Availableonlineatwww.sciencedirect.com Science Direct E噩≈RS ELSEVIER Joumal of the European Ceramic Society 29(2009)2849-2855 www.elsevier.comlocate/jeurceramsoc Investigation of thermal expansion of 3D-Stitched C-Sic composites Suresh Kumar Anil Kumar Anupam Shukla, G. Rohini Devi, A.K. Gupta Advanced Systems Laboratory, DRDO, Hyderabad 500058, India b Department of Chemical Engineering IIT Delhi, New Delhi 110016, Ina Received 28 October 2008: received in revised form 25 March 2009: accepted 3 April 2009 Available online 6 May 2009 Carbon fiber reinforced silicon carbide(C-SiC) composites are promising materials for a severe thermo-erosive environment 3D-stitched C-SiC composites were fabricated using liquid silicon infiltration. The infiltration was carried out at 1450-1650C for 10-120 min in vacuum. Coefficient of thermal expansion( CTE)of the composites was determined in in-plane and through-thickness directions in the temperature range from room range(15-4)10-6/C. The effect of siliconization conditions is higher in the through-thickness direction than in the in-plane ares in the temperature to 1050C. The in-plane Cte varies in the range(0.5-2)x10/C, while that in the through-thickness direction, it vari CTE values are lower than the values reported for chemical vapor impregnation based 3D C-SiC composites. An extensive microstructure stud was also carried out to understand the thermal expansion behavior of the composites. It was found out that Cte behavior is closely related to the position of the composite which in turn depends upon siliconization conditions. The best conditions were 1650 C and 120 min C 2009 Elsevier Ltd. all rights reserved Keywords: Composites; Thermal properties: Thermal expansion; 3D-stitched C-Sic 1. ntroduction better Cte in all the directions. These have been fabricated using coal-tar pitch as a carbon precursor. CtE depends on reinforce Liquid silicon infiltration (LSD) process was developed by ment orientation, composition(carbon-fibers, carbon-matrix, German Aerospace Centre(DLR) for processing bi-directional silicon-carbide(SiC)and silicon)and microstructure of the com- (2D)carbon-silicon-carbide( C-SiC)composites. The LSI pro- posite. The composition and microstructure of these composites cess is relatively economic and fast as compared to chemical would depend upon siliconization conditions and on the rate of vapor impregnation(CVI) and polymer impregnation processes. cooling. 6. 7 The temperature of silicon infiltration in the carbon Major applications of C-SiC composites include nose tip of performs has been reported to be in the range 1420-1600oC.3-6, 9 reusable space vehicles, leading edges of hypersonic vehicles, The infiltration time ranges from a few minutes to 2 h depending jet-vanes for thrust vectoring, and brake discs for high speed on the size of the C-C preform 5.6 CTE of composites can be automobiles. These products require uniform thermal prop- explained andestimated by the rule of mixtures but for complex erties in all the directions to meet the thermo-structural loads. fiber architectures estimation is difficult. Therefore it needs to Coefficient of thermal expansion(CTE) is an important desi be determined experimentally input. Many researchers have reported CtE data for LSI based 2D C-SiC composites. # It is much higher in the through- I.I. Aim and scope of the work thickness direction as compared to that in the in-plane direction, which may lead to thermal shock failure. It was decided to fabricate coal-tar pitch carbon precursor CTE of 2D C-Sic composites may be improved by intro- based 3D-stitched C-Sic composites which should have ver ducing carbon fibers in the third direction perpendicular to the close thermal expansion in the in-plane and in the through fabric. 3D-stitched C-Sic composites are fabricated to achieve a thickness directions For this purpose a standard procedure was developed for the preparation of 3D-stitched carbon-carbon(C-C) performs. Corresponding author. Tel: +91 40 24306994: fax: +91 40 24306498 These C-C preforms were siliconized at two different tempera E-mailaddress:sureshtanwar@rediffmail.com(SKumar) tures(1450 and 1650 C)and for two different time intervals(10 0955-2219 front matter@ 2009 Elsevier Ltd. All rights reserved. doi: 10. 1016/j-jeurceramsoc. 2009.04.005
