Carbon 41 (2003) 893–901
An RTM densification method of manufacturing carbon–carbon composites using Primaset PT-30 resin Felix Abali, Kunigal Shivakumar*, Nasrollah Hamidi, Robert Sadler Center for Composite Materials Research, Department of Mechanical Engineering, North Carolina A& T State University, Greensboro, NC 27411, USA Received 8 September 2000; received in revised form 12 September 2002; accepted 8 November 2002
Abstract This paper presents a development of carbon–carbon (C–C) composite by resin transfer molding (RTM) process. The RTM was used for both manufacturing of the resin matrix composite part as well as impregnation of the carbonized parts. Materials chosen were heat-treated T300 2-D carbon fabric and Primaset PT-30 cyanate ester. The PT-30 resin has a char yield similar to that of phenolics, very low volatiles, low viscosity at processing temperatures, and no by-products during cure, and hence, an excellent choice for RTM process. The process consists of RTM of the composite part, carbonization, RTM impregnation, and re-carbonization. The last two steps were repeated to achieve the desired density. The measured density and mechanical properties of just two times-densified C–C composite panels were superior to or nearly the same as the data in the literature by other processes. The RTM densification is about twice as fast as the resin solution method and it is environment friendly. 2002 Elsevier Science Ltd. All rights reserved. Keywords: A. Carbon / Carbon composites; B. Pyrolysis; Impregnation; Carbonization; D. Mechanical properties
1. Introduction Carbon–carbon (C–C) composites consist of carbon fibers in a carbon matrix. Unlike metals and ceramics, C–C composites retain high specific modulus, strength, and fracture strain from ambient to very high temperature. High thermal conductivity and low coefficient of thermal expansion (CTE) give the C–C composites an excellent resistance to thermal shock. A high heat of sublimation results in good ablation resistance. High specific strength at high temperatures, excellent fracture toughness, and thermal shock resistance are some of the reasons for using C–C composite in aircraft and race car brakes, heat shields for re-entry vehicles and rocket nozzles. Rocket nozzles, where temperature change from ambient to 3000 8C occurs in a few seconds of missile launch demonstrates the excellent thermal resistance of C–C composites. The other advantages of C–C composites include chemical resistance, excellent high temperature characteristics, bio-com*Corresponding author. Tel.: 11-336-334-7411; fax: 11-336256-0873. E-mail address: [email protected]
patibility, shape stability, and pseudo-plastic fracture behavior. Unlike metals and other homogeneous materials, C–C composites are highly anisotropic and they can be designed to produce highly directional properties. Applications of C–C composites to many industries are discussed in Refs. [1–4]. C–C composites are arguably the most successful material developed since 1960, with products finding many and varied applications [1–4]. Regardless of numerous promising developments, the C–C composite has not realized its full commercial potential because of excessive production cost and poor oxidation resistance. This paper addresses a method of reducing the processing cost of C–C composites through resin transfer molding (RTM) densification. Broadly, there are three known methods of manufacturing C–C composites [4,5]: (1) chemical vapor deposition (CVD), (2) pyrolysis of thermoset and thermoplastic resins, and (3) pyrolysis of hydrocarbon pitches derived from petroleum, coal tar, synthetic mesophase products. In the CVD process, methane (or high carbon content) gas is infiltrated into the carbon fiber preform under controlled cracking conditions such that the pure carbon separates and deposits in the voids of the preform. This deposition
0008-6223 / 02 / $ – see front matter 2002 Elsevier Science Ltd. All rights reserved. doi:10.1016 / S0008-6223(02)00434-7
F. Abali et al. / Carbon 41 (2003) 893–901
