Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Introduction

CMCs combine reinforcing phases with a ceramic matrix to have been developed to overcome the intrinsic brittleness and lack of reliability of monolithic ceramics and create new and superior properties. In CMCs, in another word, the primary goal of the reinforcement is to provide toughness to an otherwise brittle ceramic matrix. Reinforcements can also be added to the ceramic matrix during processing to enhance or tailor characteristics such as electrical conductivity, thermal conductivity, thermal expansion, and hardness. The desirable characteristics of thermally conductive CMCs include high thermal conductivity, high-temperature stability, high thermal shock resistance, high hardness, high corrosion resistance, light weight, nonmagnetic and nonconductive properties, and versatility in providing unique engineering solutions. The combination of these characteristics makes CMCs attractive alternatives to thermal management of electronic packaging, particularly for high-temperature electronic packaging system. Ceramic matrices can be categorized as either oxides or nonoxides and in some cases may contain residual metal after processing. Some of the more common oxide matrices include alumina, silica, mullite, barium aluminosilicate, lithium aluminosilicate, calcium aluminosilicate, and beryllium oxide. Of these, alumina and mullite have been the most widely used because of their in-service thermal and chemical stability and their compatibility with common reinforcements. Although oxide matrices are often considered more mature and environmentally stable, nonoxide ceramics are rapidly entering the marketplace with superior structural properties, hardness, and, in some environments, corrosion resistance. Some of the more common nonoxide ceramics include SiC, Si3N4, boron carbide, and AlN. Of these, SiC has been the most widely used, with AlN of increasing interest where high thermal conductivity is required and Si3N4, where high strength is desired. The reinforcements are available in a variety of forms. Early ceramic composites used discontinuous reinforcements which, when combined with ceramic matrices, could be formed using conventional monolithic ceramic processes (Richerson 1997).

Discontinuous oxide ceramic reinforcements are less prevalent because of their incompatibility with many common ceramic matrices. Advanced composites, which use the fiber in a continuous form to better optimize the structural properties, generally require more costly manufacturing processes. Continuous fiber is available in both monofilament and multifilament tow forms, with multifilament tow fiber being lower in cost on a per-pound basis and, in many cases, easier to process into complex shapes. Some of the more common continuous reinforcements include glass, mullite, alumina, carbon, and SiC. Of these, SiC fibers have been the most widely used because of their high strength, stiffness, and thermal stability. SiC matrix continuous fiber ceramic composites (CFCCs) have been successfully demonstrated in a number of applications where a combination of high thermal conductivity, low thermal expansion, light weight, and good corrosion and wear resistance is desired. SiC matrix CFCCs can be fabricated using a variety of processes, fibers, and interface coatings. Fibers widely used for industrial applications where long life is desired include SiC or mullite. Processes available to fabricate SiC matrix CFCCs and the matrix composition formed include chemical vapor infiltration (CVI) (SiC), polymer infiltration (SiCN, SiC), nitride bonding (Si–SiC–Si3N4), and melt infiltration (Si–SiC). The interface coating can be either carbon or boron nitride with a protective overcoat of SiC or Si3N4 (Richerson 1997).

For applications where temperature is lower (<1,100°C) or exposure times limited, mullite fibers have been the most widely used because of their lower cost. Continuous ceramic fibers are of increasing interest because of their ability to provide pseudoductile characteristics to otherwise brittle ceramic materials. Continuous ceramic fibers become especially important in large structures where the probability of processing- or in-service-induced flaws increases with part size and could result in catastrophic failure (Richerson 1997). As an example of applications, the use of CMCs in advanced engines will allow an increase of the temperature at which the engine can be operated and eventually the elimination of the cooling fluids, both resulting in an increase of yield. Further, the use of light CMCs in place of heavy superalloys is expected to yield significant weight saving. Although CMCs are promising thermostructural materials, their applications are still limited by the lack of suitable reinforcements, processing difficulties, sound material data bases, lifetime and cost.

With advent of carbon nanotubes (CNTs) or other naomaterials, the nanoscale reinforcements have been utilized in composites in order to improve their mechanical, thermal, and electrical properties. The combination of the nanotubes with a ceramic matrix could potentially create composites that have high-temperature stability as well as exceptional toughness and creep resistance. Increased thermal and electrical conductivity of the composites even at low volume fractions also might provide the thermal transport needed to reduce material operating temperatures and improve thermal shock resistance in applications like thermal elements and electrical igniters. The high temperature and high reactive environment during conventional fabrication methods of ceramics damage the CNTs and thus, there is a need of alternate methods of processing the composites. There is also a need for control of interface between nanotube and the matrix for better interfacial bonding. However, the research is still in the embryonic stage and there are many challenges to be resolved before it is ready for use in varied industrial applications including thermal management of electronic packaging (Samal and Bal 2008).

State of the Art in Processing of SiC Matrix Ceramic Matrix Composites

Depending on the discontinuous or continuous reinforcements, SiC matrix ceramic composites can be fabricated by a variety of processes. Discontinuous reinforced ceramic composites are produced using processes originally developed for monolithic ceramics. Processing methods commonly used include slip casting or injection molding followed by sintering to full density in a high-temperature-capable furnace. The shaping and sintering processes can also be combined using unidirectional hot pressing or hot isotatic pressing. Net or near-net shape processing can be achieved with final machining often limited to satisfying high-tolerance dimensions or surface finishes (Richerson 1997).

Ceramic composites containing continuous fiber reinforcements must be processed by methods that accommodate the continuous nature of the reinforcement. Typically, processing involves the formation of a fiber preform that contains an interface coating applied by chemical vapor deposition (CVD) or a particle-filled slurry process followed by impregnation with a second particle-filled slurry mix, preceramic polymer, precursor gases, molten metal, or other raw material that converts to a ceramic matrix when heated. The interface is a very thin layer (<5 μm total thickness), can be multiple layers to achieve the desired result, and is applied to the individual ceramic filaments. The interface serves as protection for the fibers during matrix processing and as a source of debonding during crack propagation in the brittle ceramic matrix. Some typical SiC ceramic composite microstructures are shown in Figure 7.1. Figure 7.1a illustrates the SiC–Si matrix, SiC fiber, and BN–SiC interfaces. Figure 7.1b shows a highly tailored composite structure: multilayered self-healing interphases and matrices that combine crack arrester layers and glass former layers. Figure 7.1c and d show the microstructure and fracture surface of carbon fiber reinforced SiC composite, separately.

Fig. 7.1
figure 1

Typical microstructure of silicon carbide (SiC) matrix composites. (a) SiC +Si matrix, SiC fiber and BN–SiC interface phases; (b) Multilayered self-healing matrix ceramic matrix composite (CMC), including crack arrester layers and glass-forming layers; (c) Carbon fiber reinforced SiC composite; and (d) Fracture surface of (c)

Full size image

With the CVD process, a ceramic interface and the SiC matrix are both formed by deposition from a gas precursor at elevated temperature. Multiple cycles of the matrix-forming process are required to achieve the desired final density, with each cycle taking up to a week in many cases. The number of cycles varies from two to five, depending on the part geometry and size (Richerson 1997).

