this drop becomes very rapid. In this region the averaged data are somewhat misleading as they show a gradual drop in modulus and hardness. Individual indentations do not show such a gradual transition. Instead the modulus and hardness for Fig. 3. Load–displacement data for indentations into three different regions of an SCS-6 SiC fiber. Fig. 4. Average Young’s modulus and hardness as a function of the fiber’s radius. Each average data point includes data for indentations performed to a number of different peak loads. January 1999 Radial Variations in Modulus and Hardness in SCS-6 Silicon Carbide Fibers 113
114 Vol. 82. Ne the individual indentations in this region appear to fall into two distinct categories: first, those with a modulus of=175 GPa and hardness of =22 GPa, and, second, those with a modulus of =300 GPa and a hardness of =33 GPa The observation of two distinct values of mechanical prop. erties near the core suggests that the variations in this regio are very localized. This is probably a result of the complex 300 tructure of the sic near the core. as described earlier. the crystals are equiaxed near the graphitic carbon coating hange rapidly to a textured columnar structure within the 2 um of the inner sheath, the favored radial orientation ng(111. The carbon in the inner SiC sheath forms ran- domly oriented nanometer-sized grains next to the graphitic Carbon core arbon coating and then 3 nm thick graphite layers around the - SiC grains. Additionally, the proximity to the graphitic car- 0 bon which surrounds the core means that the volume being sampled during the indentation may include both the SiC inner sheath and the graphitic carbon coating The graphitic carbon has a behavior different from that of either the Sic or the carbon core. while the average data of Fig. 4 do not show this variation, it is clearly demonstrated by he individual indentations in Fig. 5. The three load-dis lacement curves, A, B, and C, form part of a set of 20 closely paced indentations (2 um apart) performed across the edge of the carbon core Indent a is in the inner edge of the SiC sheath indent B is in the graphitic coating, and indent C is in the arbon core. The graphitic coating has a lower hardness(1.7 20 lus(21 GPa) than the carbon core. This is (C) shown in Fig. 6, which gives the calculated properties for a set of indents(including A, B, and C)across the edge of the core Note that the distinct drop in H and e at the coating(indent B) was found to be very reprod sic The graphitic carbon shows considerable anelastic deforma- tion. This is demonstrated by Fig. 7, where repeated load TTTTT unloading cycles show large hysteresis loops. The unusual 10 deformation characteristics of the graphitic carbon, including Indent no the low modulus, the low hardness, and the anelasticity, can be attributed to the orientation of the graphite(0002) planes nor Fig. 6. Calculated modulus and hardness for a set of indentations mal to the radial direction. This orientation enables graphite (including indentations A, B, and C of Fig. 5)which cross the et, he lanes to slip over each other along the axis of the fiber with relative ease, thus giving a low hardness. The anelastic behav fiber's core ior is also likely to be a result of the graphite planes slipping over each other. The modulus observed for the graphitic layer close to the Reuss(isostress)average of 25.5 GPa for poly crystalline graphite. This average is based on the elastic con- Graphitic Carbon stants given by blakeslee et al Lastly, to be sure that the measured properties are not af- fected by the polish ocess. the modulus and hardness of the core and the sic sheath were examined as a function of indentation depth(see Figs. 8 and 9). Very little variation with displacement(n Carbon(B) Fage 7. Load-displacement data for an indentation cycles show hysteresis loops which are due peated which coats the carbon core. The repeated =品 depth was seen in the modulus and hardness of the carbon core 30035040 or in the hardness of the sic. but the modulus of the sic showed a slight drop(from 360 to 305 GPa) as the indentation depth increased. This could be due to the surface preparation, but may also be a result of the method used to calculate the contact depth from the indentation data. Specifically, for hard SiC sheath, indentation B is in the graphitic carbon coating of the core, materials the assumption that the tip behaves as a parabola may and indentation C is in the core not be valid for all indentation depths