eventually fills the voids in the fiber preform with carbon to create a finished product. This method gives a high grade C–C composite. However, the CVD process has a number of limitations. The process has a tendency of not filling the inner plies of the fabric due to outer surface blockage, and therefore it is generally limited to thin structures. The outer surface of the part has to be machined and re-infiltrated several times to achieve nearly uniform density. Furthermore, infiltration of complex fiber architecture such as two dimensional (2-D) braids or weaves with interlocks and / or complex geometrical parts are very difficult and may result in high void content. In addition, the initial set-up cost is expensive and requires highly skilled professionals to operate. In the 2nd method, first a polymeric composite (thermoset or thermoplastic) part is made from a high char yield resin and then it is pyrolyzed in a furnace filled with inert gas to remove volatiles and obtain the carbon matrix. The pyrolysis process is less complicated, and the set-up is less expensive, but requires a long processing time (sometimes months) to allow pyrolyzed gases to escape without damaging the part. The carbonization of thermoplastic composites requires a highpressure chamber to contain the matrix within the carbon fiber preform or the thermoplastic should have been stabilized before carbonization. The third method skips the green part stage and directly pyrolyzes the pitch to get carbon. As in thermoplastics, this method also requires high pressure chamber. Most of the C–C composite work on thermoset resin was done using phenolic resin in the form of prepreg (fiber preform impregnated with resin) because of its excellent char yield (about 65%). Phenolic resin gives out moisture as a condensation by-product during curing and pyrolysis, hence controlled heating rates and evacuation of gases are required to avoid damaging the part. Furthermore, prepreg is not suitable for manufacturing near net shape affordable composites. RTM is an attractive technology for those applications (for example, Ref. ). It is also a high temperature 66–121 8C (150–2508F) and high-pressure injection process and it fully impregnates even the highly complex fiber preform as well as porous C–C parts. The phenolic resin is solid at temperatures required for RTM and releases moisture during cure, hence it cannot be used for RTM. In the rocket nozzle program at NCA&T [6,7], we identified a very high char yield resin system called Primaset PT-30 cyanate ester, manufactured by Lonza Corporation. It is a thermoset resin, with 65% char yield (same as phenolics), less than 0.5% volatiles, no gaseous by-products during cure, a low viscosity at RTM temperatures (80 c.p.s. at 121 8C), and post-curable. Hence it is a potential material for making C–C composites. The molecular structure of PT-30 is shown in Fig. 1. The objectives of this paper are to (1) demonstrate manufacturing of C–C composites from PT-30 resin system, (2) demonstrate RTM densification of carbonized parts, (3) demonstrate the efficiency of RTM densification,
Fig. 1. Molecular structure of PT-30 cyanate ester resin.
and finally (4) determine the mechanical properties of the manufactured C–C composite. The fiber selected was Fiberite’s T-300 PAN-based carbon fabric, woven using 3 K-tow fiber into 8-harness satin architecture. The fabric was heat stabilized at 2650 8C.
2. Development of C–C composite The processing details of C–C composites from PT-30 resin system is explained here. Fig. 2 describes two methods of manufacturing C–C composites from thermoset resin. One is the conventional or solvated resin solution method, Fig. 2a and the other is the present RTM method. In both, the primary steps are (1) fabrication of rigidized prepreg preform (green or as-cured part), (2) carbonization of prepreg preform, and (3) densification. The first two steps are common to both methods. Manufacturing of polymeric composites by various methods is found in many references.. The RTM method is described in Refs. [6–11]. The carbonization, which depends on the resin system in this case PT-30, is explained in the next section. The densification, which is different for the two methods, is further divided into resin impregnation, cure, and carbonization. The resin solution method was for materials like phenolic resin, which is commonly used in industries. In this method, the resin impregnation and cure processes are further divided into four steps: preparation of solvent
Fig. 2. Flow chart of C–C development using solvated and our process.
F. Abali et al. / Carbon 41 (2003) 893–901
solution of resin, impregnation, drying to remove solvents, and cure (see Fig. 2a). Details of this process using PT-30 as a resin system are given in Refs. and. Primary disadvantages of this method are: inefficient resin impregnation, removal of the solvent, potential exposure of workers to hazardous solvents and environmental hazards associated with disposal of resin and solvents. All of these disadvantages are eliminated in the present RTM method. Details of this method are explained later.