Available online at www.sciencedirect.com Journal of the European Ceramic Society 29 (2009) 2849–2855 Investigation of thermal expansion of 3D-stitched C–SiC composites Suresh Kumar a,∗, Anil Kumar a, Anupam Shukla b, G. Rohini Devi a, A.K. Gupta b a Advanced Systems Laboratory, DRDO, Hyderabad 500058, India b Department of Chemical Engineering IIT Delhi, New Delhi 110016, India Received 28 October 2008; received in revised form 25 March 2009; accepted 3 April 2009 Available online 6 May 2009 Abstract Carbon fiber reinforced silicon carbide (C–SiC) composites are promising materials for a severe thermo-erosive environment. 3D-stitched C–SiC composites were fabricated using liquid silicon infiltration. The infiltration was carried out at 1450–1650 ◦C for 10–120 min in vacuum. Coefficient of thermal expansion (CTE) of the composites was determined in in-plane and through-thickness directions in the temperature range from room temperature to 1050 ◦C. The in-plane CTE varies in the range (0.5–2) × 10−6/ ◦C, while that in the through-thickness direction, it varies in the range (1.5–4) × 10−6/ ◦C. The effect of siliconization conditions is higher in the through-thickness direction than in the in-plane direction. The CTE values are lower than the values reported for chemical vapor impregnation based 3D C–SiC composites. An extensive microstructure study was also carried out to understand the thermal expansion behavior of the composites. It was found out that CTE behavior is closely related to the composition of the composite which in turn depends upon siliconization conditions. The best conditions were 1650 ◦C and 120 min. © 2009 Elsevier Ltd. All rights reserved. Keywords: Composites; Thermal properties; Thermal expansion; 3D-stitched C–SiC 1. Introduction Liquid silicon infiltration (LSI) process was developed by German Aerospace Centre (DLR) for processing bi-directional (2D) carbon–silicon-carbide (C–SiC) composites.1 The LSI process is relatively economic and fast as compared to chemical vapor impregnation (CVI) and polymer impregnation processes. Major applications of C–SiC composites include nose tip of reusable space vehicles, leading edges of hypersonic vehicles, jet-vanes for thrust vectoring, and brake discs for high speed automobiles.2,3 These products require uniform thermal properties in all the directions to meet the thermo-structural loads. Coefficient of thermal expansion (CTE) is an important design input. Many researchers have reported CTE data for LSI based 2D C–SiC composites.4,5 It is much higher in the throughthickness direction as compared to that in the in-plane direction, which may lead to thermal shock failure. CTE of 2D C–SiC composites may be improved by introducing carbon fibers in the third direction perpendicular to the fabric. 3D-stitched C–SiC composites are fabricated to achieve a ∗ Corresponding author. Tel.: +91 40 24306994; fax: +91 40 24306498. E-mail address: sureshtanwar@rediffmail.com (S. Kumar). better CTE in all the directions. These have been fabricated using coal-tar pitch as a carbon precursor. CTE depends on reinforcement orientation, composition (carbon-fibers, carbon-matrix, silicon-carbide (SiC) and silicon) and microstructure of the composite. The composition and microstructure of these composites would depend upon siliconization conditions and on the rate of cooling.6,7 The temperature of silicon infiltration in the carbon performs has been reported to be in the range 1420–1600 ◦C.3–6,9 The infiltration time ranges from a few minutes to 2 h depending on the size of the C–C preform.5,6 CTE of composites can be explained and estimated by the rule of mixtures8 but for complex fiber architectures estimation is difficult. Therefore, it needs to be determined experimentally. 1.1. Aim and scope of the work It was decided to fabricate coal–tar pitch carbon precursor based 3D-stitched C–SiC composites which should have very close thermal expansion in the in-plane and in the throughthickness directions. For this purpose a standard procedure was developed for the preparation of 3D-stitched carbon–carbon (C–C) performs. These C–C preforms were siliconized at two different temperatures (1450 and 1650 ◦C) and for two different time intervals (10 0955-2219/$ – see front matter © 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jeurceramsoc.2009.04.005
