A process for the fabrication of SiC matrix composites has been developed which employs aliquid, highly branched, polycarbosilane (AHPCS). This polymer thermosets on heating at 200-400 °C (or at 100 °C with catalyst) and yields an amorphous SiC with low excess C and O content in 60-80% yield on pyrolysis to 1000 °C. Preforms consisting of C-coated Nicalon SiC fiber cloth, unidirectional Textron SiC SCS-6 fiber layups, or Mo boats packed with SiC whiskers, were infiltrated with the polymer, cured in an autoclave, and pyrolyzed to 1000 °C. Five to eight infiltration cycles gave net shape composites with final densities at 85-94 % of theoretical. The results of 4-point flexure tests on the as-prepared Nicalon composites indicate flexure strengths (aver. 378 MPa) that are comparable to or better than similarly reinforced CVI-SiC matrix composites. The whisker and SCS-6 composites showed a small weight loss (10-20%) on heating to 1500 °C in Ar and little or no weight change or obvious embrittlement in air at 1000 °C.

Ceramic matrix composites (CMCs) are being considered for a broad range of aerospace applications that include various structural components for the aircraft engine and the space shuttle main engine. Use of silicon-based CMCs which have high thermal conductivity, allows for improvements in fuel efficiency due to increased engine temperatures and pressures, which in turn generate more power and thrust. Furthermore, CMCs offer significant potential for raising the thrust-to-weight of gas turbine engines by tailoring directions of high specific reliability using design-based fiber architecture. One of the low-cost processing techniques for the silicon-based CMCs is the reactive melt infiltration [1] of silicon into the preform of carbon-coated silicon carbide fiber. However, fabrication of high performance SiC/SiC composites requires a deeper understanding of the infiltration kinetics such that fibers are protected from adverse reaction with the molten metal, that the preform is thoroughly infiltrated, and that there is no residual silicon left unreacted.

Reactive Vapor Infiltration is a convenient new process for producing high purity MoSi2 matrix composites by reacting compacted mixed Mo + MoSi2 powders ranging from 0 to 80 wt. % MoSi2 with H2 + SiCl4 vapors at temperatures from 1000 to 1400°C. Consideration must be taken of the volumetric increase that accompanies this conversion. The kinetics of the advance of the MoSi2 growth layer and the gravimetric uptake are given as a function of temperature and the composition. The parabolic growth rate constant was independent of composition at least up to 30% MoSi2 composition, but then showed a strong, linear increase with increasing MoSi2 content. There appeared to be a levelling off of rate above 70% MoSi2. The corresponding gravimetric rate constant was relatively insensitive to the starting MoSi2 content. This behavior is at variance with a simple one-dimensional diffusive growth model. Possible reasons for this variance are offered.

Silicon carbide (SiC) based ceramic composites with improved fracture toughness and increased flexure strength have been developed by incorporating some other non-oxide and oxide particles or some whiskers and fibres. Hot-Isostatic Pressing (HIP) has been identified as an important technology for strengthening carbide by surface modification. In this paper, Hot-pressed SiC-TiC with different densities and HIP-SiC/SiC(w) composites were post HIPped under a N2-pressure of 200MPa at 1850°C for lh. The results showed that the open pores were closed and physical and mechanical properties such as density, flexure strength and toughness were obviously improved. For the SiC-TiC composites, the final density can be reached above 98% theoretical density, flexure strength and fracture toughness were increased by 100% and 30-50%, respectively. For the SiC/5vol%-SiC(w) composites, the final flexure strength and fracture toughness were increased from 595 MPa and 6.7 MPa m1/2 to 920MPa and 8.5 MPa m1/2 separately. A possible reaction-HIP densification mechanism for SiC ceramics with open pores is proposed.

A set of numerical analyses of crack growth was performed to elucidate the influence of microcracking on the fracture behavior of microcracking ceramic and ceramic composites. The random nucleation, orientation and size effects of discrete nucleating microcracks and resulting interactions are fully accounted for in a hybrid finite element model. The results obtained from the finite element analysis are compared with the continuum description of the microcracking. Although continuum description can provide a reasonable estimation of shielding, it fails to resolve the details of the micromechanism of toughening resulting from microcracking, since not every shielding event during the course of crack extension corresponds to an increase in the Rcurve. Moreover, as seen in the composite cases, the local events leading to toughening behavior may not be associated with the microcracking, even in the presence of large population of microcracks.

An interactive analytical model has been developed for modeling the behavior of Continuous Fiber reinforced Ceramic matrix Composites (CFCC). The model integrates a large number of micromechanics solutions to problems associated with the microstructure of CFCC materials into an easy to use tool for predicting properties, strengths, and stress states for these materials in unidirectional and laminated forms. Particulate reinforcement and voids can be included in the material description. Inherent in the code is a model for handling the accumulation of micro cracks within the matrix as loading is increased, resulting in a nonlinear stress-strain response of the composite. Sufficient material characteristics are retained within the model to enable sensitivity studies to identify principal causes for material behavior.

