《复合材料 Composites》课程教学资源（学习资料）第五章 陶瓷基复合材料_polymer pyrolysis and liquid

ELSEVIER Journal of Nuclear Materials 219(1995)26-30 Ceramic matrix composites using polymer pyrolysis and liquid densification processing H.O. Davis a.D. R. Petrak b Kaiser Aerotech, San Leandro, CA, USA Dow Corning Corporation, midland, MI, USA Abstract The polymer precursor approach for manufacture of ceramic matrix composites( CMCs)is both flexible and tailorable to shape and engineering requirements. The tailorability includes a wide range of reinforcements, polymer matrix precursors and fillers, Processing is selected based on cure/pressure requirements to best produce the required shape radii, fiber volume and fiber orientation. Combinations of tooling used for cure/pressure applica 1. Introduction 2. Toughness Producing ceramic matrix composites(CMCs) with There are many classical ways to achieve toughness preceramic polymer derived matrices reinforced with in ceramic composites. However, the two approaches eramic and carbon continuous filaments began in the discussed for this paper are primarily (1)use of a weak early 1980s. This process is normally referred to as the interface between fiber and matrix, and (2) use of polymer impregnation pyrolysis or fiber with a significantly higher modulus than the ce Much of the development work was accomplished on ramic matrix. For the early work at dC and Ka,a company funded research and development; therefore, carbon coating was used which worked satisfactorily many efforts were not well publicized. The approach at However, for temperatures above approximately 850F Kaiser Aerotech(KA)and Dow Corning Corporation (450C)for long periods of exposure, the carbon inter (DC) began independently. These efforts were com- face will oxidize. Subsequent development has led to bined in a joint venture in 1992 under the name of use of an advanced interface. Additional work is pro- Kaiser Ceramic Composites(KCC). Two primary mate- ceeding at dC to develop other interface coatings ial systems are being produced he joint venture The PiP process also has varying percentages of The material system data contained herein is called open porosity. However, the effects of porosity on Sylramic S-202. This material system is fabricable into toughness have not been sufficiently quantified to dis- complex shapes very similar to reinforced thermosets cuss porosity contribution to toughness. It has been and thermoplastics demonstrated that without controlled interfacial bond though the total processing methodology is dis- ing, a porous composite will not provide sufficient cussed,the emphasis of this discussion is the com- toughness. This behavior was demonstrated by dr Phil paction process for achieving high fiber volume in Chen while at KA and Ron boisvert and Ross K. complex shapes particularly T sections, ribs and sharp Hutter at Rensselaer Polytechnic Institute(RPD)in the radii. Process methodology tooling for tubular struc- mid eighties. Toughness was also demonstrated by us- tures, and aerospace structures such as leading edges gh""modulus Silicon Carbide fiber and combustor places have been previously developed In summary, the CMCs discussed herein would not 0022-3115/95/$09.50@ 1995 Elsevier Science B V. All rights reserved SSDⅠ0022-3115(94)00394-7

ELSEVIER Journal of Nuclear Materials 219 (1995) 26-30 Ceramic matrix composites using polymer pyrolysis and liquid densification processing H.O. Davis a, D.R. Petrak b a Kaiser Aerotech, San Leandro, CA, USA b Dow Coming Corporation, Midland, MI, USA Abstract The polymer precursor approach for manufacture of ceramic matrix composites (CMCs) is both flexible and tailorable to shape and engineering requirements. The tailorability includes a wide range of reinforcements, polymer matrix precursors and fillers. Processing is selected based on cure/pressure requirements to best produce the required shape, radii, fiber volume and fiber orientation. Combinations of tooling used for cure/pressure applica￾tions are discussed and fabricated components are shown. 