Analysis of damage mechanisms in fiber-reinforced ceramic matrix composites under thermomechanical loading at intermediate and high temperatures

Abstract

This paper examines the damage and failure behavior of woven ceramic matrix composites (CMCs) under thermomechanical loading. The study focuses on S200H (Hi-Nicalon™ SiC fiber with a boron nitride coating in a SiNC matrix) and S400N (carbon fiber with a pyrolytic carbon coating in a SiNC matrix) CMCs. Quasi-static and creep-fatigue tests were conducted at various temperatures and stress levels to investigate failure mechanisms at intermediate and high temperatures. Creep-fatigue tests for S200H were performed at 800°C in an oxidative environment at stress levels corresponding to fractions of the ultimate tensile strength (UTS). S400N samples were tested at 600–900°C and 1200°C under low and high stress levels. Residual strength tests were conducted after cooldown to assess mechanical degradation in samples that did not fail. Fracture surface characterization using confocal microscopy, scanning electron microscopy, and energy dispersive spectroscopy provided insights into failure mechanisms. Results indicate that chemical reactions of non-stoichiometric phases in SiC/SiNC, thermal property mismatch, and pyrolytic carbon coating volatilization in C/SiNC govern quasi-static failure mechanisms. Both CMC systems exhibited a 30% reduction in UTS at elevated temperatures, with strain-to-failure decreasing by 17% in SiC/SiNC and 31% in C/SiNC. The SiC/SiNC samples exhibited an increase in yield strength with higher applied stress levels after 100 hours of creep-fatigue testing due to matrix residual compressive stresses, but a 37% reduction in UTS due to oxidation. In contrast, C/SiNC samples experienced significant diffusion-limited oxidation, leading to rapid fracture within 20 hours under creep-fatigue loading.

Keywords

Ceramic matrix composites damage creep-fatigue oxidation

Introduction

The superior strength-to-weight ratio and high temperature capabilities of ceramic matrix composites (CMCs), compared with nickel-based super alloys, make them the material of choice in a wide range of applications including aircraft and land-based turbines.¹ CMCs are also notably thermally resistant and maintain their mechanical properties even after prolonged exposure to elevated temperatures. Specific CMCs are suitable for use in temperature ranges categorized into intermediate (300°C – 900°C), moderately high (900°C – 1150°C), high (1150°C – 1650°C), and ultra-high (>1650°C).² For example, some CMCs, such as C/SiC, are used in moderately high-temperature (≤1000°C) applications,³ while other CMCs, such as SiC/SiC, are used in high-temperature (≤1650°C) applications such as turbine engines and hypersonic applications.⁴ In these high temperature applications, the material will simultaneously experience oxidation and sustained loading, and the reported literature on damage and degradation of CMCs operating in such harsh environments is limited. In recent years, there has been a growing interest in ultra-high temperature ceramic matrix composites (UHTCMCs) due to their high oxidation and ablation resistance in harsh environments.^5,6 At this ultra-high temperature range, active oxidation occurs, where SiC oxidizes into gaseous SiO, reducing the silica scale protection and allowing further oxygen ingress into the microstructure. Current research focuses on reinforcing the monolithic ultra-high temperature ceramics (UHTCs) matrix with materials that enhance toughness while maintaining superior resistance to environmental degradation.⁷ The two non-oxide CMC systems investigated in this paper are S200H (Hi-Nicalon™ SiC fiber with a boron nitride (BN) coating in a SiNC matrix) and S400N (carbon fiber with a pyrolytic carbon (PyC) coating in a SiNC matrix). C/SiNC is primarily used in aerospace applications, including nozzles, nozzle ramps, leading edges, thermal protection systems (TPS), thrusters, turbines, body flaps, and turbopumps.^3,8 However, C/SiNC is known to suffer from significant oxidation-induced damage at high temperatures.³ In contrast, SiC/SiNC exhibits greater oxidation resistance and is currently used in energy production (e.g., land-based gas turbines), propulsion (e.g., turbine and rocket engines), and structural applications (e.g., hypersonic aeroshells).^9–11 A comprehensive study is needed to evaluate the thermomechanical response of these material systems and to investigate the primary damage mechanisms and environmental degradation effects in service environments.