S Kumar et al. / Journal of the European Ceramic Society 29(2009)2849-2855 三1000 Fig. 1. A typical siliconization cycle for blocks b-7 and b-8 Fig. 2. XRD image of the C-SiC and 120 min), respectively, to see the effect of temperature and of these blocks were cut to obtain eight different blocks of time on siliconization and also on thermal expansion behavior sizes 75 mm x 50 mm x 20 mm to 150 mm x 50 mm x 50mm of the composites. and were named b-l to b-8. These eight blocks were siliconized It was also felt necessary to correlate the thermal expansion at four siliconization conditions to obtain 3DC-SiC composites behavior with the chemical composition and the other charac- b-I and b-2 were siliconized at 1450 C for 10 min, b-3 and b-4 teristics of the composites at 1450C for 120min. b-5 and b-6 at 1650C for 10 min, and b-7 and b-8 were siliconized at 1650 C for 120 min. Represen 2. Experimental tative heating and cooling cycle for blocks b-7 and b-8 is shown in Fig. 1. The specific siliconization conditions were arrived at 2.1. Preparation of 3 D-stitched C-SiC composites keeping in mind the process requirements and the siliconization furnace limitations 2. Raw materials 3k PAn fiber tow based 8H satin fabric having 24 end 2. 2. Analysis inch and an average thickness of 0.45 mm was used for making the preforms. Pan based 6k carbon-fiber tows were used for The C-Sic composites were analyzed: (i) by XRD to dete stitching Indigenously available coal-tar pitch having softening mine phases present; (ii) chemical composition was determined point between 85 and 120C and coking value 45-50% was by acid digestion method; (ii) pore structure and porosity were used as a carbon precursor Indigenously available silicon metal determined by mercury porosimetry; (iv)microstructure by opti- of purity >98%o was used for siliconization cal and scanning electron microscopy; (v)CTE was determined 2.1.2. Procedure Fibrous preforms were prepared by stitching several lay- 2.2.1. XRD analysis ers of carbon fabric by carbon-fiber tows. Number of stitches XRD was carried out for all the composite blocks. A repre was maintained in the range 350-400 per 100 cm. The pre- sentative XRD pattern of b-l is shown in Fig. 2. The patterns forms were rigidized by vacuum infiltration of coal-tar pitch were similar in all the blocks. The results of the XRD studies at 200-300 C followed by carbonization at 900-1000oC, and will be discussed later. graphitization at 2400-2600C in the nitrogen atmosphere. The rigidized preforms were further densified by a hot-iso- 2.2.2. Chemical composition determination static-pressure-impregnation-carbonization(HIPIC) process at The composite blocks contain carbon fibers and un-reacted 800C and 1000 bar. These were termed as C-C performs; their carbon matrix, SiC, and residual silicon. The un-reacted car- density was found to lie in the range 1.55-1.60 g/cm. Some bon fibers and carbon matrix were considered as C-C Relative the C-SiC composite blocks after siliconization S No. Temp, C Time, min C-C perform density, g/cm C-SiC density, g/cm Open porosity, C-C, wt/wt SiC, wt/wt Silicon, wt/wt 2.19 4.86 0.568 0.335 0.097 1450 2.19 4.81 0.592 0.335 0073 1450 2.33 3.71 0.538 0.359 0.102 1450 0.359 0.111 1650 0.348 1650 2.17 5.29 0.599 0.348 0053 b-7 1650 159 0.563 1650 0.559 0.384 0057
2850 S. Kumar et al. / Journal of the European Ceramic Society 29 (2009) 2849–2855 Fig. 1. A typical siliconization cycle for blocks b-7 and b-8. and 120 min), respectively, to see the effect of temperature and time on siliconization and also on thermal expansion behavior of the composites. It was also felt necessary to correlate the thermal expansion behavior with the chemical composition and the other characteristics of the composites. 2. Experimental 2.1. Preparation of 3D-stitched C–SiC composites 2.1.1. Raw materials 3k PAN fiber tow based 8H satin fabric having 24 ends per inch and an average thickness of 0.45 mm was used for making the preforms. PAN based 6k carbon–fiber tows were used for stitching. Indigenously available coal–tar pitch having softening point between 85 and 120 ◦C and coking value 45–50% was used as a carbon precursor. Indigenously available silicon metal of purity >98% was used for siliconization. 