2.1. Carbonization process Carbonization is the decomposition of a polymeric matrix into carbon and volatile molecules by pyrolyzing the polymeric matrix in an inert atmosphere. The volatiles escape leaving the carbon matrix in the carbon fiber preform. The amount of carbon matrix depends on char yield of the resin and the processing conditions. The PT-30 has about 65% char yield by weight confirmed independently by the controlled thermo-gravimetric analysis (TGA). Fig. 3 shows the TGA data of PT-30 resin. Starting from neat resin, it was identically cured the way as in the RTM process, post-cured, and pyrolyzed at a slow heat rate (0.1 8C / min). All weights were normalized by the post-cured weight. The data indicate that the resin loses 5% during cure, 2% during post-cure, and 35% during the carbonization (pyrolysis). This study re-confirmed the data provided by the manufacturer and tested by ARC  at 10–20 8C / min heating rate. The remaining 35% weight of matrix becomes voids that have to be filled by repeated densification to the required density or acceptable void content. The carbonization process consists of (1) establishment of the carbonization cycle, (2) set-up of the carbonization furnace, and (3) filtration and analysis of pyrolysis gases.
2.1.1. Development of carbonization cycle The carbonization consists of vaporization of volatiles in the polymeric matrix by heat. The gases released must diffuse through the composite material without damaging
Fig. 3. Complete cycle of TGA of PT-30 resin to carbonization.
the part. It is ideal if the carbonization could be initiated at the outer surface of the part and slowly progressed to the inside. This would create natural shrinkage cracks and provide a passage for the escaping gases. The heating rate determines the rate of release of these gases. Higher heating rates increase the volatile release rate and could crack and / or delaminate the composite part. Hence it is necessary to determine safe heating rates that will not damage the part. This was established from TGA (see Fig. 3) and DSC data. For PT-30, pyrolysis starts at 400 8C, the volatile release rate peaks at about 425 8C, and starts decreasing from 500 8C, and reaching nearly zero at 650 8C. Preliminary heating cycle was selected based on this data which had heating rate of 6 8C / min from room temperature to 350 8C, 3 8C / min to 400 8C, 1 8C / min to 500 8C, 2 8C / min to 600 8C, and 3 8C / min to 950 8C. The part was cooled to room temperature by turning off the furnace (see Fig. 4). This heat cycle was too severe and resulted in completely delaminated panels. Subsequently the heating rates were modified by trial studies and the safe heating rate was established. This had an average of 0.1 8C / min for the temperature range of 350 8C to 650 8C. Trial studies were conducted, which resulted in a number of interesting results. The step-hold heating helped in pushing Tg above the pyrolysis temperature. Increasing the hold time at 600 8C yielded more char. Based on this study, a modified and final carbonization cycle was established and it is shown in Fig. 4. In this cycle, the furnace was ramped from room temperature to 250 8C at a high heating rate of 5 8C / min and held for 8 h. The 8-h hold represents the post-curing of PT-30, which is necessary to achieve high char yield . The furnace temperature was raised to 350 8C and held for 4 h. Starting from 350 8C, the temperature was incremented by 25 8C and held for 4 h at each level until 650 8C was reached. At 600 8C, the hold time was increased to 8 h. From 650 to 950 8C, the furnace was ramped up at 5 8C / min and held at 950 8C for 24 h. Finally, the C–C composite part was cooled to room temperature at 5 8C / min. The composite part chosen for this study was a rectangular panel of size 305330532.6 mm. The panel temperature was measured by a thermocouple and is shown by the dashed line in the
Fig. 4. Preliminary and final carbonization cycles.
F. Abali et al. / Carbon 41 (2003) 893–901
figure. The difference in furnace and panel temperatures was due to the location of thermocouples in the furnace and the sample. The total heating cycle was about 4 full days and cooling was about one-half day. All panels made from this heating cycle were found to be of good quality. Additional studies are being conducted to optimize the heating rates.