The infiltration process for CMC fabrication can be clarified gas-phase routes or liquid-phase routes, each of them having advantages and drawbacks. In gas-phase routes, i.e., the so-called CVI processes, the reinforcements (usually as a multidirectional preform) are densified by the matrix deposited from a gaseous precursor, e.g., a hydrocarbon for carbon or a mixture of methyltrichlorosilane and hydrogen for silicon carbide. It is now well established that a fiber coating, i.e., the interphase, has to be deposited on the fiber prior to the infiltration of the matrix in order to control the fiber-matrix bonding and the mechanical behavior of the composite. Pyrocarbon (PyC), boron nitride or (PyC–SiC)n and (BN–PyC)n multilayers, with an overall thickness ranging from about 0.1 μm to about 1 μm, displaying a layered crystal structure (PyC, BN) or a layered microstructure (multilayers), are the most common interphase materials in nonoxide CMCs. The main role of the interphase is to deflect the microcracks which form in the matrix under loading and hence to protect the fiber form notch effect (mechanical fuse function). There are several versions of the CVI process. The most commonly studied and used version is isothermal/isobaric CVI (or I-CVI). It is a relatively slow process because mass transfer in the preform is mainly by diffusion and it yields some residual porosity and density gradient. Conversely, I-CVI is a clean and flexible process (it can be used to densify simultaneously a large number of preforms, eventually of different shapes). For these reasons, it is the preferred process at the plant level. It is well suited to the fabrication of relatively thin parts. In order to increase the densification rate and hence to reduce the processing times, temperature or/and pressure gradients can be applied to the preform. In temperature gradient CVI, or forced CVI, the processing time can be reduced by one order of magnitude with respect to I-CVI. A similar processing time lowering has also been reported for the film-boiling (or calefaction) process, in which the heated fiber preform is directly immersed in a liquid matrix precursor. Finally, pressure pulsed-CVI has been presented at the micrometer (or even nanometre) scale, either the interphase or the matrix. Based on this technique, multilayered self-healing interphases and matrices (combining crack arrester layers and glass former layers) have been designed and produced, through a proper selection of chemical composition of the layers (Fig. 7.1b) (Naslain 2010).

The methods of fabrication through liquid phase routes include (1) infiltration of molten ceramic; (2) slurry infiltration process (SIP); (3) reactive melt infiltration (RMI); and (4) polymer infiltration and pyrolysis (PIP). Infiltration of molten ceramic into a fiber preform is limited by low viscosity of molten ceramics and by high temperature causing chemical interaction between the molten matrix and the dispersed phase (fibers). This process is similar to liquid state fabrication of metal matrix composites (MMCs), and sometimes used for fabrication glass matrix composites. SIP involves the following operations: (1) passing fibers (tow, tape) through a slurry containing particles of the ceramic matrix; (2) winding the fibers infiltrated by the slurry onto a drum and drying; (3) stacking of the slurry impregnated fibers in a desired shape; and (4) consolidation of the matrix by hot pressing in graphite die at high temperature. The RMI process is used primarily for fabrication of SiC matrix composites. The process involves infiltration of carbon (C) containing preform with molten silicon (Si). Infiltration is usually capillary forced. Carbon of the impregnated preform reacts with liquid silicon, forming SiC. Resulting matrix consists of SiC and some residual silicon. When liquid aluminum (Al) is used for infiltration of a preform in oxidizing atmosphere, alumina–aluminum (Al2O3–Al) matrix is formed. The RMI method is fast and relatively cost effective. Materials fabricated by the RMI method possess low porosity and high thermal conductivity and electrical conductivity (Naslain 2010).

PIP involves the following operations: (1) fiber preform (or powder compact) is soaked with a soft (heated) polymer, forming polymeric precursor. (2) The polymer is cured (cross-linked) at 480°F (250°C). (3) The polymer precursor is then pyrolyzed at 1,100–1,830°F (600–1,000°C). As a result of pyrolysis the polymer converts to ceramic. Pyrolysis causes shrinkage of the matrix material and formation of pores (up to 40 vol.%). (4) The pyrolyzed polymeric cursor may be hot pressed for densification. Hot pressing is limited by possible damage of fibers. (5) Infiltration pyrolysis cycle is repeated several times until the desired density is achieved. Matrices consisting of carbon, silicon carbide, silicon oxycarbide, silicon nitride and silicon oxynitride may be fabricated by PIP method. The following materials are used as polymers in the PIP method include thermosets (thermosetting resins); Pitches or other carbon-containing liquids for fabrication of carbon matrix; polycarbosilane, polysilastyrol, dodecamethylcyclohexasilan for fabrication of SiC matrix. PIP methods are simple low temperature methods, which allow production of intrinsic parts.

The infiltration of fiber preform with preceramic polymer or particle slurry is illustrated in Figure 7.2. The interface material is applied onto the fiber, cloth, or perform by CVI or particle-filled slurry processes before coating with a matrix-forming second slurry or preceramic polymer. Once coated with a polymer or slurry, the shape is formed and heat treated at low temperatures to rigidize the shape. The rigidized shape is then heat treated at elevated temperatures to convert the polymer to a ceramic matrix or bond particles in the ceramic-filled slurry. For preceramic polymers, multiple cycles are required to achieve the desired final density, with each cycle taking 1–3 days. The number of cycles used varies from 4 to 15 depending on the polymer and the desired final density. When using particle-filled slurries to form the matrix, porosity remaining after rigidizing the preform is either filled with molten silicon metal to form a matrix of Si–SiC or Si contained in the slurry reacts during high-temperature heat treatment in nitrogen to form a matrix of Si–SiC–Si3N4. Variants of the melt infiltration process can also be used in which the molten silicon reacts with carbon present in the matrix to form additional SiC. Shaping to near-net dimensions occurs during preforming with final machining limited to high tolerance attachments to avoid surface fiber damage. Additional machining may be required for complex shapes and is generally performed early in the process when the composite is less dense and easier to machine. External coatings can be used to smooth the contact surface for assembly or enhance environmental resistance with some of the more common environmental barrier coatings including mullite, SiC, and Si3N4 (Richerson 1997).

Fig. 7.2
figure 2

Flowchart of liquid phase route processing for SiC based CMC

Full size image

SiC–Diamond Composites

The inherent brittleness of single-crystal diamond limits its application in harsh environments of dynamic impacts and high-stress concentrations. Therefore, more flaw-tolerant diamond composites have been developed through the design of their microstructures. In order to manufacture bulk materials containing diamond powder, which starts with single-crystal diamond or diamond powder converted directly from graphite, the composites must be fabricated by using a suitable bonding agent under appropriate conditions. It is difficult to directly sinter diamond because some diamond may transform into graphite because of the inhomogeneous distribution of pressure inside the specimen. To avoid this problem, two classes of diamond composites were developed, polycrystalline diamond (PCD) and SiC–diamond composites. For PCD composites, cobalt usually serves as the bonding agent. Under high temperature and high pressure, cobalt melts and infiltrates the voids among the diamond particles. Then cobalt maintains the hydrostatic condition inside the sample and thus prevents the diamond graphitization during the sintering process. The next step is to leach out cobalt with a solvent, then the PCD composites mainly consist of diamond crystallites growing together around their contacting area, and the remaining strong carbon–carbon (C–C) bonds retain about 90% modulus of elasticity of diamond. However, even a very small amount of cobalt remaining in the composites after leaching, promotes graphitization of diamond when PCD is exposed to temperature above 1,000 K. Graphite impurities reduce the strength and degrade the PCD compacts (Wang 2006).

SiC–diamond composites can be used to overcome the problem of diamond graphitization by replacing cobalt with SiC as the binding agent. There exist techniques to manufacture SiC–diamond composites, like CVD, vacuum sintering, and high-pressure and high-temperature (HPHT) sintering. Under HPHT, Si melts and fills out cavities among diamond particles, and immediately it starts reacting with diamond. The as-grown SiC serves as a matrix phase to bind all diamond particles together. At the correct ratio of diamond to Si and optimized reaction conditions, all Si will transform into SiC. Strong Si–C bonds and well bonding between SiC and diamond make it is possible for SiC–diamond composites to provide excellent thermal conductivity and mechanical properties. Besides the grain size of the precursors, the overall thermal conductivity and mechanical properties of SiC–diamond composites depend on the concentration of SiC, its structure, density of dislocation, and grain boundaries (Wang 2006).