2.1.2. Procedure Fibrous preforms were prepared by stitching several layers of carbon fabric by carbon–fiber tows. Number of stitches was maintained in the range 350–400 per 100 cm2. The preforms were rigidized by vacuum infiltration of coal–tar pitch at 200–300 ◦C followed by carbonization at 900–1000 ◦C, and graphitization at 2400–2600 ◦C in the nitrogen atmosphere. The rigidized preforms were further densified by a hot-isostatic-pressure-impregnation-carbonization (HIPIC) process at 800 ◦C and 1000 bar. These were termed as C–C performs; their density was found to lie in the range 1.55–1.60 g/cm3. Some Fig. 2. XRD image of the C–SiC composite block-1. of these blocks were cut to obtain eight different blocks of sizes 75 mm × 50 mm × 20 mm to 150 mm × 50 mm × 50 mm and were named b-1 to b-8. These eight blocks were siliconized at four siliconization conditions to obtain 3D C–SiC composites: b-1 and b-2 were siliconized at 1450 ◦C for 10 min, b-3 and b-4 at 1450 ◦C for 120 min, b-5 and b-6 at 1650 ◦C for 10 min, and b-7 and b-8 were siliconized at 1650 ◦C for 120 min. A representative heating and cooling cycle for blocks b-7 and b-8 is shown in Fig. 1. The specific siliconization conditions were arrived at keeping in mind the process requirements and the siliconization furnace limitations. 2.2. Analysis The C–SiC composites were analyzed: (i) by XRD to determine phases present; (ii) chemical composition was determined by acid digestion method; (iii) pore structure and porosity were determined by mercury porosimetry; (iv) microstructure by optical and scanning electron microscopy; (v) CTE was determined by dilatometer. 2.2.1. XRD analysis XRD was carried out for all the composite blocks. A representative XRD pattern of b-1 is shown in Fig. 2. The patterns were similar in all the blocks. The results of the XRD studies will be discussed later. 2.2.2. Chemical composition determination The composite blocks contain carbon fibers and un-reacted carbon matrix, SiC, and residual silicon. The un-reacted carbon fibers and carbon matrix were considered as C–C. Relative Table 1 Density and composition of the C–SiC composite blocks after siliconization. S. No. Temp., ◦C Time, min C–C perform density, g/cm3 C–SiC density, g/cm3 Open porosity, % C–C, wt/wt SiC, wt/wt Silicon, wt/wt b-1 1450 10 1.58 2.19 4.86 0.568 0.335 0.097 b-2 1450 10 1.59 2.19 4.81 0.592 0.335 0.073 b-3 1450 120 1.60 2.33 3.71 0.538 0.359 0.102 b-4 1450 120 1.60 2.30 3.68 0.530 0.359 0.111 b-5 1650 10 1.59 2.23 5.31 0.582 0.348 0.069 b-6 1650 10 1.58 2.17 5.29 0.599 0.348 0.053 b-7 1650 120 1.59 2.19 5.42 0.563 0.384 0.053 b-8 1650 120 1.60 2.20 5.46 0.559 0.384 0.057
S. Kumar et al. /Journal of the European Ceramic Society 29(2009)2849-2855 2851 12e-3 microscope( SEM)to understand the effect of siliconization con- 1,1e3 ditions. A few representative optical and SEM images of block E82e4 siliconized at different conditions are shown in Fig. 5. Simi lar microstructures were observed for all the blocks. However 界5.9e-4 the distribution of un-reacted silicon SiC and carbon was some- what different, i.e. residual silicon was observed highest in block 3.5e-4 b-1 while it the least in the block b-8. Discussion about the thh microstructure is given Section 3.2. 00100200501020.5125102050100200 3. Results and discussion Fig. 3. Pore size distribution of C-SiC composite siliconized at 1450.C, 3. 1. Effect of siliconization conditions on composition and other characteristics KRD images show three phases, viz., B-SiC, carbon, and sili- amount of each was determined by chemical analysis. A small con(Fig. 2 ), all the three phases were present in all the composite mass of each composite block was powdered. About 5 g powder blocks. It is evident that the infiltrated silicon reacts with the car- was digested in 50 ml acid mixture(HF: HNO3, 4: 1)for 24 h Sil- bon matrix and forms silicon carbide. However the infiltrated icon in the composite powder dissolved in the acid. The residue silicon is not fully converted to SiC. Therefore SiC, carbon, and carbon and SiC) was obtained by filtering the solution Mass of residual silicon are present in all the composite blocks Sic was determined by heating the residue in air at about 600oC thus burning off the carbon. .o Composition of all blocks is posite blocks: however pores below I um diameter account f given in Table 1 about 80% of the total pore volume. Open porosity was found 2.23 Pore structure determination to be in the range 3. 68-5.46%. It was the least in the blocks b-3 and b-4 and the highest in the blocks b-7and b-8. Pore size distribution and porosity studies were carried out The density of the composite blocks varies in the range by mercury porosimetry using Quantachrome instrument. Pore 2.19-2.33 g/cm(Table 1). It is the highest for the blocks b- structure of all the specimens was similar. A representative pore 3 and b-4(1450 C and 120 min)and the least for blocks b-7 size distribution is shown in Fig 3. Porosity values of all the and b-8(1650.C and 120 min). Porosity and density are closely blocks are given in Table 1. The results are discussed in Section related to each other. Chemical composition of all the composite blocks was also 2. 