Manufacturing-induced defects also play an essential role in damage initiation and progression. Common CMC manufacturing techniques include melt infiltration (MI), chemical vapor infiltration (CVI), and polymer infiltration pyrolysis (PIP). Depending on the manufacturing process, pre-existing flaws are induced, such as porosity and matrix microcracks^12–14 or unreacted elements or compounds, including residual silicon, free carbon, and pre-ceramic polymer.⁴ Previous studies have investigated the material degradation behavior of CMCs and quantified various failure mechanisms. Halbig et al.³ discussed the oxidation and stress effects on a C/SiC composite. The composite under investigation was manufactured using T300 carbon fiber with a PyC coating in a CVI SiC matrix. It was found that the two primary oxidation kinetic regimes were chemical reaction control and diffusion control. In an unstressed state and under purely thermal loading, the composite exhibited self-healing effects through thermal expansion, which resulted in manufacturing-induced residual thermal stress relief and partial microcrack closure. In conjunction with silica formation, this residual stress relief and partial crack closure sealed the remaining crack openings. However, various dependencies were observed under a stressed state for different temperature ranges. At intermediate temperatures, the time to failure strongly depended on the temperature. Specifically, failure rates increased with temperature. The time to failure at high temperatures was determined by applied stress rather than applied thermal load since the times to failure were relatively short in the entire high-temperature range. Morscher¹⁵ studied the manufacturing-induced defects that cause material degradation mechanisms for a C/SiC composite system where uncoated carbon fibers reinforced a CVI SiC matrix. The composite system was found to have a high volume fraction of microcracks due to the mismatch in the coefficient of thermal expansion (CTE) of the constituents and the cooldown from manufacturing processing temperatures to room temperature. These microcracks enable the oxidation of the carbon fibers, leading to fiber volatilization.

Morscher et al.¹⁶ experimentally investigated the creep and high-temperature fatigue behavior of BN-coated Sylramic silicon carbide (SiC) fibers in an MI SiC matrix. The authors discussed damage evolution, failure mechanisms, and changes in the material’s mechanical properties. At elevated temperatures, the two leading causes for strength degradation were oxidation-induced propagation of cracks that did not display fiber-bridging of matrix cracks and creep-controlled fiber strength degradation. The matrix crack density at elevated temperatures was similar to that at room temperature. Gowayed et al.¹⁷ explored the effect of creep-fatigue testing on the accumulation of strain in the same material system, demonstrating that the matrix is the constituent more susceptible to creep. Accordingly, the causes of material property degradation were matrix cracking and subsequent oxidation products, combined with the repeated unloading and reloading of the sample. Partial strain recovery was also observed when the samples were unloaded, indicating the presence of irreversible deformation, such as fiber/matrix sliding and oxidation-induced creep strain. The oxidation effects of pure thermal loading on SiC/SiC composites manufactured through CVI were also studied by Nasiri et al.,¹⁸ in which SiC/SiC samples were tested in air as an oxidative environment at high temperatures ranging from 1200°C to 1400°C. In an unstressed state, relatively high oxidation resistance was observed compared to monolithic SiC. Oxidation was found to follow parabolic reaction kinetics due to oxygen diffusion through the oxide layer. The oxide layer thickness depended heavily on the thermal environment and the duration that the sample was held under these conditions. Bhatt et al.⁴ explored the effect of creep on the material behavior of two different SiC/SiC composites with a BN fiber coating. Both CVI and PIP matrix systems were tested. PIP results in more crystallization and shrinkage in the samples during the production process than CVI,¹⁹ leading to a higher volume fraction of microcracks and porosity. PIP samples depicted more fiber-dominated creep than CVI samples since the fibers take on more load.⁴