The role of the material parameters on the fracture and creep behavior of discontinuous ductile fiber reinforced brittle matrix composite system was numerically investigated. For simulation of fracture behavior, the ductile fibers were modeled using a constitutive relationship that accounts for strength degradation resulting from nucleation and growth of the voids. The matrix is assumed to be elastic and fails according to requirements of a stress criterion. Results indicate that the contribution of ductile reinforcement to the work of fracture value (toughness) of the composite increases with less exhaustion of its work hardening capacity before the onset of matrix failure. At creep regime, for rigidly bonded interfaces, the creep rate of the composite is not significantly influenced by the material properties of the ductile reinforcing phase due to development of large hydrostatic stress and constrained deformation in the reinforcement. Significant increases in room temperature fracture toughness can be achieved without extensively sacrificing the creep strength by ductile discontinuous reinforcements.

Development of Continuous Fiber reinforced Ceramic Composite (CFCC) materials is a process of identifying components which will benefit from CFCC properties, and defining appropriate composite constructions which will provide materials which will meet the structural and thermal requirements of the application. Materials Sciences Corporation (MSC) has been providing analytical support to Textron Specialty Materials in the development of re-heated tubes for metal reheating furnaces. As part of this support, a study has been made of the sensitivity of composite properties to fiber orientation as well as a number of matrix properties which control the stress-strain behavior of the composite.

Fiber-reinforced composites consisting of combination of SiC and C have been explored as materials having thermostructural characteristics suitable for elevated temperature applications [1–4]. There are two opposing strategies. One uses C fibers and a matrix of SiC, usually produced by CVI [5, 6]. The other uses SiC fibers and a matrix consisting of C ; those are designated CERACARBTM and were studied in this paper.

Nicalon fiber is the primary reinforcement in SiC-SiC composites currently produced by a variety of techniques including CVI and polymer infiltration. Low strength retention at high temperatures of the Nicalon fibers limits the choice of manufacturing processes which can be employed to produce low cost SiC-SiC composites. MER has developed a new SiC reinforcement based upon a conversion of low cost carbon fabric to SiC via a Chemical Vapor Reaction (CVR) process. This new SiC filaments exhibit an excellent creep resistance at temperatures up to 1600°C. Several SiC-SiC composites were fabricated using graphite fabric converted to SiC fabric utilizing the CVR process combined with a slurry infiltration and CVI densification. A correlation between processing conditions, microstructure and properties of the SiC-SiC composites are discussed in detail.

Woven fabric ceramic composites fabricated by the chemical vapor infiltration method are susceptible to high void content and inhomogeneity. The condition of such materials may be characterized nondestructively with ultrasonic methods. In this work, longitudinal and shear waves were used in the quantitative determination of elastic constants of NicalonTM/SiC composites as a function of volume percent of porosity. Elastic stiffness constants were obtained for both the inplane and out-of-plane directions with respect to fiber fabric. The effect of porosity on the modulus of woven fabric composites was also modelled and compared to the measured results. Scan images based on the amplitude and time-of-flight of radio frequency (RF) ultrasonic pulses were used for evaluating the material homogeneity for the purpose of optimizing the manufacturing process and for correlation with the mechanical testing results.

The microstructure of Al/α-Al2O3 composites made by infiltrating into dense mullite preforms has been characterized using transmission electron microscopy. Observations revealed that the formation of the Al/Al2O3 composites involves three stages. Initially, infiltrates into a dense mullite preform through grain boundary diffusion, and reacts with mullite at grain boundaries to form a partial reaction zone. Then, a complete reaction takes place in the reaction region between the partial reaction zone and the full reaction zone to convert the dense mullite preform to a composite of α-Al2O3 (matrix) and an Al-Si phase (thin channels). Finally, the reduced Si from the reaction diffuses out of the Al/A2O3 composite through the metal channels, whereas from the molten pool is continuously drawn to the reaction region until the mullite preform is consumed or the sample is removed from the molten pool. Based on the observed microstructure, infiltration mechanisms have been discussed, and a growth model of the composites is proposed in which the process involves repeated nucleation of Al2O3 grains and grain growth.

The shear lag model has been used extensively to analyze the stress transfer in a singe fiberreinforced composite (i.e., a microcomposite). To achieve analytical solutions, various simplifications have been adopted in the stress analysis. Questions regarding the adequacy of those simplifications are discussed in the present study for the following two cases: bonded interfaces and frictional interfaces. Specifically, simplifications regarding (1) Poisson's effect, and (2) the radial dependences of axial stresses in the fiber and the matrix are addressed. For bonded interfaces, the former can be ignored, and the latter can generally be ignored. However, when the volume fraction of the fiber is high, the radial dependence of the axial stress in the fiber should be considered. For frictional interfaces, the latter can be ignored, but the former should be considered; however, it can be considered in an average sense to simplify the analysis. Comparisons among results obtained from analyses with various simplifications are made.