2.4. cte determination determined. SiC fraction was the highest in the blocks b-7 and Several test specimens of diameter 6 mm and length 25 mm 2(1450 C, 10min). Sic fraction increased with siliconization were cut in in-plane(or parallel)and through-thickness (r-T) time and temperature. The un-reacted silicon was the highest in directions from the composite blocks(b-I to b-8). The configu- the blocks b-3 and b-4 and the least in the blocks b-7 and b-8.The ration of the specimens is shown in Fig 4 CTE measurements un-reacted silicon increases with time at 1450 C: it decreases were conducted in nitrogen atmosphere, from room temperature with time at 1650%C. These observations indicate that the rate to 1050C using DIL 402C Dilatometer(NETZSCH Germany). of formation of SiC is much slower as compared to the rate of Temperature of the specimens was raised at the rate of 5C/min infiltration of silicon The change in the specimen length(resulting due to temperature It may be inferred that chemical composition and hence all rise)was measured. The coefficient of the linear thermal expan- the other characteristics are governed by siliconization condi sion was calculated as CTE=(△L/L△T; where△ LL is the tions fractional change in length of the specimen due to temperature change△T 3. 2. Microstructure studie 2.5. Microstructure analysis Microstructure studies were carried out for all the blocks with Two different types of microstructures are observed in all the the help of an optical microscope and also a scanning electron are embedded in the residual silicon. There is no evidence of sic formation at the carbon fiber surface. Similar microstructures have also been reported in literature. Micro-cracks were observed in the matrix of all the ite blocks: the cracks are mainly in the direction perpendicular Stitches to that of the carbon fibers The cracks in the blocks would have eveloped due to thermal expansion-mismatch of the residual Fig. 4. Configuration of test specimens for CTE determination: (a) through- silicon and the carbon fibers during cooling after siliconiza thickness (T-T)and (b)in-plane(or parallel tion. Had the cracks been present before siliconization, those
S. Kumar et al. / Journal of the European Ceramic Society 29 (2009) 2849–2855 2851 Fig. 3. Pore size distribution of C–SiC composite siliconized at 1450 ◦C, 120 min. amount of each was determined by chemical analysis. A small mass of each composite block was powdered. About 5 g powder was digested in 50 ml acid mixture (HF:HNO3, 4:1) for 24 h. Silicon in the composite powder dissolved in the acid. The residue (carbon and SiC) was obtained by filtering the solution. Mass of SiC was determined by heating the residue in air at about 600 ◦C thus burning off the carbon.9,10 Composition of all blocks is given in Table 1. 2.2.3. Pore structure determination Pore size distribution and porosity studies were carried out by mercury porosimetry using Quantachrome instrument. Pore structure of all the specimens was similar. A representative pore size distribution is shown in Fig. 3. Porosity values of all the blocks are given in Table 1. The results are discussed in Section 3.1. 2.2.4. CTE determination Several test specimens of diameter 6 mm and length 25 mm were cut in in-plane (or parallel) and through-thickness (T-T) directions from the composite blocks (b-1 to b-8). The configuration of the specimens is shown in Fig. 4. CTE measurements were conducted in nitrogen atmosphere, from room temperature to 1050 ◦C using DIL 402C Dilatometer (NETZSCH Germany). Temperature of the specimens was raised at the rate of 5 ◦C/min. The change in the specimen length (resulting due to temperature rise) was measured. The coefficient of the linear thermal expansion was calculated as CTE = (L/L)/T; where L/L is the fractional change in length of the specimen due to temperature change T. 2.2.5. Microstructure analysis Microstructure studies were carried out for all the blocks with the help of an optical microscope and also a scanning electron Fig. 4. Configuration of test specimens for CTE determination: (a) throughthickness (T-T) and (b) in-plane (or parallel). microscope (SEM) to understand the effect of siliconization conditions. A few representative optical and SEM images of blocks siliconized at different conditions are shown in Fig. 5. Similar microstructures were observed for all the blocks. However, the distribution of un-reacted silicon, SiC and carbon was somewhat different, i.e. residual silicon was observed highest in block b-1 while it the least in the block b-8. Discussion about the microstructure is given Section 3.2. 