High-temperature fatigue failure in plain-woven SiC/SiC CMCs is significantly influenced by the meso-woven architecture and porosity distribution. The presence of stitching holes and localized variations in porosity contribute to fracture initiation, with both periodic and aperiodic porosity patterns observed due to the woven structure and manufacturing defects. Notably, stepped cracks predominantly initiate at the fiber surface or boundary under high fatigue stress, accelerating crack propagation and ultimately leading to failure.²⁰ Over the past few decades, extensive research^21,22 has been conducted to develop environmental barrier coatings (EBCs) to mitigate accelerated water vapor-induced SiC surface recession and protect CMCs from environmental degradation caused by molten calcium-magnesium alumino-silicate (CMAS).²³ Yang et al.²⁴ investigated the failure mechanisms of SiC/SiC CMCs with EBCs subjected to tension-tension fatigue using a mechanism-based framework. Furthermore, they developed a novel fatigue damage model based on continuum damage mechanics (CDM) to predict damage evolution and fatigue life. Additionally, Ruggles-Wrenn et al.²⁵ examined the failure mechanisms of SiC/SiC CMCs at 1300°C under fatigue, where the BN interface oxidizes, forming brittle borosilicate, which acts as a bonding agent. This in turn effectively fuses exposed fibers together, leading to localized stress concentration and ultimately causing brittle and catastrophic CMC failure.

This paper presents a comprehensive investigation of the thermomechanical response and damage of CMCs subjected to quasi-static and cyclic loading with various hold times. In particular, CMCs with a SiNC matrix manufactured using the PIP process were explored, and a comparison was drawn between the SiC/SiNC and C/SiNC composite systems. The differences in degradation mechanisms at elevated temperatures were investigated. First, quasi-static tensile testing was performed to obtain the tensile behavior and mechanical properties of the samples. Creep-fatigue tests were conducted at elevated temperatures to capture the material response under typical operating conditions. Creep-fatigue tests involved loading a sample to a specified stress level, allowing the sample to dwell at the stress level for a given hold-time, unloading sample completely, and immediately reloading the sample. These cycles were repeated until either sample failure was achieved or until a previously decided runout time was reached. Residual tensile strength testing was subsequently performed to gather additional information regarding the material degradation. Finally, microstructure characterization and fractography were conducted to analyze the failure mechanisms.

Materials and methodology

Materials

The material systems investigated in this study are S200H (SiC/BN/SiNC) and S400N (C/PyC/SiNC) sourced from COI Ceramics, Inc. The S200H material comprises eight-harness satin (8HS) woven Hi-Nicalon^TM SiC fabric in an amorphous SiNC matrix with a BN-based interphase in a layup of 8 plies in a standard [0/90]_2s configuration. The S400N material comprises five-harness satin (5HS) woven T300 carbon fiber fabric in an amorphous SiNC matrix with a PyC interphase in a layup of 8 plies in a standard [0/90]_2s configuration. Both material systems are non-oxide CMCs manufactured using the PIP process, which involves impregnating the woven fabric with a pre-ceramic polymer solution before being pyrolyzed. The S200H material is pyrolyzed in a nitrogen environment to achieve the BN coating on the fibers, and the S400N material is pyrolyzed in a carbon environment to achieve the PyC coating meant to protect them from oxidation. Once sufficiently coated, the fibers are re-impregnated with the pre-ceramic polymer and subjected to pyrolysis in iterations until a finished composite is produced. Manufacturing-induced defects, such as microcracks and porosity, arise from the shrinkage of the pre-ceramic polymer during its conversion into an amorphous SiNC matrix.

Quasi-static testing

As depicted in Figure 1, the test setup for quasi-static testing utilized a two-zone Ametco furnace with a 1400°C capacity, mounted on an MTS load frame with a hydraulic gripping system and a built-in internal linear variable differential transformer (LVDT) for displacement measurement. K-type thermocouples from Omega Engineering were attached to the sample surfaces using Hi-Strength Ceramic Adhesive from Cotronics, which withstands the studied temperature range. These thermocouples were used to analyze the furnace temperature profile and determine the time required for the sample to reach the desired temperature. This measurement was performed initially on one sample, after which subsequent tests were conducted following the waiting period determined from the first sample. The thermocouple wires were threaded through a port at the top of the furnace and kept separate before being connected to thermocouple plugs, which were then inserted into thermocouple readers. Two thermocouple readers, each with two ports, were used to monitor the temperature recorded by all three thermocouples.