2.1.2. Carbonization furnace set-up The set-up used for the carbonization is shown in Fig. 5. It consists of a furnace capable of temperatures up to 1200 8C. A continuous flow of nitrogen gas into the furnace was maintained to create a non-oxidizing environment. A positive nitrogen pressure (of about 51 mm of water) was maintained inside the furnace to avoid any air from getting into the furnace during the carbonization process. Composite panels were embedded in coke granules to further avoid exposure to any trace of oxygen during the carbonization cycle. The gaseous by-products of the pyrolysis were filtered through a water bath and an activated carbon filtering system before exhausting to the atmosphere. 2.1.3. Analysis PT-30 pyrolysis gas Primaset PT-30 was a new resin and no pyrolysis data was available in literature. Therefore, constituent gases were measured by GC mass spectroscopy by collecting pyrolysis gas samples from 350 to 950 8C at increments of
50 8C. The predominant gases before filtration were acetonitrile (,0.3 ppm), trichloroethane (,0.5 ppm) and toluene (,1.1 ppm). Pyrolysis gases after passing through the filtration system were benonitrile (,1 ppm) and phenol (,22 ppm) and all other gases were less than 0.1 p.p.m. These gases were found to be environmentally safe according to the local OSHA office, and were released to the atmosphere.
3. RTM densification process Densification consists of impregnation of a porous C–C composite part with high char yield resin followed by re-carbonization. Several alternative polymer / solvent solutions and impregnation procedures could be used to fill shrinkage cracks in the carbonized composite material. The end product’s density and the mechanical properties depend on the number of densification cycles, which in turn depends on the efficiency of resin impregnation and its carbon yield. As mentioned previously in the resin solution method, PT-30 was dissolved in MEK with 60:40 ratio as described in Refs. [8–10]. Details of RTM impregnation are presented here (see Fig. 6). The carbonized panel was placed in the mold (see Fig. 6). A Graco pump was loaded with degassed Primaset PT-30 cyanate ester resin and preheated to 79 8C. The pump, mold and vacuum pump were connected together
Fig. 5. Carbonization furnace with filtering arrangement for exhaust gases.
F. Abali et al. / Carbon 41 (2003) 893–901
Fig. 6. RTM set-up for part fabrication and re-impregnation.
with vacuum-tight copper plumbing and preheated with flexible strip heaters. The inlet side (between the pump and the mold) of the plumbing was heated to 79 8C (175 8F). A vacuum of 690 mm (27 inches) of mercury was maintained during the resin injection phase. The outlet (between the mold and the resin trap) was heated to 93 8C (220 8F). The press platen were heated to 177 8C (350 8F) and pressed to close the mold (about 2.1 MPa clamping pressure). The mold and plumbing were checked periodically to make sure all connections and seals were air-tight. For RTM impregnation to be successful, a good vacuum is necessary. Once the mold and the carbonized part are evacuated the vacuum line was shut-off and hot resin was pumped at 15 ml / min at a pressure of 1.24 MPa (180 p.s.i.). The plumbing line was burped several times to ascertain the air pockets in the mold are removed. After the resin flowed freely without visible air bubbles or splatter, the burp valve was turned off and the resin pressure was maintained at 1.24 MPa until the filling was completed. Then the pump pressure was increased to 2.1 MPa. The evacuation process takes about 15 min. The filling process takes about 45 min. Once the filling was complete, the panel was cured at 177 8C (350 8F) and 2.1 MPa pressure for 5 h. The impregnated panels were re-carbonized, as explained in the Carbonization section. Fig. 7 shows scanning electron micrographs of the cross-section of panels at various stages of carbonization. Fig. 7a shows the as-cured composite part, notice the dense dark color of PT-30 resin. Fig. 7b is after the 1st carbonization. Many shrinkage cracks due to pyrolysis of the resin are created. Complete fill-up of cracks by PT-30 is shown in Fig. 7c. This demonstrates the efficiency of RTM impregnation in filling up of cracks throughout the body of the material. Finally, healing of shrinkage cracks are demonstrated in the fully densified part in Fig. 7d. A US patent  has been issued in 2001. The RTM densification process is beneficial to both designer and the manufacturer. To a designer, it provides greater freedom of geometric configuration, fiber architecture, and lower cost in processing and machining. Benefits
to a manufacturer include the use of established RTM process along with 100% resin for deep impregnation; reduced equipment cost because the same mold is used for fabrication as well as impregnation of the carbonized part; absence of solvents in the process eliminates employees’ exposure to hazardous fumes; and finally, fewer densification steps (one-half) to achieve the same density as the resin solution impregnation method.