1. Introduction 2. Toughness Producing ceramic matrix composites (CMCs) with preceramic polymer derived matrices reinforced with ceramic and carbon continuous filaments began in the early 1980s. This process is normally referred to as the polymer impregnation pyrolysis or the PIP process. Much of the development work was accomplished on company funded research and development; therefore, many efforts were not well publicized. The approach at Kaiser Aerotech (KA) and Dow Corning Corporation (DC) began independently. These efforts were com￾bined in a joint venture in 1992 under the name of Kaiser Ceramic Composites (KCC). Two primary mate￾rial systems are being produced by the joint venture. The material system data contained herein is called Sylramic S-202. This material system is fabricable into complex shapes very similar to reinforced thermosets and thermoplastics. Although the total processing methodology is dis￾cussed, the emphasis of this discussion is the com￾paction process for achieving high fiber volume in complex shapes particularly T sections, ribs and sharp radii. Process methodology, tooling for tubular struc￾tures, and aerospace structures such as leading edges and combustor places have been previously developed. 0022-3115/95/$09.50 © 1995 Elsevier Science B.V. All rights SSDI 0022-3115(94)00394-7 There are many classical ways to achieve toughness in ceramic composites. However, the two approaches discussed for this paper are primarily (1) use of a weak interface between fiber and matrix, and (2) use of a fiber with a significantly higher modulus than the ce￾ramic matrix. For the early work at DC and KA, a carbon coating was used which worked satisfactorily. However, for temperatures above approximately 850°F (450°C) for long periods of exposure, the carbon inter￾face will oxidize. Subsequent development has led to use of an advanced interface. Additional work is pro￾ceeding at DC to develop other interface coatings. The PIP process also has varying percentages of open porosity. However, the effects of porosity on toughness have not been sufficiently quantified to dis￾cuss porosity contribution to toughness. It has been demonstrated that without controlled interracial bond￾ing, a porous composite will not provide sufficient toughness. This behavior was demonstrated by Dr. Phil Chen while at KA and Ron Boisvert and Ross K. Hutter at Rensselaer Polytechnic Institute (RPI) in the mid eighties. Toughness was also demonstrated by us￾ing "high" modulus Silicon Carbide fiber. In summary, the CMCs discussed herein would not reserved

H.O. Davis, D, R. Petrak /Joumal of Nuclear Materials 219(1995)26-30 Table 1 TENSILE STRESS (MPa) temperature properties of Sylramic S-202 ramic matrix composites(Values are for warp direction, plain ave,0.90 lay-up) Density, g/ 2.15 lexural strength, Ksi 12-15 Tensile strength, Ksi 10-13 02030405060708090100110120 Compression strength, Ksi TIME TO FAILURE (HOUR) Fig. 2. Tensile failure under constant load at 1000 C. CTE, ppm/C Emissivity at 200 nm STRESS (MPa) have high mechanical properties without the interface coating on the Nicalon filaments, further development is needed to reduce cost and obtain better stabilit 3. CMC properties 10+03100+04100E+05100+06 Typical room temperature properties of Sylramic ER OF FREQUENCY(Hz) S-202 are listed in Table 1. These and other properties discussed below are for 2D laminates of 8 Harness Fig 3. Low cycle fatigue properties at 1000/C. atin CG Nicalon fiber cloth reinforcement Test pan els were warp aligned symmetrical lay-ups. The poly- stress for this material was over 200 MPa at 0.4%0 mer-derived matrix has an SiNc chemistry with strain. The proportional limit was 80 MPa and 0.08% advanced, proprietary, interface chemistry. Properties strain. Similar stress-strain curves have been observed f Sylramic S-202 CMC were developed for aerospace applications where high temperature mechanical prop- at Rt and 1200 C. However, ultimate strength values are lower (180 Mpa) at RT and higher (240 MPa)at erties in air are of primary concern. It is expected that 1200 C. Time dependent properties such as stress rup- need to be made to develop optimum PIP CMC prop- zawada [1] and Linsay [2]. Tensile stress rupture be. A typical stress-strain curve for Sylramic $-202 at havior is shown in Fig. 2. No stress rupture was found 1000C in air is shown in Fig. 1. The ultimate tensile at 100 MPa(110 h). It should be noted that this stress level is well above the proportional limit shown in Fi 1. a similar result was found for fatigue life at 100 Mpa as shown in Fig 3 with failure not detected at 105 UTS 213 MPa cycles at 100 MPa. Although shorter life was observed for fatigue condition than for constant stress condition (stress rupture), it is significant that this CMc does survive beyond the proportional limit in oxidizing con 4. Manufacture of complex shapes Data bases can be generated for CMCs from flat laminates or panels. This has occurred within the in- 000.100,20030040 dustry; however, many applications, if not most, re- Strain (%) quire that shapes be made from the material system Fig. 1. Tensile stress/strain for Sylramic S-20 Sylramic S-202 is a formable system which can be made