the individual indentations in this region appear to fall into two distinct categories: first, those with a modulus of ≈175 GPa and a hardness of ≈22 GPa, and, second, those with a modulus of ≈300 GPa and a hardness of ≈33 GPa. The observation of two distinct values of mechanical properties near the core suggests that the variations in this region are very localized. This is probably a result of the complex structure of the SiC near the core. As described earlier, the b-SiC crystals are equiaxed near the graphitic carbon coating, but change rapidly to a textured columnar structure within the first 0.2 mm of the inner sheath, the favored radial orientation being 〈111〉. The carbon in the inner SiC sheath forms randomly oriented nanometer-sized grains next to the graphitic carbon coating and then 3 nm thick graphite layers around the b-SiC grains.7 Additionally, the proximity to the graphitic carbon which surrounds the core means that the volume being sampled during the indentation may include both the SiC inner sheath and the graphitic carbon coating. The graphitic carbon has a behavior quite different from that of either the SiC or the carbon core. While the average data of Fig. 4 do not show this variation, it is clearly demonstrated by the individual indentations in Fig. 5. The three load–displacement curves, A, B, and C, form part of a set of 20 closely spaced indentations (2 mm apart) performed across the edge of the carbon core. Indent A is in the inner edge of the SiC sheath, indent B is in the graphitic coating, and indent C is in the carbon core. The graphitic coating has a lower hardness (1.7 GPa) and modulus (21 GPa) than the carbon core. This is shown in Fig. 6, which gives the calculated properties for a set of indents (including A, B, and C) across the edge of the core. Note that the distinct drop in H and E at the coating (indent B) was found to be very reproducible. The graphitic carbon shows considerable anelastic deformation. This is demonstrated by Fig. 7, where repeated loading/unloading cycles show large hysteresis loops. The unusual deformation characteristics of the graphitic carbon, including the low modulus, the low hardness, and the anelasticity, can be attributed to the orientation of the graphite (0002) planes normal to the radial direction. This orientation enables graphite planes to slip over each other along the axis of the fiber with relative ease, thus giving a low hardness. The anelastic behavior is also likely to be a result of the graphite planes slipping over each other. The modulus observed for the graphitic layer is close to the Reuss (isostress) average of 25.5 GPa for polycrystalline graphite. This average is based on the elastic constants given by Blakslee et al.18 Lastly, to be sure that the measured properties are not affected by the polishing process, the modulus and hardness of the core and the SiC sheath were examined as a function of indentation depth (see Figs. 8 and 9). Very little variation with depth was seen in the modulus and hardness of the carbon core or in the hardness of the SiC, but the modulus of the SiC showed a slight drop (from 360 to 305 GPa) as the indentation depth increased. This could be due to the surface preparation, but may also be a result of the method used to calculate the contact depth from the indentation data. Specifically, for hard materials the assumption that the tip behaves as a parabola may not be valid for all indentation depths. Fig. 5. Load–displacement data for indentations across the edge of the carbon core and the SiC sheath. Indentation A is just in the inner SiC sheath, indentation B is in the graphitic carbon coating of the core, and indentation C is in the core. Fig. 6. Calculated modulus and hardness for a set of indentations (including indentations A, B, and C of Fig. 5) which cross the edge of the core. The inset shows the location of the indentations relative to the fiber’s core. Fig. 7. Load–displacement data for an indentation into the graphitic layer which coats the carbon core. The repeated loading/unloading cycles show hysteresis loops which are due to anelastic deformation. 114 Journal of the American Ceramic Society—Mann et al. Vol. 82, No. 1