However, conventional HPHT infiltration sintering is not suitable when the starting materials consist of submicron or nano-size powders due to the so-called self-stop effect. During the sintering process, the newly formed SiC may decrease the open space of tiny pores or block the entrance to a void. This phenomenon is apt to limit or even completely block the infiltration of liquid Si and hence hinder the entire sintering process. To overcome this problem, an alternative method has been developed, which is based on the simultaneous ball milling of diamond and silicon powder mixtures prior to HPHT sintering. After ball milling Si became amorphous and uniformly coated on diamonds. The synthesized composites had a unique nanostructure of diamond embedded into a matrix of nanocrystalline SiC, with large improvement in fracture toughness. These composites showed as much as 50% enhanced fracture toughness while maintaining superhard and superabrasive properties when SiC matrix size decreased from 10 μm to 20 nm. It is in contradiction with the regularly held belief of an inverse correlation between hardness and fracture toughness of materials. Further reduction of SiC grain sizes by the ball milling technique seems very unlikely. Another method of producing nanostructure SiC matrix is to use CNT in SiC–diamond composites manufacturing. This novel composite will consist of diamond imbedded into SiC/CNTs matrix obtained by sintering/infiltration technique under HPHT. Ideally the CNTs will be distributed along the grain boundary of diamond to form a network structure. These materials have great potential to promote thermal management of electronic packaging.

For instance, ceramic heat spreaders (e.g., AlN) and MMC heat spreaders (e.g., SiC/Al) have been used to cope with the increasing amounts of heat generation. However, such materials have a thermal conductivity that is not greater than that of Cu, hence, their ability to dissipate heat from semiconductor chips is limited. A typical semiconductor chip contains closely packed metal conductor (e.g., Al, Cu) and ceramic insulators (e.g., oxide, nitride). The thermal expansion of metal is typically five to ten times that of ceramics. When the chip is heated to above 60°C, the mismatch of thermal expansions between metal and ceramics can create microcracks. The repeated cycling of temperature tends to aggravate the damage of the chip. As a result, the performance of the semiconductor will deteriorate. Furthermore, when temperatures reach more than 90°C, the semiconductor portion of the chip may become a conductor so the function of the chip is lost. In addition, the circuitry may be damaged and the semiconductor is no longer usable (i.e., becomes “burned out”). Thus, in order to maintain the performance of the semiconductor, its temperature must be kept below a threshold level (e.g., 90°C). A conventional method of heat dissipation is to contact the semiconductor with a metal heat sink. A typical heat sink is made of aluminum that contains radiating fins. These fins are attached to a fan. Heat from the chip will flow to the aluminum base and will be transmitted to the radiating fins and carried away by the circulated air via convection. Heat sinks are therefore often designed to have a high heat capacity to act as a reservoir to remove heat from the heat source. The above heat dissipation methods are only effective if the power of the central processing unit (CPU) is less than about 60 W. For CPUs with higher power, more effective means must be sought to keep the hotspot of the chip below the temperature threshold. Alternatively, a heat pipe may be connected between the heat sink and a radiator that is located in a separated location. The heat pipe contains water vapor that is sealed in a vacuum tube. The moisture will be vaporized at the heat sink and condensed at the radiator. The condensed water will flow back to the heat sink by the wick action of a porous medium (e.g., copper powder). Hence, the heat of a semiconductor chip is carried away by evaporating water and removed at the radiator by condensing water. Although heat pipes and heat plates may remove heat very efficiently, the complex vacuum chambers and sophisticated capillary systems prevent designs small enough to dissipate heat directly from a semiconductor component. As a result, these methods are generally limited to transferring heat from a larger heat source, e.g., a heat sink. Thus, removing heat via conduction from an electronic component is a continuing area of research in the industry. One promising alternative that has been explored for use in heat sinks is diamond-containing materials. Diamond can carry heat away much faster than any other material. The thermal conductivity of diamond at room temperature (about 2,000 W/mK) is much higher than either copper (about 400 W/mK) or aluminum (250 W/mK), the two fastest metal heat conductors commonly used. Additionally, the thermal capacity of diamond (1.5 J/cm3) is much lower than copper (17 J/cm3) or aluminum (24 J/cm3). The ability for diamond to carry heat  away without storing it makes diamond an ideal heat spreader. In contrast to heat sinks, a heat spreader acts to quickly conduct heat away from the heat source without storing it. In addition, the thermal expansion coefficient of diamond is one of the lowest of all materials. The low thermal expansion of diamond makes joining it with a low thermally expanding silicon semiconductor much easier. Hence, the stress at the joining interface can be minimized. The result is a stable bond between diamond and silicon that does not delaminate under the repeated heating cycles (Sung 2007).

As a result, diamond heat spreaders have been used to dissipate heat from high power laser diodes, such as that used to boost the light energy in optical fibers. However, large area diamonds are very expensive; hence, diamond has not been commercially used to spread the heat generated by CPUs. In order for diamond to be used as a heat spreader, its surface must be polished so it can make intimate contact with the semiconductor chip. Furthermore its surface may be metallized (e.g., by Ti/Pt/Au) to allow attachment to a conventional metal heat sink by brazing. Many current diamond heat spreaders are made of diamond films formed by CVD. The raw CVD diamond films are now sold at over $10/cm2, and this price may double when it is polished and metallized. This high price would prohibit diamond heat spreaders from being widely used except in those applications (e.g., high-power laser diodes) where only a small area is required or no effective alternative heat spreaders are available. In addition to being expensive, CVD diamond films can only be grown at very slow rates (e.g., a few micrometers per hour); hence, these films seldom exceed a thickness of 1 mm (typically 0.3–0.5 mm). However, if the heating area of the chip is large (e.g., a CPU), it is preferable to have a thicker (e.g., 3 mm) heat spreader. As such, cost-effective devices that are capable of effectively conducting heat away from a heat source, continue to be sought through ongoing research and development efforts (Sung 2007).

Because of their exceptionally high hardness, excellent wear resistance, and thermal stability, SiC–diamond composites have been conventionally used in various industrial applications such as machining, grinding, drilling, and mining. SiC–diamond composites have been prepared by a variety of methods that include CVD, HPHT liquid-phase sintering, and low vacuum liquid-phase infiltration. Most available SiC–diamond composites are composed of microcrystalline diamond held together by microcrystalline SiC. Figure 7.3 shows an examples of SiC–diamond heat spreaders as fabricated or with Si coating for improving surface roughness. The thermal conductivity of the SiC–diamond spreader is approximately 600 W/m K, CET is 1.8 ppm/K. These SiC–diamond heat spreaders are now in production for IBM-server heat spreaders (Zweben 2006).