4. Evaluation of C–C composites
4.1. Development plan and test methods Polymeric composite panel of size 305330532.6 mm was manufactured by the RTM process. The panel was divided into four quarters (see Fig. 8). The first quarter was used for determining the mechanical properties of the as-cured polymeric laminate. The remaining three-quarters were carbonized. The second quarter was carbonized and the mechanical properties determined. The third quarter was carbonized and densified (impregnation1 carbonization) and then the mechanical properties determined. The fourth quarter was carbonized and densified twice, and then the mechanical properties determined. The properties measured were bulk density, tensile modulus and strength, Poisson’s ratio, compression strength, inplane shear modulus and strength, interlaminar shear strength, and interlaminar tensile strength. The density, tension and interlaminar shear properties were measured for each carbonization cycle whereas the remaining properties were measured at the end of all densification cycles. Each of the quarters was divided into specimens for tension (type A), interlaminar shear (type B) and interlaminar tension (type C) tests as shown in Fig. 9.
4.2. Bulk density The bulk density was measured from the overall volume
F. Abali et al. / Carbon 41 (2003) 893–901
Fig. 7. Micrographs of panel cross-sections at various stages of carbonization and impregnation.
Fig. 9. Specimen layout in the first quarter of the panel.
Fig. 8. Carbonization plan for a panel of 30330 cm.
and mass of the quarter panel. After the panel was made and the edges trimmed, length and width were measured at three points (two ends and one mid-section) and the
F. Abali et al. / Carbon 41 (2003) 893–901
Table 1 Properties of T300 carbon–carbon composites Properties
T300 / PT-30 composite
C–C composite Quarter panel number 2
Density, g / ml
In-plane tensile test Modulus, GPa (S.D.)
Strength, MPa (S.D.)
Fracture strain, % (S.D.)
In-plane shear test Strength, MPa (S.D.) Modulus, GPa (S.D.) Compression strength, MPa (S.D.) Interlaminar shear strength , MPa (S.D.) Interlaminar tension strength, MPa (S.D.) a
34.70 (0.27) 3.24 (0.06)
SOL RTM a SOL RTM SOL RTM
3 1.42 1.41
80.10 (4.8) – 163.50 (14.5) – 0.43 (0.12) –
4 1.53 1.56
126.90 (8.3) 104.80 (4.8) 233.40 (6.2) 251.20 (7.6) 0.20 (0.05) 0.40 (0.10)
15.7 (4.3) – –
127.00 (11.0) 118.30 (6.9) 308.10 (4.8) 308.10 (1.4) 0.28 (0.03) 0.25 (0.03)
27.70 (2.1) 5.78 (0.32) 158.60 (8.2) 20.60 (3.3) 3.40 (1.04) 2.00 (0.17)
10% PT-60 tackifier was used.
thickness was measured at eight points (four corners and four mid-points along the edges). The volume was calculated by multiplying the average length, width, and thickness. The density of the panel was obtained by dividing the mass by volume. The computed density is the bulk density, the true density will be larger than the bulk density. Densities were measured at the end of each carbonization cycle. Table 1 summarizes the results. The Table includes results of solvated resin (represented by SOL) as well as by RTM impregnation methods. The C–C composite density increased with the number of densifications. The solvent impregnation process took more (three or more) cycles to achieve the same density as the RTM impregnation. The maximum density achieved by resin solution impregnation by three densifications was about 1.54 g / ml whereas RTM impregnation yielded 1.61 g / ml in two densifications. The theoretical limit for T300 fabric was 1.7 g / ml, but the practical values reported by the industry were between 1.60 and 1.65 g / ml (as reported by the Atlantic Research Corporation).