3. Results and discussion 3.1. Effect of siliconization conditions on composition and other characteristics XRD images show three phases, viz., -SiC, carbon, and silicon (Fig. 2), all the three phases were present in all the composite blocks. It is evident that the infiltrated silicon reacts with the carbon matrix and forms silicon carbide. However the infiltrated silicon is not fully converted to SiC. Therefore SiC, carbon, and residual silicon are present in all the composite blocks. Pore size varies in the range 0.001–380 m in all the composite blocks: however pores below 1 m diameter account for about 80% of the total pore volume. Open porosity was found to be in the range 3.68–5.46%. It was the least in the blocks b-3 and b-4 and the highest in the blocks b-7and b-8. The density of the composite blocks varies in the range 2.19–2.33 g/cm3 (Table 1). It is the highest for the blocks b- 3 and b-4 (1450 ◦C and 120 min) and the least for blocks b-7 and b-8 (1650 ◦C and 120 min). Porosity and density are closely related to each other. Chemical composition of all the composite blocks was also determined. SiC fraction was the highest in the blocks b-7 and b-8 (1650 ◦C, 120 min) and the least in the blocks b-1 and b- 2 (1450 ◦C, 10 min). SiC fraction increased with siliconization time and temperature. The un-reacted silicon was the highest in the blocks b-3 and b-4 and the least in the blocks b-7 and b-8. The un-reacted silicon increases with time at 1450 ◦C; it decreases with time at 1650 ◦C. These observations indicate that the rate of formation of SiC is much slower as compared to the rate of infiltration of silicon. It may be inferred that chemical composition and hence all the other characteristics are governed by siliconization conditions. 3.2. Microstructure studies Two different types of microstructures are observed in all the blocks: (i) carbon is surrounded by the SiC and (ii) SiC crystals are embedded in the residual silicon. There is no evidence of SiC formation at the carbon fiber surface. Similar microstructures have also been reported in literature.11,12 Micro-cracks were observed in the matrix of all the composite blocks: the cracks are mainly in the direction perpendicular to that of the carbon fibers. The cracks in the blocks would have developed due to thermal expansion-mismatch of the residual silicon and the carbon fibers during cooling after siliconization. Had the cracks been present before siliconization, those
S Kumar et al. / Journal of the European Ceramic Society 29(2009)2849-2855 would have been filled by the molten silicon during siliconiza- 3.3. Coefficient of thermal expansion tion. It is reported that the initial formation of SiC layer is very fast, and its subsequent growth is controlled by diffusion of sil- CtE of the composite would depend on the CTEs of indi- icon through it. There is a volume misfit between carbon and vidual constituents, viz., carbon fibers, carbon matrix, Sic the product SiC 3. Therefore, some of the SiC formed may and residual silicon. CTE of silicon and Sic are very close break into small particles and move to liquid silicon pool. It and have been reported to be in the range(4-6)x 10-6/oCI4 is also evident that the carbon fibers remain unaffected by sili- and(3-5.1)x10-broC, respectively. 5-I8C-Ccomposites show conization(Fig. 5e and f). The fibers are surrounded by graphitic anisotropic CTE. For coal-tar pitch based 2D C-C composites it carbon, which protects the carbon fibers from the carbon-silicon is reported to be in the range(0-0.27)x 10-b/C in the in-plane reaction direction while(3.0-9. 1)x 10/'c in the through-thickness SHicon Carbon siC Silicon Carbon 50 Hm 50 Am 产 :50pm e Perfect fibe nd ne traces of surface Fig. 5. Microstructures of C-SiC composites processed at different conditions: (a)block-1,(b)block-3. (c)block-5 and(d-f) block-8