A 3D-unit cell for 0/90 laminated composites has been developed to predict the composite behavior under longitudinal tensile loading condition. 3D contact element has been used to model the fiber matrix interface. Two interface conditions, namely, infinitely strong and weakly bonded, are considered in the analysis. Both large displacement and plastic strain behavior for the matrix are considered to account for the geometric and material non-linearities. Investigations were carried out at three temperatures to compare the composite response obtained from mechanical tests at those temperatures. Stress-strain behavior and the local stress distributions at the fiber as well as at the matrix are presented, and their effects on the failure of the interface are discussed in the paper. The material under investigation was SiCf/Si3N4.

The mechanical properties of fiber reinforced composites depends strongly upon the properties of the fiber/matrix interface. Enhanced fracture resistance and strain to failure are synonymous with debonding and sliding of the reinforcement phase. Thus, the two key properties of the composite are the interfacial toughness and the post-debond sliding stress. After debonding a variety of interfacial sliding phenomena are noted, including: stick-slip, steady-state, and roughness dominated sliding. The interfacial properties, including the coefficient of friction, the radial clamping pressure, asperity amplitude, the elastic properties of the constituents, and the compliance of the test machine, each play a role in the operative sliding phenomenon. Experiments have been conducted to explore each of these phenomena. In addition, models have been developed that rationalize all of the observed behavior.

Matrix cracking from controlled through cracks with bridging filaments was studied in a model unidirectional composite of SiC filaments in an epoxy-bonded alumina matrix. An unbonded, frictional interface was produced by moderating the curing shrinkage of the epoxy with the alumina filler and coating the filaments with a releasing agent. Uniaxial tension test specimens (2.5 × 25 × 125 mm) with filament-bridged through cracks were fabricated by a novel two-step casting technique involving casting, precracking and joining of cracked and uncracked sections. Distinct matrix-cracking stresses, corresponding to the extension of the filamentbridged cracks, were measured in uniaxial tension tests using a high-sensitivity extensometer. The crack-length dependence of the matrix-cracking stress was found to be in good agreement with the prediction of a fracture-mechanics analysis that employed a new crack-closure force - crack-opening displacement relation in the calculation of the stress intensity for fiber-bridged cracks. The prediction was based on independent experimental measurements of the matrix fracture toughness (Kcm), the interfacial sliding friction stress (τ) and the residual stress in the matrix (σmI). The matrix-cracking stress for crack lengths (2a) greater than 3 mm was independent of the crack length and agreed with the prediction of the steady-state theory of Budiansky, Hutchinson and Evans[2]. Tests on specimens without the deliberately introduced cracks indicated a matrix-cracking stress significantly higher than the steady-state stress.

The redistribution of internal stresses in a composite with linear viscoelastic constituents was calculated when the composite is subjected to a constant stress at a temperature where the phases would exhibit time-dependent deformation. It was found for the case of an elastic fiber embedded in a matrix that behaves as a Burgers material under distortion and elastically under dilation, that the normal interfacial stress and the axial stress in the matrix undergo complete relaxation at long times. The implications of these findings are discussed in relation to the behavior of ceramic matrix composites.

The design and performance of an SEM-based microindentor, for interfacial property measurements in CMCs, is described. It enables high resolution imaging and simultaneous load/displacement monitoring with capacity and resolution of 20N ± lmN (load) and 100μm ± 10nm (displacement). Its application to measurement of interface debond and shear stresses for a wide range of fibres and monofilaments is described.

Interfacial properties of Nicalon™-reinforced brittle matrix composites were determined from single-fiber load-controlled push-in tests using a Mechanical Properties Microprobe (Nanoindenter) and single-fiber displacement-controlled push-out tests using the Interfacial Test System (ITS). A comparison between the results from these two tests is presented along with a discussion of data analysis techniques.

The potential use of fiber-reinforced composite materials for high temperature applications makes the development of interface test methodology at those high temperatures very desirable. A facility for performing high temperature fiber push-out tests will be described with emphasis on critical issues in experimental procedure. Examples from several composite systems illustrate the temperature dependence and environmental sensitivity of fiber debonding and sliding. Interpretation of the temperature dependence will be made primarily in terms of changes in residual stresses along with additional effects due to changes in matrix ductility and interfacial wear. Examples will show that high temperature fiber push-out testing can often distinguish between chemical and frictional fiber/matrix bonding in cases where room temperature only testing cannot.