Fig. 7.3
figure 3

Diamond particle-reinforced SiC composite heat spreader (a) with silicon coating that improves surface roughness; (b) without silicon surface coating (Zweben 2006)

Full size image

Despite their extraordinary hardness and wear resistance, SiC–diamond composites have relatively low fracture toughness (<6 MPa·m1/2), which limits their potential applications. Fracture toughness of SiC–diamond composites has been improved by incorporating nanocrystalline diamond into the composites. It is believed that the nanocrystalline diamond and SiC hinder dislocation growth and microcrack propagation in the composite better than microcrystalline diamond and SiC do, which improves fracture toughness (Ekimov et al. 2000). The composite was prepared by the liquid silicon infiltration of nanocrystalline diamond powder under HPHT conditions HPHT (7.7 GPa, 1,700–2,300 K). The composite displayed high fracture toughness (10 MPa·m1/2) but was only partially densified; infiltration depth was only 1–2 mm because the pores closed very quickly during infiltration due to the “self-stop process;” as silicon infiltrates through the pores, it reacts rapidly with diamond to form a silicon carbide phase that seals the pores and prevents further infiltration. Alternative methods that are not limited by the self-stop process may be required to overcome problems relating to the self-stop process in order to provide fully dense composites with high fracture toughness. Such a method would also minimize graphitization of the nanocrystalline diamond, which has also been a problem in the past. Fully dense, SiC–diamond composites having high fracture toughness have been fabricated with a method involving ball-milling microcrystalline diamond powder and microcrystalline silicon powder to form a powder mixture of microcrystalline diamond and amorphous silicon, then sintering the ball-milled powder mixture at a pressure of about 5 GPa to about 8 GPa and a temperature of about 1,400 K to about 2,400 K for a sufficient period of time to form a fully dense diamond-silicon carbide composite of microcrystalline diamond and nanocrystalline silicon carbide having a fracture toughness of at least 10 MPa·m1/2 and with minimal graphitization (Qian and Zhao 2005).

SiC–Carbon Composites

C–C composites are resistant to heat and chemicals and lightweight, and have a high strength. Accordingly, C–C composites are useful as heat-resistant materials used in nonoxidizing atmosphere, and are particularly used for brake discs and pads or the like of airplanes and automobiles because of their superior resistance to heat attack. Unfortunately, carbon materials are generally oxidized at about 500°C, and therefore, are not suitable to use in high-temperature atmospheres except for being used for a very short time. In order to prevent the physical or chemical degradation of the C–C composite resulting from oxidation, or, if it is used as a sliding member, in order to prevent the reduction of mechanical strength resulting from oxidation and abrasion, the reduction of frictional properties at low temperatures, and the reduction of frictional properties caused by rain drops and other water attached to the sliding member, the C–C composite is impregnated with melted Si to react with the Si and thus to transform the carbon of the C–C composite into C/SiC. In general, polyacrylonitrile  (PAN)-based carbon fiber is used as the carbon fiber of the C–C composite that is to be impregnated with melted Si to form SiC. One method to produce C–C composite includes a step for preparing a material comprising a bundle of carbon fiber wherein a powdery binder pitch acting as a matrix in the material comprising the carbon fiber bundle and ultimately acting as free carbon being free from the carbon fiber bundle is added to the carbon fibers aligned in a single direction, and then phenol resin powder, for instance, is added thereto. The carbon fibers thus prepared are covered with a flexible coating made of a resin such as a thermoplastic resin to produce a preformed yarn used as a flexible intermediate material. The preformed yarn is formed into sheets. The C–C composite is produced by stacking a desired number of the sheets one on top of another in such a manner that the directions of the carbon fiber orientations are perpendicular to each other, and subsequently performing predetermined process steps. The resulting C–C composite is impregnated with molten silicon, and thus a SiC–C/C composite material having a porosity of 5% is produced. Since the SiC–C/C composite material is a stack of carbon fiber sheets whose fiber orientations are perpendicular to each other, it is considered that the composite material has a high tensile strength in the longitudinal direction of the fibers (in the 0° direction or the 90° direction), but a low tensile strength in directions of 45° with respect to the perpendicular carbon fibers. In addition, in a SiC–C/C composite prepared by impregnating a C–C composite using PAN-based carbon fibers with a melted Si for transformation into SiC, the reinforcing effect of the carbon fibers is degraded, so that the SiC–C/C composite has no more than a tensile strength close to that of carbon materials not reinforced with carbon fibers. Another modified method to produce the carbon fiber-reinforced SiC composite is by impregnating a carbon fiber-reinforced carbon composite with melted Si. The C–C composite includes short carbon fibers prepared from a pitch-based carbon fiber. The pitch-based short carbon fibers are oriented in two-dimensional random directions in the carbon fiber-reinforced SiC composite (Fukagawa and Kubo 2009).

The methods described above are all aimed at enhancing the toughness of SiC ceramics by the addition of C or transformed SiC fibers to ceramics. Especially for the impartation of high toughness, the production of continuous SiC fiber reinforced SiC ceramics is carried out. None of the continuous SiC fibers heretofore developed is capable of retaining its strength intact at elevated temperatures exceeding 1,400°C. In the circumstances, there has developed a need for practicable fibers capable of retaining strength at such elevated temperatures. Carbon fibers are capable of retaining strength at high temperatures (2,000°C) as compared with SiC fibers. When such carbon fibers are combined with SiC matrix, the strength of the carbon fibers is degraded by a reaction which occurs between the matrix and the fibers. When continuous carbon fibers are used, it is necessary to curb the reaction of the carbon fibers with the matrix. To eliminate the degradation, a method for the production of a continuous carbon fiber reinforced SiC composite, which comprises impregnating continuous carbon fibers coated with at least one member selected from the group consisting of SiC, TiC, TiB2, and TiN with a slurry mixture which comprises at least one member selected from the group consisting of SiC, Si3N4, SiO2, and Si with a thermosetting resin or a high-carbon caking agent thereby obtaining impregnated continuous carbon fibers, then molding the impregnated continuous carbon fibers thereby obtaining a shaped article, curing the shaped article, subsequently heating the shaped article in an inert gas to make a carbon fiber–carbon composite having the interstices between the carbon fibers filled with carbon and the aforementioned silicon or silicon compound, further impregnating the carbon fiber–carbon composite with melted Si, and thereafter heat treating the impregnated composite (Nakano et al. 1998).

On the other hand, carbon fiber-reinforced silicon carbide matrix (SiC–C) composites processed by CVI are candidate materials for aerospace thermal structures. Carbon fibers can retain properties at very high temperatures, but they are generally known to have poor oxidation resistance in adverse, high-temperature environments. Nevertheless, the combination of CVI process and SiC matrix with higher stiffness and oxidation resistance, the interfacial coating, and additional surface-seal coating provides the necessary protection to the carbon fibers, and makes the material viable for high-temperature space applications operating under harsh environments. Furthermore, SiC–C composites, like other CMCs, exhibit graceful noncatastrophic failure because of various inherent energy-dissipating mechanisms. The material exhibits nonlinearity in deformation even at very low stress levels. This is the result of the severe matrix microcracking present in the as-processed composite because of large differences between the coefficients of thermal expansion of the fiber and the matrix.

In addition, a comingled blend of carbon and SiC fibers is used to eliminate process-induced matrix microcracking that is typical in SiC–fiber reinforced SiC–matrix composites. The blending of thermoelastically dissimilar fibers allows for tailoring the fiber preform thermal expansion to be more favorably compatible with the matrix, while improving the high-temperature mechanical performance and lowering cost over C–fiber reinforced SiC–matrix composites. Silicon carbide matrix composite thrust chambers utilizing hybrid fiber preforms were shown to exhibit virtually zero through-thickness permeability over typical C–fiber reinforced SiC–matrix composites resulting from a hybrid preform comprising an optimized blend of carbon and SiC fiber reinforcement.