H.O. Davis, D.R. Petrak /Journal of Nuclear Materials 219 (1995) 26-30 27 Table 1 Typical room temperature properties of Sylramic S-202 ce￾ramic matrix composites (Values are for warp direction, plain weave, 0.90 ° lay-up) Sylramic ® S-202 Density, g/cm 3 2.15 Open porosity < 7% Flexural strength, Ksi 55 Flexural modulus, Msi 12-15 Tensile strength, Ksi 28-35 Tensile modulus, Msi 10-13 Compression strength, Ksi 65 Bearing strength, Ksi 50 CTE, ppm/°C 4.3 Interlaminar shear, Psi 4000 Emissivity at 200 nm 0.93 have high mechanical properties without the interface coating on the Nicalon filaments. Further development is needed to reduce cost and obtain better stability. 3. CMC properties Typical room temperature properties of Sylramic S-202 are listed in Table 1. These and other properties discussed below are for 2D laminates of 8 Harness Satin CG Nicalon fiber cloth reinforcement. Test pan￾els were warp aligned symmetrical lay-ups. The poly￾mer-derived matrix has an SiNC chemistry with an advanced, proprietary, interface chemistry. Properties of Sylramic S-202 CMC were developed for aerospace applications where high temperature mechanical prop￾erties in air are of primary concern. It is expected that some changes in CMC chemistry and processing would need to be made to develop optimum PIP CMC prop￾erties for nuclear applications. A typical stress-strain curve for Sylramic S-202 at 1000°C in air is shown in Fig. 1. The ultimate tensile -fi￾Ct. 250 .... i .... , .... i .... i .... UTS - 213 MPa E, = 104 GPa 200 Load Rate = 0.01 mm/s / 150 I~ 100 50 O r .... i .... ~ .... ~ .... i .... 0.0 0.10 0.20 0.30 0.40 0.50 Strain (%) Fig. 1. Tensile stress/strain for Sylramic S-202. TENSILE STRESS (MPa) I 190i Io 170[ i 1501 130i u~ 11o i 90 L 0 10 20 30 40 50 60 70 80 90 100 110 120 TIME TO FAILURE(HOUR) Fig. 2. Tensile failure under constant load at 1000°C. STRESS(MPa) 200- 150 100 50 0 t~ 1,000E+03 1.000E+04 1.000E+05 1.000E +06 NUIVlBER OF FREQUENCY (Hz) Fig. 3. Low cycle fatigue properties at 1000°C. stress for this material was over 200 MPa at 0.4% strain. The proportional limit was 80 MPa and 0.08% strain. Similar stress-strain curves have been observed at RT and 1200°C. However, ultimate strength values are lower (180 Mpa) at RT and higher (240 MPa) at 1200°C. Time dependent properties such as stress rup￾ture and low cycle fatigue have also been evaluated by Zawada [1] and Linsay [2]. Tensile stress rupture be￾havior is shown in Fig. 2. No stress rupture was found at 100 MPa (110 h). It should be noted that this stress level is well above the proportional limit shown in Fig. 1. A similar result was found for fatigue life at 100 Mpa as shown in Fig. 3 with failure not detected at 105 cycles at 100 MPa. Although shorter life was observed for fatigue condition than for constant stress condition (stress rupture), it is significant that this CMC does survive beyond the proportional limit in oxidizing con￾ditions. 4. Manufacture of complex shapes Data bases can be generated for CMCs from flat laminates or panels. This has occurred within the in￾dustry; however, many applications, if not most, re￾quire that shapes be made from the material system. Sylramic S-202 is a formable system which can be made

H.O. Davis, D.R. Petrak/ Journal of Nuclear Materials 219(1995)26-30 into complex shapes similar to glass and carbon rein formation of the radii for t sections, ribs or rectangu- forced epoxy and /or phenolic composites. Special tool lar shapes. ng, mold dies or mandrels are required, Hard tooling There are two forming mechanisms which must materials with high expansion and /or flexible bags are occur for the radii to achieve the proper contour and needed to make the required shapes for turbine en fiber volume. These mechanisms are"compaction"and gines or other structures. Examples of fabricated stru ply slippage". Both must occur simultaneously and tures made from Sylramic S-202 are shown in Fig 4 over a period of time, usually minutes. The compaction To manufacture these shapes, the preimpregnated of the fabric to approximately 0.013 in. per ply is fabrics must have characteristics which allow these typical. This process may vary depending on resin shapes to be formed. On the other hand, laminates content filler loading and interface coating This mech merely have to be compacted in one direction and anism is the same as that which occurs in a flat lami generally very high pressures, up to 5000 psi, can be nate. The evolution of this mechanism is depicted in used if necessary. Conversely, pressures achievable for