To achieve repeatable results in high-temperature tests, two key factors must be considered in the test setup: (i) accurate temperature control and (ii) minimizing bending stress on the test samples.²⁶ While the furnace has built-in thermocouples that measure the ambient temperature in the furnace, a temperature profile showed that the temperature gradient is not constant. The thermocouples were attached to the sample’s surface, ensuring that the tensile test was only started once the sample reached the desired temperature. Fiberglass sheets were utilized to thermally isolate the furnace from the hydraulic grips of the MTS load frame to ensure the hydraulic grips did not overheat from the furnace. The fiberglass sheets also served to isolate the thermocouple wires to ensure they did not cross each other after exiting the furnace and to prevent the thermocouple wires from coming in contact with the metal of the furnace itself. In addition to these procedures aimed at facilitating a uniform temperature profile, further steps were taken to enhance axial alignment of the test samples. Both measures help mitigate the potential for residual bending stresses.

Quasi-static tensile tests were performed on S200H material at room temperature, 800°C, and 1200°C, using a total of seven samples. For S400N material, tests were conducted at 300°C, 600°C, 900°C, and 1200°C, with a total of 18 samples. All tests used dogbone-shaped specimens with a gauge length of 100 mm. The tests were carried out under a constant crosshead displacement rate of 0.06 mm/min, yielding a nominal strain rate of approximately $1 \times 10^{- 5}$ s⁻¹. The displacement rate and the corresponding strain rate are within the range reported in the literature.^10,26–31 Because only crosshead (actuator) displacement was measured (via an internal LVDT), the strain in the gauge section may differ from this nominal value if any compliance (deformation) occurs within the load train or if there is slippage at the grips. This approach does not provide the same direct measurement that a clip-on or extensometer-based strain gauge would. In future work, additional compliance corrections or the use of an external extensometer could help refine the accuracy of strain measurements.

Creep-fatigue testing

The high-temperature test setup for the creep-fatigue tests was the same as that of the quasi-static tests; however, the creep-fatigue tests were load-controlled. CMC creep tests are conventionally performed at one stress level below the proportional limit, at the proportional limit, and at least one stress level above the proportional limit. The tests conducted below the proportional limit help isolate the effects of creep from those due to the combined effects of damage, creep, and oxidation at a stress level above the proportional limit. However, the proportional limits for S200H and S400N are very low (approximately 50 MPa), which would only be about 13% of the ultimate tensile strength (UTS). Such a low stress level would not result in significant creep behavior.

The four stress levels at which creep-fatigue tests were performed for the S200H samples were 20%, 30%, 40%, and 50% of the UTS at an intermediate temperature of 800°C. The UTS depends on temperature and was determined from the previously conducted quasi-static testing. For the S200H samples, the load frame was programmed to apply the desired load in 1 minute with a hold-time of 10 hours, followed by 1 minute to completely unload the sample before immediately reloading again. This was repeated 10 times for a total run time of 100 hours. The tests were conducted for two different SiC/SiNC samples at each stress level. A stress power creep law was used to fit the secondary creep strain data at 800°C, as described in the Results section (equations (1)–(3)). Residual strength tests were also performed on the samples after creep-fatigue testing. The procedure for the residual strength test was identical to that of the quasi-static test, allowing for straightforward comparison of the results.

For the S400N, the samples were tested under one low and one high stress level at various temperatures. The proportional limit was found to be around 35% of the UTS at 900°C. Accordingly, two samples were tested at high-stress levels of 72.5% and 60% of the UTS at 900°C. However, both samples failed within an hour. For samples tested at 600°C, the high-stress level was set to 34.2%, and the low-stress level was set to 20% of the UTS at 600°C. Even under these drastically reduced stress levels, the S400N samples failed after only one full 10-hour hold. To obtain true creep-fatigue behavior, the hold time was reduced from 10 to 2 hours. The subsequent tests were conducted using a low-stress level of 20% and a high-stress level of 34.2% of the UTS at the corresponding temperature. The test parameters for S400N are summarized in Table 1. The damage and failure mechanisms of the two CMC systems under creep-fatigue, particularly those related to oxidation and microstructural features, are comprehensively discussed in the Results and Microstructural analysis sections.

Table 1.

Creep-fatigue test parameters for C/SiNC CMC.