Often, some of the desirable property characteristics allow composites to offer advantages over conventional structural materials, such as tailoring of composite properties. However, the complexity is in fact responsible for their greater statistical variability and the requirements for more characterization tests. Composite properties are anisotropic as well, having different properties in different directions. This means that characterization of a property such as stiffness, which will vary greatly depending on the orientation of the fiber relative to the direction of the testing, must be repeated for several different directions and loading conditions. The fabrication process for composites also introduces statistical variations in properties and geometry. A composite part is produced in a number of steps, each of which introduces statistical variability. The matrix is usually produced from a combination of raw materials; and the fiber, which has its own set of properties, is often coated or surface treated, introducing yet another source of variability. The processes are usually performed by various vendors and are not under the control of the fabricator of the composite part. Additional irregularities are introduced by the influence of temperature and moisture. Composites are usually more susceptible to environmental conditions. Changes in environmental conditions produce a significant change in properties, leading not only to a source of property variability, but also to a requirement for additional testing to characterize the effects of these variables. In general, CMCs are complex, have brittle constituents, and are potentially flaw-sensitive materials. They inherently have considerable scatter in their properties. It is important to note that because of the flaw-sensitivity of brittle materials, additional characterization is required to characterize the volume effect. Thus, the advantages that composites bring must be weighed against increased material testing costs. Any CMC material characterization effort based solely on a large test matrix is simply impractical because of time and cost considerations.

Reaction-Bonded SiC Composites

Although SiC can be densified with high temperature and pressure, the process is not a viable commercial process. Reaction bonded SiC is made by infiltrating compacts made of mixtures of SiC and carbon with liquid silicon. The silicon reacts with the carbon forming SiC. The reaction product bonds the SiC particles. Any excess silicon fills the remaining pores in the body and produces a dense SiC–Si composite. The ratio of SiC to carbon and particle size distribution varies widely in practice. Articles are produced with a wide range of compositions and properties. At one extreme, carbon fiber felt or cloth can be infiltrated with liquid silicon, while at the other extreme, an impervious SiC body can be made with a small amount of carbon. Most reaction bonded SiC is made with formulations that contain an organic plasticiser, carbon, and SiC particles. This mixture is ideally suited to near net formation by pressing, injection molding, or extruding. Furthermore, because the reaction process typically gives a dimensional change of <1%, manufacturers have excellent control of component tolerances (Ceram 2001).

Reaction bonded SiC ceramics combine the advantageous properties of high performance traditional ceramics, with the cost effectiveness of net shape processing. Table 7.1 shows properties of some typical reaction bonded SiC compared with CVD SiC composites (Karandikar 2003). These materials provide high surface hardness, very high specific stiffness, high thermal conductivity, and very low CTE. The processing consists of two steps. First, a carbon containing near net shape porous preform is fabricated; and second, the preform is reactively infiltrated with molten Si to form a primarily SiC body. Reaction bonded SiC ceramic offers extremely high levels of mechanical and thermal stability. It possesses low density (similar to Al alloys) and very high stiffness (~70% greater than steel). These properties lead to components that show little deflection under load, allow small distances to be precisely controlled with fast machine motion, and do not possess unwanted low frequency resonant vibrations. In addition, due to the high stiffness and hardness of the material, components can be ground and lapped to meet stringent flatness requirements. Furthermore, as a result of very low CTE and high thermal conductivity, reaction bonded SiC components show little motion with temperature changes and are resistant to distortion if localized heating occurs. Also, due to an excellent CTE match with Si, reaction bonded SiC ceramics are well suited as substrates for Si handling operations. Furthermore, both Si and SiC possess refractory properties, which yield a composite with good performance in many high-temperature and thermal shock applications. Finally, dense, high purity SiC coatings can be applied when extremely high purity and/or superior resistance to corrosion are required (Evans et al. 2003).

Table 7.1 Properties of some typical reaction bonded SiC compared with CVD SiC composites
Full size table

Many processes have been developed to form reaction-bonded SiC. Figure 7.4 shows a schematic of this process. In this process, the preform containing the reinforcement and a carbon precursor or binder is “carbonized” in an inert atmosphere above 600°C to convert the precursor to carbon. Next, the preform is placed in contact with Si metal or alloys of Si in an inert or vacuum atmosphere and heated to above the melting point of the alloy. As a result of the spontaneous wetting and reaction between carbon and molten Si, the preform is infiltrated completely. The carbon in the perform reacts with the Si forming SiC and in the process bonds the reinforcement together. Some residual Si remains (Karandikar et al. 2007). Typically, a molten infiltrant containing silicon and one or more sources of boron are contacted to a porous mass that contains at least some boron carbide, and also containing at least some free carbon. The molten infiltrant infiltrates the porous mass without a pressure or vacuum assist to form a composite body of near theoretical density. The silicon component of the infiltrant reacts with the free carbon in the porous mass to form in situ SiC as a matrix phase. Furthermore, the tendency of the molten silicon to react with the boron carbide component can be suppressed or at least greatly, attenuated by the alloying or doping of the silicon with the boron source. The resulting composite body thus comprises boron carbide dispersed or distributed throughout the SiC matrix. Typically, some residual, unreacted infiltrant phase containing silicon and boron is also present and similarly distributed or interspersed throughout the matrix (Si–SiC–B4C). Reaction formed SiC composites featuring a boron carbide reinforcement possess stiffness (e.g., elastic or Young’s Modulus) comparable with their counterparts featuring the usual SiC reinforcement, but exhibit a lower specific gravity for the same volumetric filler loading. Accordingly, such B4C reinforced SiC composites will find utility in applications requiring low mass and high stiffness, such as equipment requiring precise motion control, often at high accelerations. Additionally, because of the extreme hardness and low specific gravity of boron carbide, such composites are attractive armor material candidates (Evans et al. 2003).

Fig. 7.4
figure 4

Schematic of a reaction bonding process for reaction-bonded SiC composite (Karandikar et al. 2007)

Full size image

Figure 7.5 shows typical microstructure and property of reaction-bonded SiC with laser sintered process (SLS). Three main manufacturing phases: preform SLS, binder burnout, and reactive infiltration. Preform formation utilizes SiC powder of an appropriate average particle size mixed with a multicomponent binder. The preform or green part is then placed in a vacuum furnace to carbonize the binder. The binder chemistry must support accurate component shapes and acceptable surface roughness, a strong green part and maintenance of the part shape during the first furnace operation. Finally, the physics and chemistry of the infiltration process, based on the microstructure of the initial green preform, determine the viability of the manufacturing process and the characteristics of the final composite material. The governing principles of binder function, carbonization, and infiltration must be established to move beyond functional prototyping and into manufacturing arenas of thermal managing components in electronic packaging industry (Evans et al. 2003).

Fig. 7.5
figure 5

Microstructure and properties of reaction-bonded SiC composite by laser sintering process (Evans et al. 2003)

Full size image

The properties of SiC components depend on the material grade. In the case of a fully dense SiC–Si composite, the material demonstrates good bend strength at room temperature (typically 400 MPa), which is maintained to the melting point of silicon (1,410°C) where it decreases to around 250 MPa. Young’s modulus is typically in the range 350–400 GPa. The properties that lead to selection of the material are resistance to wear, resistance to corrosion; the material tolerates a wide range of acids and alkalis, resistance to oxidation, abrasion resistance, good thermal shock resistance due to low thermal expansion coefficient and high thermal conductivity, strength at high temperature, and good dimensional control of complex shapes. The negligible volume change after reacting with liquid silicon means that components can be formed with complex shapes and to exacting tolerances (Evans et al. 2003). Examples of components made by this route include laser mirror blanks, wafer handling devices, optical benches, and active cooling components such as heat pipe. The components are light weight and stiff with excellent thermal stability.

Aluminum-Toughened SiC

Reaction bonded SiC ceramics have many outstanding properties, including high-specific stiffness, low CTE, and high thermal conductivity. However, they have low fracture toughness. As an alternative, Al-toughened SiC composite maintains many of the advantageous properties of reaction bonded SiC while providing higher toughness, higher thermal conductivity, and more tailorable CTE. The composite is produced using a reactive infiltration process, which allows near-net-shape components to be fabricated.