January 1999 Radial variations in Modulus and Hardness in SCS-6 Silicon Carbide Fibers Carbon Core SiC Outer sheath 480 6g0 on Core SiC Outer Sheath H 600 30Q Fig. 8. Modulus and hardness as a function of indentation depth in Fig. 9. Modulus and hardness as a function of indentation depth in the outer siC sheath. V. Discussion the graphitic carbon layer which lies at the edge of the carbon The layer consists of small of textured turbostratic The modulus and hardness data in Figs 4 and 6 clearly show arbon. Note that the magnitude of the measured modulus and that the mechanical properties of scs-6 Sic fibers vary dra hardness are less than the values recorded for the carbon core matically with distance from their centers. Youngs modulus The texturing in the coating is believed to produce these lower increases from 28 GPa at the inner core to approximately 360 values of E and H compared to the carbon filament. It is most GPa in the outer sheath, and hardness increases from roughly 4 likely due to the(0002) planes buckling and sliding over each to 34 GPa over the same range. These 13x and 9x increases are other very significant and suggest that the fibers will behave as com It is interesting to note that due to the extreme anisotropy of ite materials. This result matches DiCarlo's predictions graphite the Reuss (isostress)modulus for polycrystalline and it also helps to explain Zywicz et al's XTM report that aphite is many times smaller than the Voigt(isostrain )modu- SCS fibers fail first in their cores prior to their SiC sheaths lus(e.g, Ref. 19). Since the Reuss value is closest to the t is also worth noting that the graphitic layer between the core measured moduli of both the carbon core and the graphitic core and the Sic has a lower indentation modulus and hardness than coating, the indentation process must test the carbon under any other part of the fiber, and hence, it must be regarded as a conditions of isostress. In real applications of the fibers, the potential site of failure. The mechanical properties of each of loading is unlikely to resemble conditions of isostrain, and ons will now be considered in relation to their chem- hence, indentation testing may not be the most effective way of istry and microstructure examining the modulus of the carbon core and the graphiti The inner cores of the fibers consist of carbon filaments that layer uniform grain sizes and are expected to be isotropic and com- s the next region within the SCS fibers is the inner Sic measure 16.5 um in radii. These carbon filaments have small pliant. Note that the data points that fall within the inner core caath. This region consists of B-SiC crystals imbedded in a ig. 4 give consistent values of 25-30 GPa for Youngs nificantly with radius&, 10 The section closest to the carbon core modulus and 4.0-4.4 GPa for hardness. These results imply is strongly carbon-rich while the section bordering the outer that the carbon core is very uniform in its elastic and plastic sheath has a nearly equal mix of C and Si atoms as shown in properties. The magnitude of the measured modulus is lowe Fig. 2. The data of Fig. 4 show that the modulus and hardness an, but similar to, at least three other reports in the literature. increase as the indentations move away from the core. Most of DiCarlo reported an approximate value of 41 GPa for E of the the increase is confined to the first few micrometers of the core, Eldridge et al. cited an estimated value of 35 GPa for E, inner SiC sheath where the chemistry and structure of the Sic while most recently Sathish et al. 15 reported an experiment change very sharply. As stated earlier, close to the carbon core ralue of 40 GPa. The difference between our value and those the indentation data fall into two types, the first with a high of other researchers probably results from the unique geometry modulus(=300 GPa)and hardness(=33 GPa) and the second of the indentation test. As stated previously, the Ye ith an intermediate modulus(=175 GPa) and hardness(=22 otained is very he Reuss, isostress, GPa). Sathish et al. have reported a region in the SiC, close of 25.5 GPa to the carbon core. where the modulus has an intermediate Before moving on to the inner sheath, it is usefi onsider value of 141-215 GPa. Further from the core. but still within