Fig. 5. The pressure compacts the yarn bundles, pushes complex shapes without using all hardened tooling are out air and volatiles and squeezes out a required 15 to 300 psi. Therefore, fibers as well as interface amount of resin/filler. Temperature is used to lower coatings must be pliable and formable with a mat the viscosity of the resin and significantly reduces the which reaches a low viscosity at some time during the amount of pressure required to achieve the target fiber cure/pressure process. Naturally, the modulus of a volume. Most of the data discussed above was taken ceramic fiber such as Nicalon prevent radial bends less from panels compacted in this manner than about 3/8 to 1/4 in. In addition, this bending and The second mechanism which must occur before forming must take place without damage to the fabric the first mechanism can occur in a radii, is depicted in or coating Fig. 6. For a rectangle as shown, each ply has to be cut Many shapes have been made by KCC including or applied as shown, i. e by the lines representing plies tubes, leading edges, bolts and nuts. Some of the more of material. The approach is for a female mandrel with complex ones are the combustors and deflectors being he compaction pressure being internally applied. The developed for aircraft engines. One of the most diffi- plies of material must slip from the as laid-up condi cult tasks in making complex structures has been the tion to the required"final"cured position. The amount Fig 4. Typical components manufactured from Sylramic S-202

28 H. O. Davis, D.I~ Petrak /Journal of Nuclear Materials 219 (1995) 26-30 into complex shapes similar to glass and carbon rein￾forced epoxy and/or phenolic composites. Special tool￾ing, mold dies or mandrels are required. Hard tooling, materials with high expansion and/or flexible bags are needed to make the required shapes for turbine en￾gines or other structures. Examples of fabricated struc￾tures made from Sylramic S-202 are shown in Fig. 4. To manufacture these shapes, the preimpregnated fabrics must have characteristics which allow these shapes to be formed. On the other hand, laminates merely have to be compacted in one direction and generally very high pressures, up to 5000 psi, can be used if necessary. Conversely, pressures achievable for complex shapes without using all hardened tooling are 15 to 300 psi. Therefore, fibers as well as interface coatings must be pliable and formable with a matrix which reaches a low viscosity at some time during the cure/pressure process. Naturally, the modulus of a ceramic fiber such as Nicalon prevent radial bends less than about 3/8 to 1/4 in. In addition, this bending and forming must take place without damage to the fabric or coating. Many shapes have been made by KCC including tubes, leading edges, bolts and nuts. Some of the more complex ones are the combustors and deflectors being developed for aircraft engines. One of the most diffi￾cult tasks in making complex structures has been the formation of the radii for T sections, ribs or rectangu￾lar shapes. There are two forming mechanisms which must occur for the radii to achieve the proper contour and fiber volume. These mechanisms are "compaction" and "ply slippage". Both must occur simultaneously and over a period of time, usually minutes. The compaction of the fabric to approximately 0.013 in. per ply is typical. This process may vary depending on resin content filler loading and interface coating. This mech￾anism is the same as that which occurs in a fiat lami￾nate. The evolution of this mechanism is depicted in Fig. 5. The pressure compacts the yarn bundles, pushes out air and volatiles and squeezes out a required amount of resin/filler. Temperature is used to lower the viscosity of the resin and significantly reduces the amount of pressure required to achieve the target fiber volume. Most of the data discussed above was taken from panels compacted in this manner. The second mechanism, which must occur before the first mechanism can occur in a radii, is depicted in Fig. 6. For a rectangle as shown, each ply has to be cut or applied as shown, i.e., by the lines representing plies of material. The approach is for a female mandrel with the compaction pressure being internally applied. The plies of material must slip from the as laid-up condi￾tion to the required "final" cured position. The amount Fig. 4. Typical components manufactured from Sylramic S-202