4.3. Mechanical properties The mechanical properties measured were in-plane tension, compression and shear, and interlaminar tension and shear. The specimen configurations are shown in Fig. 10. The specimen locations in the first quarter-panel are shown in Fig. 9. In all other quarters, specimens were
located symmetrically to the first quarter. Separate panels were made for both compression and in-plane shear test specimens. The specimen configurations of in-plane tension, compression, and shear and interlaminar tension were as given in ASTM D-30 handbook . The interlaminar shear strength (ILSS) was determined using modified short beam shear test . All tests were conducted by displacement controlled loading in an MTS test machine. The loading rate for all tests was 0.5 mm / min except for transverse tension which had 0.05 mm / min. The specimens were appropriately strain-gagged and the load, displacements, and strains were recorded continuously at 1 s intervals. Loads at failure initiation and at fracture were recorded. Strengths were calculated from the ultimate load and modulus from the linear portion of the stress–strain curve. At least four of the specimens were tested for each case. The measured average values and the standard deviation (S.D.) for the resin solution (SOL) and RTM methods are listed in Table 1. The Table includes the data for resin solution and RTM impregnation methods. As mentioned previously, resin solution impregnated panels had at least three densifications versus two for the RTM process. Both tension modulus and strength increased with number of densifications. Furthermore, data scatter generally decreased with number densifications. Properties of the SOL and RTM impregnated panels had nearly the same values. The inplane tension modulus, strength and fracture strain of
F. Abali et al. / Carbon 41 (2003) 893–901
Fig. 10. Test specimen configurations.
densified panels were nearly the same as that of the as-cured polymer composite panel. The C–C composite modulus was about 118 GPa, strength of 308 MPa, and fracture strain of 0.25%. The compression strength of C–C composite (158.6 MPa) is about three-quarter of the as-cured polymeric composite (199.8 MPa). The in-plane shear modulus and strength of the as-cured polymeric composites were 3.24 GPa and 34.7 MPa respectively, while those of C–C composite were 5.78 GPa and 27.7 MPa. Both interlaminar
tension strength (ILTS) and shear strength increased with number of densifications. Almost a four-fold increase in ILSS was achieved by densifications. After two densifications ILSS was 20.6 MPa and ILTS was 2.0 MPa by the RTM method and 3.4 MPa by resin solution method. The ILTS test was very sensitive to misalignment of the loading, which is reflected by large data scatter. Table 2 compares the properties of C–C composites manufactured by the RTM densification method (2-densification cycle) with CVD process  and Bimatrix C–C
Table 2 Comparison of carbon–carbon composite properties Mechanical properties
ARC a T300 and ACC-2 bimatrix
In-plane tensile test Modulus, GPa Strength, MPa Fracture strain, %
118.30 308.10 0.25
103.4 275.8 –
37.90 314.40 0.83
In-plane shear test Modulus, GPa Strength, MPa 0.2% strength, MPa Compression strength, MPa Interlaminar shear strength, MPa Interlaminar tension strength, MPa
5.78 27.70 24.80 158.60 20.60 c 2.00
17.2 34.7 – 165.5 10.3 6.2
– 54.70 b – 111.70 8.01 4.10 b
ARC—Atlantic Research Corporation. SORI—Southwest Research Institute (ACC-4). c Measured from the MSBS test . b
F. Abali et al. / Carbon 41 (2003) 893–901
composite from Atlantic Research Corporation (ARC). The Bimatrix method involved pyrolysis of a thermoset composite made of heat-treated T300 carbon fiber followed by CVD densification. Most of the properties of RTM method are better than the properties reported in the two references. The in-plane shear properties of the present method are not as good as in Ref. . Our ILTS had alignment problems, hence the measured strength is low due to out-of-plane bending of the specimen.