V. Discussion The modulus and hardness data in Figs. 4 and 6 clearly show that the mechanical properties of SCS-6 SiC fibers vary dramatically with distance from their centers. Young’s modulus increases from 28 GPa at the inner core to approximately 360 GPa in the outer sheath, and hardness increases from roughly 4 to 34 GPa over the same range. These 13× and 9× increases are very significant and suggest that the fibers will behave as composite materials. This result matches DiCarlo’s predictions,3,4 and it also helps to explain Zywicz et al.’s XTM report that SCS fibers fail first in their cores prior to their SiC sheaths.11 It is also worth noting that the graphitic layer between the core and the SiC has a lower indentation modulus and hardness than any other part of the fiber, and hence, it must be regarded as a potential site of failure. The mechanical properties of each of the regions will now be considered in relation to their chemistry and microstructure. The inner cores of the fibers consist of carbon filaments that measure 16.5 mm in radii. These carbon filaments have small, uniform grain sizes and are expected to be isotropic and compliant. Note that the data points that fall within the inner core in Fig. 4 give consistent values of 25–30 GPa for Young’s modulus and 4.0–4.4 GPa for hardness. These results imply that the carbon core is very uniform in its elastic and plastic properties. The magnitude of the measured modulus is lower than, but similar to, at least three other reports in the literature. DiCarlo reported an approximate value of 41 GPa for E of the core, Eldridge et al. cited an estimated value of 35 GPa for E, while most recently Sathish et al.15 reported an experimental value of 40 GPa. The difference between our value and those of other researchers probably results from the unique geometry of the indentation test. As stated previously, the Young’s modulus that we obtained is very close to the Reuss, isostress, average of 25.5 GPa. Before moving on to the inner sheath, it is useful to consider the graphitic carbon layer which lies at the edge of the carbon core. The layer consists of small grains of textured, turbostratic carbon. Note that the magnitude of the measured modulus and hardness are less than the values recorded for the carbon core. The texturing in the coating is believed to produce these lower values of E and H compared to the carbon filament. It is most likely due to the (0002) planes buckling and sliding over each other. It is interesting to note that due to the extreme anisotropy of graphite the Reuss (isostress) modulus for polycrystalline graphite is many times smaller than the Voigt (isostrain) modulus (e.g., Ref. 19). Since the Reuss value is closest to the measured moduli of both the carbon core and the graphitic core coating, the indentation process must test the carbon under conditions of isostress. In real applications of the fibers, the loading is unlikely to resemble conditions of isostrain, and hence, indentation testing may not be the most effective way of examining the modulus of the carbon core and the graphitic layer. The next region within the SCS fibers is the inner SiC sheath. This region consists of b-SiC crystals imbedded in a carbon matrix and its chemistry and microstructure vary significantly with radius.8,10 The section closest to the carbon core is strongly carbon-rich while the section bordering the outer sheath has a nearly equal mix of C and Si atoms as shown in Fig. 2. The data of Fig. 4 show that the modulus and hardness increase as the indentations move away from the core. Most of the increase is confined to the first few micrometers of the inner SiC sheath where the chemistry and structure of the SiC change very sharply. As stated earlier, close to the carbon core the indentation data fall into two types, the first with a high modulus (≈300 GPa) and hardness (≈33 GPa) and the second with an intermediate modulus (≈175 GPa) and hardness (≈22 GPa). Sathish et al.15 have reported a region in the SiC, close to the carbon core, where the modulus has an intermediate value of 141–215 GPa. Further from the core, but still within Fig. 8. Modulus and hardness as a function of indentation depth in the carbon core. Fig. 9. Modulus and hardness as a function of indentation depth in the outer SiC sheath. January 1999 Radial Variations in Modulus and Hardness in SCS-6 Silicon Carbide Fibers 115