Figure 7.6 shows typical microstructures of aluminum-toughened 12%Si/70%SiC (a); and 70%SiC (b) (Karandikar 2003). Al-toughened SiC provides a nominally 37% increase in fracture toughness relative to standard reaction bonded SiC. In addition, the composite can be used in thin-walled component designs that would be difficult to produce with a low toughness ceramic. The addition of the aluminum to the microstructure does lead to a slight decrease in stiffness; however, the Al-toughened SiC is still well stiffer than comparable Al–SiC MMCs. Thus, the composite is still well suited to applications where very high levels of mechanical stability are required. The presence of the aluminum results in an increase in thermal conductivity relative to the standard SiC product, which is valuable in heat sink applications or in components where localized heating can occur. In addition, the thermal conductivity is in excess of that of most MMCs because no metallic alloying elements are used. The CTE of Al-toughened SiC is low (between those of SiC ceramic and Al–SiC MMCs), and thus provides a high level of thermal stability. In addition, it very closely matches the CTE of AlN, and thus has utility in applications where interfacing with AlN components occurs. Finally, Al-toughened SiC ceramic represents a family of materials. By changing the respective volume fractions of Al, Si, and SiC, the properties can be tailored to meet specific application needs (MMMT 2004).

Fig. 7.6
figure 6

Microstructure of aluminum-toughened 12%Si/70%SiC (a); and 70%SiC (b) (Karandikar 2003)

Full size image

Al-toughened SiC components can be fabricated with a near-net-shape reactive infiltration process. These manufacturable ceramic-metal composites offer high levels of mechanical and thermal stability, with higher fracture toughness relative to conventional ceramic materials, and will be promising for future thermal management applications.

Ceramic Nanocomposites

Since the discovery of CNTs, vast areas of research have opened which also include nanoscale reinforcements in composites in order to improve their mechanical, thermal, and electrical properties. Although the focus of the research in nanotube-based composites has mostly been on polymer-based composites, the unique properties of CNTs have also be exploited in CMCs. The combination of these nanotubes with a ceramic matrix could potentially create composites that have high temperature stability as well as exceptional toughness and creep resistance. Increased thermal and electrical conductivity of the composites even at low volume fractions might also provide the thermal transport needed to reduce material operating temperatures and improve thermal shock resistance in applications such as thermal elements and electrical igniters. The high temperature and high reactive environment during conventional fabrication methods of ceramics damage the CNTs and thus, there is a need of alternate methods of processing the composites. There is also a need for control of interface between nanotube and the matrix for better interfacial bonding. However, the research is still in the embryonic stage and there are many challenges to be resolved before it is ready for use in varied industrial applications (Samal and Bal 2008).

The physical properties of material critically depend on the size of grains due to the fact that for different grain sizes a significantly different fraction of atoms forms the surface of the grain. When the grain size is reduced to the nanoscale, the corresponding nanostructure shows remarkably different properties from those of microsize or bulk materials. For instance, the strength of the one-dimensional (1-D) SiC nanostructure is substantially greater than that of larger SiC structures, and it approaches theoretical values. Consequently, as the reinforcing phase, 1-D nanosize SiC is a promising candidate to fabricate superhard nanocrystalline composites. SiC nanorods were first synthesized successfully in 1995, in the process in which CNTs were converted to SiC nanorods by reaction with SiO vapor. Subsequently, a number of methods of synthesis of 1-D SiC nanostructures have been developed. They fall broadly into the following categories: CNTs confined reaction; carbothermal reduction of SiO by C nanocapsules or amorphous activated C; CVD; reaction between SiCl4 and CCl4 and metal Na as coreductant; and annealing CNTs covered with Si. However, all of the above methods either operated at high temperature via vapor–vapor or solid–vapor reactions, or needed some kind of metal, e.g., Fe, Cu, Ni, as a catalyst to cause and speed up the reaction (Guo et al. 2002; Wang 2006).

The most promising solution is the dispersion of nanoparticles, especially CNTs into the ceramic matrix. Apart from structural applications, nanocomposite technology can also be applied for fabricating functional ceramics like ferroelectric, piezoelectric, varistor, and ion-conducting materials. Incorporating a small amount of ceramic and metallic nanoparticles into barium titanate, zinc oxide or cubic zirconia can significantly improve their strength, hardness, and toughness, which are very important in creating highly reliable electric devices operating in severe environmental conditions. In addition, dispersing conducting metallic nanoparticles or nanowires can enhance the electrical properties. For example, the range of electroceramic nanocomposites for information and charge storage has revolutionized the electronic industry. In the nanoscale, quantum effects can be utilized to modify energy states and electronic structures of components. Conducting nanosized particles dispersed in a dielectric matrix (i.e., Ni in PZT) can improve dielectric properties. The solution sol–gel technique, in particular, offers an opportunity to produce not only ultrahomogeneous materials but also heterogeneous or nanocomposite materials. After crystallization and densification, those materials are appropriate for numerous applications, such as electronic or structural materials. The most common methods for processing nanocomposites are mechanical alloying, sol–gel synthesis, and by thermal spray synthesis. Mechanical alloying occurs as a result of repeated breaking up and joining of the component particles which can prepare highly metastable structures such as amorphous alloys and nanocomposite structures with high flexibility. Scaling up of synthesized materials to industrial quantities is easily achieved in this process, but purity and homogeneity of the structures produced remains a challenge. In addition to the erosion and agglomeration, high-energy milling can provoke chemical reactions that are induced by the transfer of mechanical energy, which can influence the milling process and the properties of the product. This is used to prepare magnetic oxide-metal nanocomposites via mechanically induced displacement reactions between a metal oxide and a more reactive metal. Sol–gel is another process in which aerogels are the precursors. Aerogels are made by solgel polymerization of selected silica, alumina or resorcinol-formaldehyde monomers in solution and are extremely light but very porous, mostly nano-sized pores. In the nanocomposites derived from aerogels, the product consists of a substrate like silica gel and one or more additional phases with one of the phases with dimension of the order of nanometers. Thermal spray coating is another commercially relevant, proved technique for processing nanostructured coatings. Thermal spray techniques are effective because agglomerated nanocrystalline powders are melted, accelerated against a substrate and quenched very rapidly in a single step. This rapid melting and solidification promotes retention of a nanocrystalline phase and even amorphous structure. Retention of nanocrystalline structure leads to enhanced wear behavior, greater hardness, and sometimes a good coefficient of friction compared to conventional coatings. As the research in nano-reinforced ceramics is still in its infancy, it is not clear which, if any, of the above toughening and strengthening mechanisms apply to ceramic nanocomposites (Ajayan et al. 2003).