2852 S. Kumar et al. / Journal of the European Ceramic Society 29 (2009) 2849–2855 would have been filled by the molten silicon during siliconization. It is reported that the initial formation of SiC layer is very fast, and its subsequent growth is controlled by diffusion of silicon through it. There is a volume misfit between carbon and the product SiC13. Therefore, some of the SiC formed may break into small particles and move to liquid silicon pool. It is also evident that the carbon fibers remain unaffected by siliconization (Fig. 5e and f). The fibers are surrounded by graphitic carbon, which protects the carbon fibers from the carbon–silicon reaction. 3.3. Coefficient of thermal expansion CTE of the composite would depend on the CTEs of individual constituents, viz., carbon fibers, carbon matrix, SiC and residual silicon. CTE of silicon and SiC are very close and have been reported to be in the range (4–6) × 10−6/ ◦C14 and (3–5.1) × 10−6/ ◦C, respectively.15–18 C–C composites show anisotropic CTE. For coal–tar pitch based 2D C–C composites it is reported to be in the range (0–0.27) × 10−6/ ◦C in the in-plane direction while (3.0–9.1) × 10−6/ ◦C in the through-thickness Fig. 5. Microstructures of C–SiC composites processed at different conditions: (a) block-1, (b) block-3, (c) block-5 and (d–f) block-8
S Kumar et al. Journal of the European Ceramic Sociery 29(2009)2849-2855 2853 1,6E03 Block-3 IL 50E06 Block-5II -.-Block-7ll 一2 DC-Cl Ref19 S30E06 60E04 2.0E06 Silicon Ref. 14 4.0E04 2.0E·04 CVD ID CSIC‖ 0.0E+00 10001200 Fig 8. Relationship between in-plane CTE and temperature. Fig. 6. Relationship between in-plane linear expansion and temperature 2. x 10-6T-9x 10-4. It shows that the expansion of the direction. The expansion of the PAN based carbon fibers C-Sic composite blocks is about twice that of 2D C-Ccom- treated at 2800 C is negative below 400C, and positive at tem- posites at entire temperature range peratures higher than 400 C20. 2. It varies from-1 x 10-b/oC at 150.C to 1.5x 10-bC at 1000C2. Also, CTE of such 3.3.2. CTE composites is affected by the thermal-treatment history and the From the AL/L vs temperature data, CTE was calculate rate of cooling nd compared with the values of (a)constituents of the C-SiC composites reported in the literature, with 2D C-C composite 3.3.1. Expansion in the in-plane direction and with CTEs of CVD based C-Sic composites(Fig 8) In the present study expansion was measured for four blocks CTE is about 1.25 x 10-6/C for all the blocks in the present b-1,b-3,b-5 and b-7(Fig. 6). It is very similar in all the four study, it is fairly close to that of PAN carbon fibers as blocks in spite of the variation in their chemical composition. mentioned earlier. It is much lower than that for silicon as well as The fiber architecture of these blocks in in-plane direction is very Sic. itis also lower than that for cvd based ID and 2DC-Sic similar to that of the 2DC-Ccomposites9. Therefore expansion However, the CTE of C-Sic composite blocks is always higher of the above composite blocks has been compared with that of than that of 2D C-C composite. They decrease in both the cases a 2D C-C composite in the same figure up to 250C. The values are almost constant between 250 and It is evident that the expansion of 3D-stitched composites is 650C. Beyond 650C, both increase marginally more than that for 2D composite over the entire temperature The observed reduction in CTE up to 250C may be due to: (1) range. The expansion is small up-to 600C and relatively higher the expansion is absorbed by the micro-cracks and(ii)negative at higher temperatures expansion(or contraction) of carbon fibers in this temperature For quantitative comparison, the thermal expansion data were well correlated by straight line equations in two temperature segments:(1)between 100 and 600C and(2)between 600 and (1)During siliconization the C-C blocks get filled with molten 1050°C(F1g.7) silicon. Some part of the silicon gets converted into SiC In segment-1, expansion of the 2D C-C composite while the remaining is present as such into the pores of the is almost zero while that of C-Sic composite blocks blocks. SiC and the residual silicon shrink uniformly dur- is 5x 10-7T+5x 10-, where T is the temperature in ing cooling. Silicon is brittle as compared to SiC and carbon oC. In segment-2, expansion of 2D C-C composites is fibers. The matrix of the C-SiC composite blocks expe 1 x 10 T while expansion of C-Sic composite blocks ences thermal stress due to cte mismatch which results into micro-cracks in the residual silicon(Fig. 5(a-d). A 1.5E-03 low expansion in the temperature range 100-300C shows y=204E06901E04aL R2=9.86E01 at the thermal stresses are released and the micro-cracks 10E03 3D-stitched ret filled 50E04 (2)As shown in Fig. 4b, the direction of the thermal expansion measurement is parallel to the axis of warp fibers and axis of 0 the weft fibers is in perpendicular direction. Third direction y=-BE-08X+ 5E-D fibers hold the stack of the fabrics together. volume frac- 50E04 tion of the third direction carbon fibers is about 0.07 of total Y=1E06X0.000 2DC-C composite reinforcement. Carbon-fibers shrink more in the radial direc- 10E03 tion and negligibly along the axis. ,0The expansion of the 10001200 composite blocks in the axial direction is mainly controlle by the expansion of the warp carbon-fibers as it the least Fig. 7. Thermal expansion comparison of 2D C-C and C-sic among all the constituents of C-SiC composites. Therefore