116 Journal of the American Ceramic Society-Mann et al. Vol. 82. No he inner SiC sheath, Sathish et al. report modulus values Youngs modulus of 28 GPa and a hardness of 4.2 GPa. The m a range of 220-310 GPa to a range of modulus and hardness appeared to be uniform in this core,as 315-405 GPa. Thus, our values for Youngs modulus across expected. The graphitic coating on the core appeared to be the SiC sheath(shown by Fig 4)are in good agreement with weaker, though, with a modulus of 21 GPa and a hardness of recent results of sathish et al. Is 1.7 GPa. The inner sheath, which contained a varying chem- an earlier paper Sathish et al. 4 used Auger spectroscopy stry, showed a sharp increase in stiffness and hardness from from Ning and Pirouz(Fig. 2) to predict variations of he inner core. The modulus and hardness increased radially bulk and shear moduli across the inner and outer sheaths of within this region to 360 and 34 GPa, respectively. Most of this SCS SiC fibers. They predicted that both moduli increase rap increase was confined to the first few micrometers of the inner idly across the inner sheath as reported here. However, the SiC sheath, and thereafter there were no signs of local minima so predicted a large, local minima in the inner SiC sheath or decreases, just a gradual increase to the maximum values This is something which they did not see in their subsequent observed. The outer sheath, which contained a relatively uni- xamination of the fibers 15 The results of these indentation form chemistry, was consistently stiff and hard. The average periments suppo arp nse in modulus and hardness were measured to be 360 and 34 GP like Sathish et al.'s later results they show no sign of a local which are similar to previous measurements or estimates. The minima. To help verify this finding, large areas of a fiber were average modulus for the full fiber was 333 GPa. The measured canned using an atomic force microscope(AFM) and an ac values of modulus and hardness clearly showed that the me technique that yields a measure of material stiffness 20, 21 No chanical properties of ScS-6 fibers changed by an order abrupt changes or local minima in stiffness were observed in magnitude across their diameters. Such dramatic variations he region of the inner sheath. Thus, all of the results reported should be considered in modeling the behavior of these here suggest that the variation in mechanical rtes across materials the inner sheath is continuous everywhere, except close to the carbon core. with no identifiable local minima Acknowledgments: We gratefully acknowledge Textron Speciality The last region that was investigated within the SCS fibers Materials for supplying the fibers, Professor W. D. Nix, Jerry English and Bob he outer sheath. This section consists of strongly textured SiC Kershaw for help during the initial testing grains and is relatively uniform in chemistry. The ratio of C to egion, as shown in Fig References 2. Young's modulus shows some scatter but is basically con- TExtron Specialty Materials, Textron CO., Lowell, MA. tant at an average value of 360 GPa. The distinct plateau S R Nutt and F E wawner. Silicon Carbide Filaments: Microstructure dulus in Fig. 4 is attributed to the uniform chemistry and J.Mer.Sc.,20,1953-60(1985) microstructure of the region. Hardness also shows a plateau J. A. DiCarlo and w. williams, ""Dynamic Modulus and Damping of Boron, Silicon Carbide, and Alumina Fibers, " Ceram. Eng. Sct. Proc., 1, 671-92 with an average value of 34 GPa. The magnitude of youngs modulus, 360 GPa, is a little lower than expected for SiC. For of CVD Silicon Carbide Fi- single crystals of B-SiC the modulus varies from 379 GPa Ceramics. Edited by R.A. the [100] crystallographic direction to 547 GPa in the [1ll Bradley. ASM International direction. 