Figure 7.7 shows typical microstructures of ceramic nanocomposites reinforced with dispersed and aligned CNTs. The CNT reinforcements promise to increase the fracture toughness of the composites by absorbing energy through their highly flexible elastic behavior during deformation, which will be especially important for nanotube-reinforced CMCs. However, materials fabrication difficulties have limited research on the composites. The major concern is to obtain a uniform dispersion of nanotubes in the matrix. Damage or destruction of the nanotubes is often observed because of the high temperatures and highly reactive environments associated with many methods of processing ceramic and metal matrices. Lack of optimized fabrication and processing of nanotubes has also restricted research efforts, although this seems to have become less of a problem in recent years as more groups produce their own nanotubes with catalyzed CVD methods. Powder processing methods have been used to fabricate ceramic matrices. The small diameter and large aspect ratio of the nanotubes can make it difficult to obtain a good mixture of the two phases prior to sintering or hot pressing. Some success has been achieved with conventional milling techniques, primarily with the use of low to moderate nanotube volume fractions. By minimizing the ceramic particle size, large surface CNT materials can be achieved. Nano-SiC ceramic-based CNT composites have been successfully fabricated by mixing SiC nanoparticles with 10 wt% CNTs and hot pressing at 22,730 K. The CNTs played a strengthening and toughening role in this CMC. Both bend strength and fracture toughness of this SiC–10% CNTs composite increase by about 10% over monolithic SiC ceramic. Another interesting approach is to use a mixture of ceramic powder and catalyst for nanotube growth (Fe, Ni, or Co), followed by a CVD process that grows nanotubes inside the particulate perform. This can produce a uniform mixture of the two phases as long as mass transport of the carbon-containing gas (e.g., C2H2, CH4, etc.) into the powder bed does not produce significant gradients. In some cases, the nanotubes are damaged by high-temperature sintering. However, spark-plasma sintering has been used to produce materials where the nanotube structure is retained. The success of this method apparently results from the use of lower temperatures and shorter firing times. While the sintering of traditional whisker and fiber reinforced composites is typically plagued by differential densification rates, this may be less of a problem with nanotubes because of their small dimensions. Conventional hot pressing and hot isostatic pressing methods have also been used to produce nanotube-reinforced composites. Tape casting and the lamination method have been used to achieve better dispersion of CNTs in an alumina matrix. The tribological studies of the fabricated composites showed that tape casting helped in achieving higher dispersion. The presence of CNTs is also expected to inhibit grain growth during high-temperature processing. This beneficial effect is also known to occur during high-temperature processing of a variety of other composite materials. Several methods have been used to form nanotubes inside preexisting porous matrices. One method of producing nanotubes inside a porous ceramic is to use anodic alumina templates with controlled pore structures. This approach is generally limited to coatings or membranes up to ~100 μm thick. The resulting materials consist of aligned nanotubes whose diameter is determined by the preexisting cylindrical (or nearly cylindrical) pores in the alumina. The dimensions of these materials make them potentially applicable as coatings. Also, the unidirectional microstructure is an advantage for fundamental investigations of mechanical properties. Nanoscale powders with CNTs provide another opportunity for creating dense ceramic-matrix powders with enhanced mechanical properties. The strength and fracture toughness of hot pressed α-alumina is typically much greater than that of conventional grain size polycrystalline alumina. The addition of CNTs to the alumina results in lightweight composites with even greater strength and fracture toughness. The mechanical properties of such composites depend strongly on the processing methods and surface treatment of the CNTs. The most interesting and challenging applications for ceramic-CNT composites are as tough materials. Toughening in CMCs is typically achieved by a weak fiber–matrix interface coupling that permits debonding and sliding of the fibers within the matrix. The closing forces exerted by fibers on matrix cracks that propagate around the fibers constrains the crack growth, and the work required to pull broken fibers out against residual sliding friction at the fiber–matrix interface imparts significant fracture toughness to CMCs. Weak interfaces between a CNT and many ceramic matrices is expected, and pyrolytic carbon is the most common interface material deposited onto fibers to induce such debonding. Thus the mechanisms of interface debonding and sliding, and associated toughening should be operative in CNT-CMCs unless the sliding resistance is too low (Curtin and Scheldon 2004).

Fig. 7.7
figure 7

Typical microstructure of ceramic nanocomposites reinforced with (a) dispersed carbon nanotubes (CNTs); and (b) aligned CNTs

Full size image

CNTs have been successfully used to enhance toughness of reaction bonded SiC ceramics. Fracture toughness of reaction bonded SiC was increased from 4 to 7 MPa m1/2 (a 73% increase) using CNT reinforcement. Table 7.2 gives a comparison of properties between reaction-bonded SiC–CNT and standard reaction-bonded SiC (Karandikar et al. 2007).

Table 7.2 Comparison of properties between reaction-bonded Carbon nanotube (CNT) reinforced silicon carbide (SiC) and standard reaction-bonded SiC
Full size table

Ceramic-CNT composite systems continue to hold promise but with significant challenges to real success. Given the costs of the materials and processes involved, it is not sufficient to obtain marginal improvements in properties over traditional micron-scale composites or virgin matrices. Yet, with few exceptions, notable enhancements have not been observed. The traditional interplay of careful processing and evaluation, coupled with mechanistic assessment of properties, remains a valid paradigm at the nanoscale and should be assiduously applied to future research in CNT composite systems (Samal and Bal 2008).

Ceramic Matrix Composite Thermal Protection System

TPS and hot structures are required for a range of hypersonic vehicles ranging from ballistic reentry to hypersonic cruise vehicles, both within Earth’s atmosphere and non-Earth atmospheres. There are multiple options for dealing with the severe thermal environments encountered during hypersonic flight. Passive, semipassive, and actively cooled approaches can be utilized. The differences between rockets and air-breathers, for instance, can have a significant impact on the TPS and hot structures. As moving toward air-breathing hypersonic vehicles, the severe thermal structural challenges require a new approach to thermal management, one that includes both TPS and hot structures (Glass 2008).

Thermal structural challenges can be quite severe on air-breathing vehicles. One of the primary thermal structural challenges results from large thermal gradients. For example, in a cryogenic tank containing liquid hydrogen as a fuel, the liquid hydrogen will be −423°F and the outer surface of the TPS may be between 2,000 and 3,000°F. With different materials operating at a wide range of temperatures, attaching the various components (tank, insulation, structure, TPS, etc.) that are growing and shrinking at different magnitudes is challenging. Control surfaces are often hot and are often connected to an actuator inside the vehicle which is much cooler. On some structures, there are thin cross sections (due to the need for low drag) at high mechanical loads. These high mechanical loads are often imposed at elevated temperatures. The stability of the outer mold line is important because it can impact performance. For example, sharp nose leading edges generate shocks which are necessary to maximize airflow into the engines. As a result, leading edges should not ablate, both to generate the desired shocks, and to enable reuse. Steps and gaps can be a problem. Gaps may potentially allow sneak flow, where hot plasma leaks into the structure. Forward facing steps may result in local hot spots, thus locally increasing the surface temperature. Thermal expansion of the propulsion system creates other issues. On air breathers, the propulsion system includes much of the undersurface of the vehicle, is very long, and can grow several inches. It must be attached to the airframe and allow for the differential growth between the propulsion system and the airframe. Other vehicle and structural challenges include affordability. Production costs, life cycle cost, and inspection and maintenance costs are all important. Other issues to consider are damage tolerance, low speed impact such as tool drop, foreign object damage on a runway, hypervelocity impact, weather, and reuse potential. All of these requirements lead to a new approach to thermal protection that move away from the Shuttle Orbiter-type of insulated airframe. The Space Shuttle Orbiter is an aluminum “airplane” with ceramic tiles and blankets, which were developed to enable this type of vehicle to be flown. The required material attributes for hypersonic air-breathing vehicles are high-temperature capability (2,000–4,000°F), high strength at those elevated temperatures, high toughness, light weight, and environmental durability. High specific strength at elevated temperature is the goal, which is obtained with high strength and low density. Metallic options include MMC, super alloys, and titanium. These all have good specific strength, but it drops off by the 2,000°F range. Grouped together as CMC, the SiC–C material, advanced carbon–carbon, and SiC–SiC provide high strength at elevated temperatures which are key for air-breathing vehicles. A CMC is illustrated in Figure 7.8a, where the matrix is reinforced with fibers. The fibers carry the load and the matrix transfers the external load to the fibers. Depending on the processing and fabrication approaches, the fiber–matrix interface could have an interface coating which, if present, is the key to the toughness of the structure. It provides a weak mechanical interface for increased toughness and graceful failure. One of the key components of the CMC is the environmental barrier coating which protects the materials from oxidation at high temperatures. The CMC hot structure control surface approach provides the lowest weight and thinnest cross section, minimal thermal expansion mismatch problems, and a good thermal margin. This approach is also advantageous because of its sufficient strength and stiffness and there is no external insulation required. Disadvantages to this approach include the high manufacturing and tooling costs for the box structure and scale up. This approach may not be recommended for very large structures. Its repair capability is limited and the manufacturing risk is high in cases of production failure or damage. Access for coating, inspection, and maintenance of internal areas is also a disadvantage. A mechanically assembled CMC control surface can be fabricated and assembled using multiple smaller parts, as shown in Fig. 7.8b. This control surface used SiC–C fastened joints and a thin ply torque tube and box structure along with gusset members for load transfer. Advantages to this approach are that the tooling is relatively simple and damaged components can be replaced without a complete scrap of the entire control surface. A disadvantage to this approach is that tolerance buildup can be problematic in assembly of numerous separate parts. High part count is another disadvantage of this approach. Figure 7.8c shows a SiC–Cf load bearing aeroshell that carries the aerodynamic and vehicle axial loads. It is to design and fly a prototype ramjet capable of sustained hypersonic flight. The vehicle weighs approximately 30 kg, and is 1.5 m long, with a diameter of 7 in. The aeroshell is being utilizing CVI. Fuel flows in an annulus between two SiC–Cf tubes. The fin roots are bonded to the aeroshell. Insulation may either be incorporated in the structure or separate, below the aeroshell. This type of aeroshell has the potential for reduced weight (Glass 2008).