HO.Davis,DR.Petrak/Journal of Nuclear Materials 219(1995)26-30 NO TEMPERATURE TEMPERATURE TEMPERATURE NO PRESSURE PRESSURE CERAMIC COMPOSITE FABRICATION POLYMER LAY-UP FILLER 1.INCH FABRIC CURE 50 GRAMS OF 30 GRAMS OF RESIN/FILLE DS FILLED WITH AIR WITH IS- DENSIFICATIONPYROLYSIS REPEAT AS NEEDED Fig. 5. Compaction/polymerization for flat laminates with ig. 7. Process flows for CMC fabrication pressure perpendicular to plies 5. Pyrolysis and densification of movement is dependent upon several parameters, After compaction and cure, the remainder of the and it is different for each ply, i.e., the innermost ply processing is the same for flat laminates or complex has to move further than the outermost ply. In me structures. The basic process consists of pyrolysis fol instances, the outermost ply is laid against the mandrel lowed by repeated densification pyrolysis cycles. The the fabric must compact into the radii to preve -ver, process flow diagram with individual steps is shown in and only a change in thickness is required. However, g. 7. The target densities and/or open porosity limits resin rich radius are established based upon the component and its Bladders or solid rubber tools have been used dication.The“char” yield of the fabricate a number of ribstiffened panels and plenum number of cycles are predetermined; however, periodic shapes. bladders have the advantage of better response ensity checks and open porosity measurements are to pressure changes in autoclave processing. However. made bladders can be subject to rupturing during the form The pyrolysis cycles are based upon the volume of ing process. Solid rubber tooling can avoid the possibil- material being decomposed and the temperatures of ity of bladder leaks. In addition, the use of high ther- maximum outgassing. Just as for carbon-carbon pro mal expansion rubber is a good process technique to cessing, there are temperatures where the polymer produce internal pressure against lower thermal expan- decomposes rapidly. Large volumes of gas must exit he structure as well as the furnace at these critical points. In general, these temperatures are determined by TGA measurements on small samples. The mechanisms for densification is rather straight- AS LAID-UP forward, i. e, the resin and solvent fill the open poros- AS COMPACTED ity. As heat is applied the volatile pass through the pores and the remaining polymer decomposes and shrinks with some bonding to the existing pore walls This repeated process builds up a ceramic matrix around the filaments and within the porous structure 0.100-INcH The matrix is generally amorphous and consists of Sicn for the starting matrix of the Sylramic S-202. The increased use of ceramics to replac 0.168-INCH 0.060-INC depends to a great extent on the ability of the Fig.6.Sketch showing fabric movement in a multiple chamber and manufacturers to lower cost. Cost can be signifi cantly lower if large numbers of parts can be processed

H.O. Davis, D.R. Petrak /Journal of Nuclear Materials 219 (1995) 26-30 29 NO TEMPERATURE TEMPERATURE NO PRESSURE PRESSURE L T T 1.5-1NCIt I-INClt 150 GR.~MS OF 150 GRAMS OF RESIN/FILLER RESIN/FILLER AND A:~D LESS VOIDS FILLED VOIDS FILLED WITH ?dR WITH MR TEMPERATURE PRESSURE l/? -f- 0.5-INCH 90 GRA.MS OF RESI N/FILLER .AND NO VOIDS Starting Partially Fully Lay-Up Formed Compacted Fig. 5. Compaction/polymerization for flat laminates with pressure perpendicular to plies. CERAMIC COMPOSITE FABRICATION ~~rpREPREG~ ~_ LA_Y-U P [ [ cuRE ~~ DENSIFICATION~ P~(R~LYslS'_ REPEAT AS NEEDED J Fig. 7. Process flows for CMC fabrication. 5. Pyrolysis and densification of movement is dependent upon several parameters, and it is different for each ply, i.e., the innermost ply has to move further than the outermost ply. In most instances, the outermost ply is laid against the mandrel and only a change in thickness is required. However, the fabric must compact into the radii to prevent a resin rich radius. Bladders or solid rubber tools have been used to fabricate a number of ribstiffened panels and plenum shapes. Bladders have the advantage of better response to pressure changes in autoclave processing. However, bladders can be subject to rupturing during the form￾ing process. Solid rubber tooling can avoid the possibil￾ity of bladder leaks. In addition, the use of high ther￾mal expansion rubber is a good process technique to produce internal pressure against lower thermal expan￾sion metal molds. AS LAID-UP AS COMPACTED __I t 0.168-1NCH 0.280 -INCH ~ i07100-1NCH 0.060-INCH Fig. 6. Sketch showing fabric movement in a multiple chamber component. After compaction and cure, the remainder of the processing is the same for flat laminates or complex structures. The basic process consists of pyrolysis fol￾lowed by repeated