2 A Voigt(isostrain)average for the modulus, based a SCS-6 Fiber/Ti-6Al on single-crystal elastic constants, is 459 GPa, while the Reuss (isostress)average is 414 GPa. Thus, all of the expected values bJ. Chung: M.Sc. Thesis. Case Western Reserve University, Cleveland, OH, X J. Ning, P. Pirouz, K. P. D. Lagerlof, and J. DiCarlo, " The Structure of may be due to nonideal stoichiometry throughout the outer Car-76(1990). aly Vapor Deposited SiC Monofilaments, "JMater Res, 5, Young,s modulus averaged over the full cross-sectional area X J Ning and P. Pirouz, "The Microstructure of ScS-6 SiC Fibers, "J. of the fiber was calculated to be 333 GPa. which is close to the Mater.Res,6,2234-48(1991) T. Bhatt and D. R Hull, ""Microstructural and Strength Stability of CVD average moduli measured by Sathish et al 5 of 335-365 GP SiC Fibers in Argon Environment, Ceram. Eng. Sci Proc 12, 1832-44 but lower than the other reported values of 400 GPa and 390 (1g91) GPa. The disparity between the current measurements of M. L. Sattler, J. H, Kinney, E. Zywicz, R. Alani, and M, C. Nichols, "The modulus and earlier reports may stem from the difficulty of Microstructure of SCS-6 and SCS-8 SiC Reinforcing Fibers, ""Ceram. Eng. Sci Proc,l3,227-37( accurately determining contact areas and therefore propertie IE. Zywicz, J. H. Kinney, M. L. Sattler, T. M. Breunig, and M. C. Nichol the absolute magnitudes of modulus and hardness in Fig. 4 crast, 169, 247-53(1993)structures and Their Influence on Failure,"JMi- during indentations with significant elastic recovery. However, show a clear trend which includes dramatic variations in me 12X. J. Ning, P Pirouz, and S. C. Farmer, " Microchemical Analysis of the SCS-6 Silicon Carbide Fiber, " J. Am. Ceram Soc., 76, 2033-41(1993). chanical properties across the SCS SiC fibers Within the core SJ. I. Eldridge, J. P. Wining, T S. Davison, and M.-J. Pindera, the modulus and hardness are consistently low, but then drop in Strength of SCs-6 Silicon Carbide Fibers,"J.Am. Ceram Soc., 76, 3151-54 the graphitic core coating before rising by an order of magni-(943). tude within the first few micrometers of the sic sheath. this 4S, Sathish, J H. Cantrell, and W. T. Yost, "Radial Variations of Elastic sharp increase is attributable to the dramatic change in chem- Sathish, J H. Cantrell, and w. T. Yost, ""Scanning Acoustic Microscopy istry and structure in the SiC sheath close to the core. Across of SCS-6 Silicon Carbide Fiber, 'J. Am. Ceram Soc., 79, 209-212(1996) the bulk of the SiC sheath, where the chemistry and structure ary slowly, the modulus and hardness are smooth and con- w. C. Oliver and G M. Pha, "An Improved Technique for Determining tinuous and largely mimic the trend in fiber chemistry tion Experiments, J. Mater. Res, 7, 1564(1992 O. L. Blakslee, D. G. Proctor, E J. Seldin, G. B. S andT. w Elastic Constants of Compression-Annealed Pyrolytic Graphite, J. App. VI Conclusions Phs,4l,3373-82(1970 J. R. Dryden and G. R. Purdy, " The Effect of Graphite on the Mechanical The mechanical properties of SCS-6 Sic fibers were mea- Proper msaovald. H.J. Butt, S.A.C. Gould, C.B. Prater, and B. Drake, J.A. niques. Polished cross sections of fibers Berkovich diamond tip to depths between 50 and 450 nm. The Surface Elasticities with the Atomic Force Microscope, Nanotechnology, 2, It data from each indentation were analyze all. G. W. Marshall. Jr. and T P to determine youngs modulus and the hardness of the material urements of the Hardness and Elas. all of the major regions of these fibers: the carbon core, the an Dentin, J. Biomech. Eng, 118, graphitic core coating, the inner sheath, and the outer sheath The carbon core of the fibers was extremely compliant with a and Technology, Lit. S