Fig. 7.8
figure 8

CMC for thermal protection system and hot structure: (a) Schematic of a CMC reinforced with fibers; (b) mechanically assembled SiC–Cf body flap including SiC–Cf fastened joints and a thin ply torque tube and box structure along with gusset members for load transfer; and (c) SiC–Cf load bearing aeroshell utilized on sustained hypersonic flight missile

Full size image

CMCs can be used to enable many of the components on air-breathing hypersonic vehicles, such as leading edges, acreage, hot structures, and in the propulsion system. For most CMC structures, there are two primary materials and structures technical challenges: fabrication and environmental durability. There are several processes for fabricating SiC–C, and SiC–SiC, each having their own challenges. Small coupons can be fabricated with relative consistency. However, a coupon built today may or may not be similar to one that is built 6 months from now. The key point is that a state-of-the-art material is not the same as a state-of-the-art structure. Just because the material can be fabricated, does not at all guarantee that a structure can successfully be fabricated and that it would survive the required mission life. The following fabrication challenges, again process dependent, arise in going from a small material coupon to a large flight structure, which include thickness (density uniformity through out the structure), complex curvature, large scale, low interlaminar properties, delaminations, critical flaw size, nondestructive inspection, tooling, assembly methods and tolerances, reproducibility, fabrication modeling, manufacturability of structures design, and affordable (cost and schedule) fabrication techniques. Meeting the flight requirements is another challenge that impacts the structure maturity on a component. Operation has a significant impact on the ability to use these materials as structures on flight vehicles. Challenges to consider for flight requirements include thermal loads, thermal gradients, mechanical loads, acoustic and vibration loads, pressure (oxidation), combined loads, and number of cycles. During flight, many of the loads are combined, thus generating potentially severe thermal-structural loading. In addition, the primary environment durability challenge is oxidation resistance, which has a major impact on mission life. The number of cycles required under combined loads, inspection and repair, and the ability to predict mission life all present challenges. Hypersonic air-breathing vehicles will require moving beyond an insulated aluminum “airplane,” such as the Space Shuttle Orbiter, to a vehicle with multiple TPS and hot structure approaches. The ability to build and fly these vehicles successfully will depend on the ability to use multiple types of CMC structures, but only after solving the environmental durability and fabrication challenges (Glass 2008).

Summary

Ceramic composites are being actively developed in many research establishments primarily for structural and load-bearing applications. Fiber, whisker, or particulate reinforced composites can be made tougher and stronger than traditional unreinforced ceramics. Among the key factors responsible for the improved stress resistance are differences in Young’s moduli between the phases and the nature of the interaction at interfaces between matrix and the reinforcement. As user confidence continues to grow and as energy savings, increased productivity, and reduced maintenance are confirmed, the need has emerged for advanced ceramics with improved toughness. Discontinuous reinforced ceramic composites have partially filled this need, but their application is limited in both size and geometry, and their toughness is less than desired for a risk-adverse industry. Continuous fiber reinforced ceramic composites are viewed as the ultimate solution with many applications rapidly becoming commercially viable. Thermally conductive ceramic composites have been actively developed thermal management of electronic packaging. Fiber, whisker, particulate, or nanotube reinforced composites can be made tougher, stronger, and more effective for thermal management with tailored CTE. The desirable characteristics of CMCs include high-temperature stability, high thermal shock resistance, high hardness, high corrosion resistance, light weight, nonmagnetic and nonconductive properties, and versatility in providing unique engineering solutions. The combination of these characteristics makes CMCs attractive alternatives to thermal management of electronic packaging, particularly for high temperature electronic packaging system.

Of these, SiC fibers have been the most widely used because of their high strength, stiffness and thermal stability. SiC matrix CFCCs have been successfully demonstrated in a number of applications where a combination of high thermal conductivity, low thermal expansion, light weight, and good corrosion and wear resistance is desired. SiC matrix CFCCs can be fabricated using a variety of processes, fibers, and interface coatings. Fibers widely used for industrial applications where long life is desired include SiC or mullite. Processes available to fabricate SiC matrix CFCCs and the matrix composition formed include Sic, polymer infiltration (SiCN, SiC), nitride bonding (Si–SiC–Si3N4), and melt infiltration (Si–SiC). The interface coating can be either carbon or boron nitride with a protective overcoat of SiC or Si3N4.

SiC–diamond composites are composed of microcrystalline diamond held together by microcrystalline SiC. The thermal conductivity of the SiC–diamond composite spreader can reach around 600 W/m K, CET is 1.8 ppm/K. These SiC–diamond heat spreaders have been commercialized. Reaction bonded SiC ceramics combine the advantageous properties of high performance traditional ceramics, with the cost effectiveness of net shape processing. These materials provide high surface hardness, very high specific stiffness, high thermal conductivity, and very low CTE. Al-toughened SiC composite maintains many of the advantageous properties of reaction bonded SiC while providing higher toughness, higher thermal conductivity and more tailorable CTE. The composite is produced using reactive infiltration process, which allows near-net-shape components to be fabricated. CNTs have been successfully used to enhance toughness of reaction bonded SiC ceramics. Fracture toughness of reaction bonded SiC was increased from 4 to 7 MPa m1/2 (a 73% increase) using CNT reinforcement.

Hypersonic air-breathing vehicles will require moving beyond an insulated aluminum “airplane,” such as the Space Shuttle Orbiter, to a vehicle with multiple TPS and hot structure approaches. CMCs provide desirable characteristics to meet these requirements, with high-temperature stability, high thermal shock resistance, high hardness, high corrosion resistance, light weight, nonmagnetic and nonconductive properties, and versatility in providing unique engineering solutions. The ability to build and fly these vehicles successfully will depend on the ability to use multiple types of CMC structures, but only after solving the environmental durability and fabrication challenges.