5. Concluding remarks This paper presented a development of carbon–carbon composite by resin transfer molding using Primaset PT-30 cyanate ester and T300 heat-treated carbon fabric. The PT-30 has a char yield about 65% which is the same as that of phenolics (a commonly used resin for C–C composites), very low volatiles, low viscosity, and no gaseous by-products during cure. Thus PT-30 is an excellent choice for resin transfer molding and C–C composite manufacturing. Starting with a RTM PT-30 / carbon fiber composite part a complete carbonization and densification cycles were established. The duration of each carbonization cycle was about 4.5 days and can be improved through optimization. The RTM densification has about twice the efficiency of the resin solution method. The RTM densification has other advantages including the absence of drying, disposal of solvents, and the part can be cured and impregnated in the same mold under pressure. The measured density and mechanical properties of just two times densified C–C composite materials were comparable to the data in the literature by CVD processes. The measured in-plane tension properties of C–C and polymer composites were nearly the same, because the property is primarily controlled by the fiber properties. The RTM densified C–C composite’s tensile modulus was 118.3 GPa, strength was 308.1 MPa, and fracture strain was 0.25%; the compression strength was 158.6 MPa, in-plane shear modulus and the shear strength were 5.78 GPa and 27.7 MPa, respectively; the interlaminar tensile and shear strengths were 2.0 MPa and 20.6 MPa, respectively.
Acknowledgements The authors acknowledge the financial support of the Air Force Research Laboratory’s MLBC Directorate through the contract No. F33615-96-C-5057.
References  Schmidt DL, Davidson KE, Theibert LS. Unique applications of carbon–carbon composite materials. SAMPE J 1999;35(3):27–39.
 Schmidt DL, Davidson KE, Theibert LS. Unique applications of carbon–carbon composite materials. SAMPE J 1999;35(4):51–63.  Schmidt DL, Davidson KE, Theibert LS. Unique applications of carbon–carbon composite materials. SAMPE J 1999;35(5):47–55.  Savage G. In: Carbon / Carbon composites, New York: Oxford University Press, 1993, pp. 117–56.  Rand B. Matrix precursors for carbon–carbon composites. In: Thomas CR, editor, Essentials of carbon–carbon composites, London: Royal Society of Chemistry, 1993, pp. 67–102.  Sadler RL, Morgan BD, Avva VS. Resin transfer molding of 3-D braided through the thickness Pan-based carbon fiber preforms with a cyanate ester. In: 1995 JANNAF Rocket Nozzle Technology Subcommittee Meeting. Tampa, Florida, USA, Chemical Propulsion Information Agency, 1995, pp. 93–101.  Shivakumar KN, Cozart AB, Krishnamurthy R, Avva VS, Spear G, Hennessey KP. Aero, thermal and structural of 3-D braided composite nozzle. In: 1995 JANNAF Rocket Nozzle Technology Subcommittee Meeting. Tampa, Florida, USA, Chemical Propulsion Information Agency, 1995, pp. 261–74.  Sunderesan MJ, Abali FL, Haq S, Avva VS, McPherson DJ, Shivakumar KN. Development of carbon / carbon composites using PT-30 precursor matrix. In: 39th AIAA /ASME / ASCE /AHS /ASC SDM Conference, Atlanta, GA, USA, 1998, AIAA Paper No. 98-1901.  Shivakumar KN, Sadler RL, McPherson DJ. Air Force Research Laboratory Final Report. Development of resin transfer molding technology and material characterization applicable to composite inlets, 2001, Contract No. F3361596-C-5057.  Shivakumar K, Abali F, Sadler R, McCoy J. Development of cyanate ester based carbon / carbon composites. In: Proceedings of ICCM-12, Paris, France, 1999.  Shivakumar KN, Abali FL, Suderesan MJ, Avva VS, Sadler RL, Cunningham A. Reimpregnation of carbon–carbon composites by the resin transfer molding process, 2001, US patent Number 6,325,608.  Annual book of ASTM Standards. Volume 15.03 space simulation; aerospace and aircraft; high modulus fibers and composites. Test standard D-30, 1999.  Abali F, Pora A, Shivakumar K, Ghante S. Modified short beam shear test for the measurement of ILSS of carbon / carbon composites. In: 41st AIAA /ASME /ASCE /AHS /ASC SDM Conference, Atlanta, GA, USA, 2000, AIAA Paper 2000-1480.  Hsu SR, Chen C. Processing and properties of some C–C systems. In: Superalloys, supercomposites and superceramics, New York: Academic Press, 1989, pp. 721–44.