the inner SiC sheath, Sathish et al. report modulus values which increase from a range of 220–310 GPa to a range of 315–405 GPa. Thus, our values for Young’s modulus across the SiC sheath (shown by Fig. 4) are in good agreement with the recent results of Sathish et al.15 In an earlier paper Sathish et al.14 used Auger spectroscopy data from Ning and Pirouz8 (Fig. 2) to predict variations of bulk and shear moduli across the inner and outer sheaths of SCS SiC fibers. They predicted that both moduli increase rapidly across the inner sheath as reported here. However, they also predicted a large, local minima in the inner SiC sheath. This is something which they did not see in their subsequent examination of the fibers.15 The results of these indentation experiments support the presence of a sharp rise in moduli, but like Sathish et al.’s later results they show no sign of a local minima. To help verify this finding, large areas of a fiber were scanned using an atomic force microscope (AFM) and an AC technique that yields a measure of material stiffness.20,21 No abrupt changes or local minima in stiffness were observed in the region of the inner sheath. Thus, all of the results reported here suggest that the variation in mechanical properties across the inner sheath is continuous everywhere, except close to the carbon core, with no identifiable local minima. The last region that was investigated within the SCS fibers is the outer sheath. This section consists of strongly textured SiC grains and is relatively uniform in chemistry. The ratio of C to Si atoms is close to 1.0 throughout the region, as shown in Fig. 2. Young’s modulus shows some scatter but is basically constant at an average value of 360 GPa. The distinct plateau in modulus in Fig. 4 is attributed to the uniform chemistry and microstructure of the region. Hardness also shows a plateau with an average value of 34 GPa. The magnitude of Young’s modulus, 360 GPa, is a little lower than expected for SiC. For single crystals of b-SiC the modulus varies from 379 GPa in the [100] crystallographic direction to 547 GPa in the [111] direction.22 A Voigt (isostrain) average for the modulus, based on single-crystal elastic constants, is 459 GPa, while the Reuss (isostress) average is 414 GPa. Thus, all of the expected values are slightly higher than the measured modulus. This difference may be due to nonideal stoichiometry throughout the outer sheath. Young’s modulus averaged over the full cross-sectional area of the fiber was calculated to be 333 GPa, which is close to the average moduli measured by Sathish et al.15 of 335–365 GPa, but lower than the other reported values of 400 GPa1 and 390 GPa.3 The disparity between the current measurements of modulus and earlier reports may stem from the difficulty of accurately determining contact areas, and therefore properties, during indentations with significant elastic recovery. However, the absolute magnitudes of modulus and hardness in Fig. 4 show a clear trend which includes dramatic variations in mechanical properties across the SCS SiC fibers. Within the core the modulus and hardness are consistently low, but then drop in the graphitic core coating before rising by an order of magnitude within the first few micrometers of the SiC sheath. This sharp increase is attributable to the dramatic change in chemistry and structure in the SiC sheath close to the core. Across the bulk of the SiC sheath, where the chemistry and structure vary slowly, the modulus and hardness are smooth and continuous and largely mimic the trend in fiber chemistry. VI. Conclusions The mechanical properties of SCS-6 SiC fibers were measured as a function of fiber radius using nanoindentation techniques. Polished cross sections of fibers were indented with a Berkovich diamond tip to depths between 50 and 450 nm. The force–displacement data from each indentation were analyzed to determine Young’s modulus and the hardness of the material in all of the major regions of these fibers: the carbon core, the graphitic core coating, the inner sheath, and the outer sheath. The carbon core of the fibers was extremely compliant with a Young’s modulus of 28 GPa and a hardness of 4.2 GPa. The modulus and hardness appeared to be uniform in this core, as expected. The graphitic coating on the core appeared to be weaker, though, with a modulus of 21 GPa and a hardness of 1.7 GPa. The inner sheath, which contained a varying chemistry, showed a sharp increase in stiffness and hardness from the inner core. The modulus and hardness increased radially within this region to 360 and 34 GPa, respectively. Most of this increase was confined to the first few micrometers of the inner SiC sheath, and thereafter