densification pyrolysis cycles. The process flow diagram with individual steps is shown in Fig. 7. The target densities and/or open porosity limits are established based upon the component and its application. The "char" yield of the resin mix and the number of cycles are predetermined; however, periodic density checks and open porosity measurements are made. The pyrolysis cycles are based upon the volume of material being decomposed and the temperatures of maximum outgassing. Just as for carbon-carbon pro￾cessing, there are temperatures where the polymer decomposes rapidly. Large volumes of gas must exit the structure as well as the furnace at these critical points. In general, these temperatures are determined by TGA measurements on small samples. The mechanisms for densification is rather straight￾forward, i.e., the resin and solvent fill the open poros￾ity. As heat is applied, the volatile pass through the pores and the remaining polymer decomposes and shrinks with some bonding to the existing pore walls. This repeated process builds up a ceramic matrix around the filaments and within the porous structure. The matrix is generally amorphous and consists of SiCN for the starting matrix of the Sylramic S-202. 6. Applications The increased use of ceramics to replace metal depends to a great extent on the ability of the suppliers and manufacturers to lower cost. Cost can be signifi￾cantly lower if large numbers of parts can be processed

H.O. Davis, D.R. Petrak/ Journal of Nuclear Materials 219(1995)26-30 at the same time. The densification/ pyrolysis process plications. a data base has been generated to show the readily lends itself to batch processing. This can signifi- stability of PIP materials in various environments cantly reduce the cost of tubular structures or similar Toughness is achieved by fiber matrix interface coat components. As the usage of fabric and resin is in- ings. The compaction cure process of the ceramic rein creased, the cost of material should decrease signifi- forced polymer prepreg utilizes many of the thermoset cantly. Combining batch processing with lower cost ra composite processing such as hard tooling(press mold- materials as well as reducing cost through manufacture ing), flexible bag/hard tooling(autoclave hydroclave), innovations provide a viable path for the increased use and expandable tooling materials of hard tooling(auto of PIP CMCs clave, hydroclave and press). A viable process and Complex shapes can be made. Therefore, those tooling approach has been developed for comple applications where CMCs are enabling technology the structures with multiple radii and T stiffeners use will increase. The increased technology and volume from these uses will also lower cost References 7. Conclusions L1 Provided by L. Zawada, Wright Laboratories WPAFB OH4543 The PIP process is a viable technology for produc- [2]G. Linsay, Pratt and Whitney Engine Company, East ing complex shapes for aerospace and commercial ap- Hartford, CT, unpublished creep rupture

30 H. O. Davis, D.I~ Petrak /Journal of Nuclear Materials 219 (1995) 26-30 at the same time. The densification/pyrolysis process readily lends itself to batch processing. This can signifi￾cantly reduce the cost of tubular structures or similar components. As the usage of fabric and resin is in￾creased, the cost of material should decrease signifi￾cantly. Combining batch processing with lower cost raw materials as well as reducing cost through manufacture innovations provide a viable path for the increased use of PIP CMCs. Complex shapes can be made. Therefore, those applications where CMCs are enabling technology, the use will increase. The increased technology and volume from these uses will also lower cost. plications. A data base has been generated to show the stability of PIP materials in various environments. Toughness is achieved by fiber matrix interface coat￾ings. The compaction cure process of the ceramic rein￾forced polymer prepreg utilizes many of the thermoset composite processing such as hard tooling (press mold￾ing), flexible bag/hard tooling (autoclave/hydroclave), and expandable tooling materials of hard tooling (auto￾clave, hydroclave, and press). A viable process and tooling approach has been developed for complex structures with multiple radii and T stiffeners. References 7. Conclusions The PIP process is a viable technology for produc￾ing complex shapes for aerospace and commercial ap- [1] Provided by L. Zawada, Wright Laboratories WPAFB, OH 45433, unpublished data. [2] G. Linsay, Pratt and Whitney Engine Company, East Hartford, CT, unpublished creep rupture