S400N creep-fatigue tests
Hold time (hours)	Temperature (°C)	Max stress	Runout time (hours)	# of repeat tests
2	600	62.23 MPa (20% UTS)	20	1
10	600	62.23 MPa (20% UTS)	20	1
2	900	42.58 MPa (20% UTS)	10	2
2	1200	42.93 MPa (20% UTS)	5	1
2	600	106.48 MPa (34.2% UTS)	15	1
10	600	106.48 MPa (34.2% UTS)	15	1
2	900	127.78 MPa (60% UTS)	1	2
2	900	127.78 MPa (60% UTS)	1	1
2	1200	73.40 MPa (34.2% UTS)	5	1

Confocal microscopy, SEM and EDS characterization

The fractured samples were analyzed using a combination of characterization techniques, where micrographs were captured and analyzed to identify failure mechanisms. The characterization techniques implemented include confocal microscopy, scanning electron microscopy (SEM), and energy dispersive spectroscopy (EDS). The confocal laser scanning microscope used to obtain micrographs of the fracture surfaces was the Zeiss LSM 900, and the SEM and EDS were performed using a Zeiss Auriga Focused Ion Beam/SEM. The fracture surface was left untreated to characterize the observable damage mechanisms accurately. For comparison with a pristine surface, a portion of a sample not exposed to the environmental conditions was cut, cold-mounted, polished, and characterized.

Results

The thermomechanical response of SiC/SiNC and C/SiNC is analyzed at various operating temperatures under quasi-static and creep-fatigue loading conditions.

Quasi-static tests

SiC/SiNC

The results of the SiC/SiNC quasi-static tests conducted over a temperature range from room temperature to 1200°C are presented in Figure 2. A summary of the corresponding average UTS, elastic modulus, and strain-to-failure percentage is provided in Table 2. The room temperature and 1200°C quasi-static test results correlate well with S200H results from the literature by Artz.³² The test results at 800°C show good agreement with one another as shown in Figure 2.

Table 2.

Quasi-static SiC/SiNC CMC results.

Temperature (°C)	Sample #	Ultimate tensile strength (MPa)	Elastic modulus (GPa)	Strain-to-failure (%)
Room temperature	1	491.29	135.82	0.572
	2	483.29	134.96	0.510
	Average	487.3	135.39	0.541
800	3	487.45	135.74	0.556
	4	440.77	122.08	0.473
	5	410.86	125.04	0.447
	Average	446.36	127.62	0.492
1200	6	348.14	109.03	0.521
	7	324.91	112.64	0.375
	Average	336.53	110.83	0.448

With reference to Figure 2 and Table 2, the tensile strength of both samples tested at 800°C is slightly lower than that of the room temperature samples. Moreover, a drastic decrease in tensile strength was found from the intermediate temperature of 800°C to the high temperature of 1200°C, which has been documented before in Hi-Nicalon^TM/BN/SiC CMCs and Hi-Nicalon^TM fibers.^33,34 The strain-to-failure decreased with increasing temperature, reaching a 17% reduction at 1200°C. The degradation in the thermomechanical performance is attributed to chemical changes that occur within the fibers at intermediate and high temperatures. The Hi-Nicalon^TM fiber contains free carbon at the grain boundaries,¹⁴ which is susceptible to oxidation. The free carbon encounters oxygen via preexisting microcracks, which evolve under an applied load, and volatilizes to carbon dioxide.³⁵ The volatilization of the free carbon at grain boundaries is just one factor that contributes to the sharp reduction in strength. The higher temperature also accelerates the rate of diffusion and free carbon consumption. This results in a marked change in strength, even for short-duration thermomechanical testing.³⁶ Another contributing factor to the weakening of the Hi-Nicalon^TM fiber is grain growth and void formation, attributed to decomposition reactions facilitated by the elevated temperatures.^30,33 The variability in the test results is attributed to both experimental and manufacturing-induced factors. Experimental variability includes possible specimen misalignment, while manufacturing variability arises from inconsistencies in the as-produced samples.