there were no signs of local minima or decreases, just a gradual increase to the maximum values observed. The outer sheath, which contained a relatively uniform chemistry, was consistently stiff and hard. The average modulus and hardness were measured to be 360 and 34 GPa, which are similar to previous measurements or estimates. The average modulus for the full fiber was 333 GPa. The measured values of modulus and hardness clearly showed that the mechanical properties of SCS-6 fibers changed by an order of magnitude across their diameters. Such dramatic variations should be considered in modeling the behavior of these materials. Acknowledgments: We gratefully acknowledge Textron Speciality Materials for supplying the fibers, Professor W. D. Nix, Jerry English and Bob Kershaw for help during the initial testing. References 1 Textron Specialty Materials, Textron CO., Lowell, MA. 2 S. R. Nutt and F. E. Wawner, ‘‘Silicon Carbide Filaments: Microstructure,’’ J. Mater. Sci., 20, 1953–60 (1985). 3 J. A. DiCarlo and W. Williams, ‘‘Dynamic Modulus and Damping of Boron, Silicon Carbide, and Alumina Fibers,’’ Ceram. Eng. Sci. Proc., 1, 671–92 (1980). 4 J. A. DiCarlo, ‘‘High Temperature Properties of CVD Silicon Carbide Fibers’’; pp. 1–8 in Whisker and Fiber Toughened Ceramics. Edited by R. A. Bradley. ASM International, Metal Park, 1988. 5 M. Lancin, J. Thibault-Desseauc, and J. S. Bour, ‘‘Structure of the Interface in a SCS-6 Fiber/Ti-6A1-4V Composite,’’ J. Microsc. Spectrosc. Electron., 13, 503 (1988). 6 J. Chung; M.Sc. Thesis. Case Western Reserve University, Cleveland, OH, 1990. 7 X. J. Ning, P. Pirouz, K. P. D. Lagerlof, and J. DiCarlo, ‘‘The Structure of Carbon in Chemically Vapor Deposited SiC Monofilaments,’’ J. Mater. Res., 5, 2865–76 (1990). 8 X. J. Ning and P. Pirouz, ‘‘The Microstructure of SCS-6 SiC Fibers,’’ J. Mater. Res., 6, 2234–48 (1991). 9 R. T. Bhatt and D. R. Hull, ‘‘Microstructural and Strength Stability of CVD SiC Fibers in Argon Environment,’’ Ceram. Eng. Sci. Proc., 12, 1832–44 (1991). 10M. L. Sattler, J. H. Kinney, E. Zywicz, R. Alani, and M. C. Nichols, ‘‘The Microstructure of SCS-6 and SCS-8 SiC Reinforcing Fibers,’’ Ceram. Eng. Sci. Proc., 13, 227–37 (1992). 11E. Zywicz, J. H. Kinney, M. L. Sattler, T. M. Breunig, and M. C. Nichols, ‘‘Heterogeneous Fibre Microstructures and Their Influence on Failure,’’ J. Microsc., 169, 247–53 (1993). 12X. J. Ning, P. Pirouz, and S. C. Farmer, ‘‘Microchemical Analysis of the SCS-6 Silicon Carbide Fiber,’’ J. Am. Ceram. Soc., 76, 2033–41 (1993). 13J. I. Eldridge, J. P. Wiening, T. S. Davison, and M.-J. Pindera, ‘‘Transverse Strength of SCS-6 Silicon Carbide Fibers,’’ J. Am. Ceram. Soc., 76, 3151–54 (1993). 14S. Sathish, J. H. Cantrell, and W. T. Yost, ‘‘Radial Variations of Elastic Properties of SCS-6 Silicon Carbide Fiber,’’ J. Mater. Res., 9, 2298–303 (1994). 15S. Sathish, J. H. Cantrell, and W. T. Yost, ‘‘Scanning Acoustic Microscopy of SCS-6 Silicon Carbide Fiber,’’ J. Am. Ceram. Soc., 79, 209–212 (1996). 16Nano Instruments, Inc., Knoxville, TN 37914. 17W. C. Oliver and G. M. Pharr, ‘‘An Improved Technique for Determining Hardness and Elastic Modulus Using Load and Displacement Sensing Indentation Experiments,’’ J. Mater. Res., 7, 1564 (1992). 18O. L. Blakslee, D. G. Proctor, E. J. Seldin, G. B. Spence, and T. Weng, ‘‘Elastic Constants of Compression-Annealed Pyrolytic Graphite,’’ J. Appl. Phys., 41, 3373–82 (1970). 19J. R. Dryden and G. R. Purdy, ‘‘The Effect of Graphite on the Mechanical Properties of Cast Irons,’’ Acta Metall., 31, 1999–2006 (1989). 20P. Maivald, H. J. Butt, S. A. C. Gould, C. B. Prater, and B. Drake, J. A. Gurley, V. B. Elings, and P. K. Hansma, ‘‘Using Force Modulation to Image Surface Elasticities with the Atomic Force Microscope,’’ Nanotechnology, 2, 606–17 (1991). 21J. H. Kinney, M. Balooch, S. J. Marshall, G. W. Marshall, Jr., and T. P. Weihs, ‘‘Atomic Force Microscope Measurements of the Hardness and Elasticity of Peritubular and Intertubular Human Dentin,’’ J. Biomech. Eng., 118, 133 (1996). 22Landolt-Borstein, Numerical Data and Functional Relationships in Science and Technology, Lit. S. 36. Springer-Verlag, New York, 1969. h 116 Journal of the American Ceramic Society—Mann et al. Vol. 82, No. 1