S. Kumar et al. / Journal of the European Ceramic Society 29 (2009) 2849–2855 2853 Fig. 6. Relationship between in-plane linear expansion and temperature. direction.19 The expansion of the PAN based carbon fibers treated at 2800 ◦C is negative below 400 ◦C, and positive at temperatures higher than 400 ◦C20,21. It varies from −1 × 10−6/ ◦C at 150 ◦C to 1.5 × 10−6/ ◦C at 1000 ◦C21. Also, CTE of such composites is affected by the thermal-treatment history and the rate of cooling. 3.3.1. Expansion in the in-plane direction In the present study expansion was measured for four blocks b-1, b-3, b-5 and b-7 (Fig. 6). It is very similar in all the four blocks in spite of the variation in their chemical composition. The fiber architecture of these blocks in in-plane direction is very similar to that of the 2D C–C composites19. Therefore expansion of the above composite blocks has been compared with that of a 2D C–C composite in the same figure. It is evident that the expansion of 3D-stitched composites is more than that for 2D C–C composite over the entire temperature range. The expansion is small up-to 600 ◦C and relatively higher at higher temperatures. For quantitative comparison, the thermal expansion data were well correlated by straight line equations in two temperature segments; (1) between 100 and 600 ◦C and (2) between 600 and 1050 ◦C (Fig. 7). In segment-1, expansion of the 2D C–C composite is almost zero while that of C–SiC composite blocks is 5 × 10−7 T + 5 × 10−5, where T is the temperature in ◦C. In segment-2, expansion of 2D C–C composites is 1 × 10−6 T while expansion of C–SiC composite blocks is Fig. 7. Thermal expansion comparison of 2D C–C and C–SiC. Fig. 8. Relationship between in-plane CTE and temperature. 2.04 × 10−6 T − 9 × 10−4. It shows that the expansion of the C–SiC composite blocks is about twice that of 2D C–C composites at entire temperature range. 3.3.2. CTE From the L/L vs. temperature data, CTE was calculated and compared with the values of (a) constituents of the C–SiC composites reported in the literature, with 2D C–C composite and with CTE’s of CVD based C–SiC composites (Fig. 8). CTE is about 1.25 × 10−6/ ◦C for all the blocks in the present study, it is fairly close to that of PAN based carbon fibers as mentioned earlier. It is much lower than that for silicon as well as SiC. It is also lower than that for CVD based 1D and 2D C–SiC. However, the CTE of C–SiC composite blocks is always higher than that of 2D C–C composite. They decrease in both the cases up to 250 ◦C. The values are almost constant between 250 and 650 ◦C. Beyond 650 ◦C, both increase marginally. The observed reduction in CTE up to 250 ◦C may be due to: (i) the expansion is absorbed by the micro-cracks and (ii) negative expansion (or contraction) of carbon fibers in this temperature range.19 (1) During siliconization the C–C blocks get filled with molten silicon. Some part of the silicon gets converted into SiC while the remaining is present as such into the pores of the blocks. SiC and the residual silicon shrink uniformly during cooling. Silicon is brittle as compared to SiC and carbon fibers. The matrix of the C–SiC composite blocks experiences thermal stress due to CTE mismatch which results into micro-cracks in the residual silicon (Fig. 5(a–d)). A low expansion in the temperature range 100–300 ◦C shows that the thermal stresses are released and the micro-cracks get filled. (2) As shown in Fig. 4b, the direction of the thermal expansion measurement is parallel to the axis of warp fibers and axis of the weft fibers is in perpendicular direction. Third direction fibers hold the stack of the fabrics together. Volume fraction of the third direction carbon fibers is about 0.07 of total reinforcement. Carbon-fibers shrink more in the radial direction and negligibly along the axis.19,20 The expansion of the composite blocks in the axial direction is mainly controlled by the expansion of the warp carbon-fibers as it the least among all the constituents of C–SiC composites. Therefore