C/SiNC

Compared to the quasi-static results for SiC/SiNC, the tensile strength of C/SiNC at the same temperature is significantly lower, as illustrated in Figure 3 and summarized in Table 3. The tensile strength decreases by 30%, while the strain-to-failure decreases by 31% with increasing temperature. The degradation in thermomechanical response is attributed to the significant mismatch of the CTE between the fiber and matrix constituents in C/SiNC compared to SiC/SiNC. The difference in CTE leads to increased internal stresses within the C/SiNC composite during thermal loading, causing damage and hence reducing tensile strength. These stresses exacerbate preexisting microstructural defects, such as porosity and voids, further degrading the material’s mechanical properties.

Table 3.

Quasi-static C/SiNC CMC results.

Temperature (°C)	Sample #	Ultimate tensile strength (MPa)	Elastic modulus (GPa)	Strain-to-failure (%)
300	1	309.01	75.69	0.671
	2	307.08	82.27	0.559
	3	316.95	88.79	0.665
	4	314.26	83.53	0.687
	5	301.18	83.12	0.643
	Average	309.7	82.68	0.645
600	6	293.08	81.37	0.621
	7	335.44	88.95	0.735
	8	319.91	86.71	0.712
	9	319.2	82.17	0.796
	10	288.2	80.38	0.703
	Average	311.17	83.92	0.715
900	11	212.2	79.26	0.480
	12	219.45	78.11	0.566
	13	216.68	80.94	0.529
	14	208.88	85.98	0.358
	15	206.47	81.17	0.462
	Average	212.73	81.1	0.479
1200	16	212.93	75.11	0.535
	17	241.27	89.48	0.405
	18	216.33	84.46	0.382
	Average	223.51	83.02	0.441

The SiNC matrix in the S400N test samples has a significantly higher CTE than the carbon fibers.³⁷ Specifically, the SiNC matrix has a CTE similar to β-SiC, approximately 4.7 × 10⁻⁶°C⁻¹,³⁸ whereas carbon fibers exhibit a CTE ranging from ∼-0.1 × 10⁻⁶°C⁻¹ to ∼ -1.1 × 10⁻⁶°C⁻¹ in the longitudinal direction and approximately ∼7.0 × 10⁻⁶°C⁻¹ in the transverse direction.^39,40 As a result of the residual stresses induced during the PIP manufacturing process at elevated temperatures and subsequent cooling to room temperature, the SiNC matrix contracts more than the carbon fibers, inducing compressive stresses on the fibers and forming pores within the matrix.⁴¹ Not only do these pores act as nucleation points for matrix cracking during thermomechanical testing, facilitating rapid crack propagation and compromising the composite’s strength and durability, but they also serve as oxidation pathways.⁴² The PyC coating on the carbon fibers is rapidly volatilized, leading to exposure of the fibers. The exposed fibers quickly degrade, which, combined with the early matrix cracking, leads to abrupt fracture.

The full variability of the results is provided in Table 3. While most of the scatter is attributed to microstructural variability, the furnace heating time can also influence the thermomechanical response of C/SiNC samples. For example, one of the tests conducted at 1200°C (sample 16 in Table 3) exhibited a lower yield point and higher strain-to-failure compared to the other samples. This discrepancy is attributed to the sample being held at 1200°C for a slightly longer duration before loading. Unstressed oxidation of C/SiNC over extended periods leads to microcrack closure due to silica formation and thermal stress relief.³ While the holding time for the anomalous sample at 1200°C in an unstressed state was not sufficient for significant silica formation, the manufacturing-induced thermal stresses in the matrix were relieved before the quasi-static procedure was performed, resulting in significantly higher strain-to-failure.

Creep-fatigue tests

SiC/SiNC

The results for the eight creep-fatigue tests are depicted in Figure 4. Linear fits were applied to the secondary creep stages of the creep-fatigue test results to obtain the strain rates at the various stress levels. Only the segments easily identifiable as the secondary creep stage were included in the strain rate calculation. The strain rate was accurately calculated using a linear regression fit for each 10-hour stress hold segment, while the unloading and reloading segments were ignored.

Strain rates were obtained for each segment and then averaged to obtain an overall strain rate. The widely accepted secondary creep formulation in the literature^11,43 was used to determine the secondary creep parameters $B$ and $n$ .

ε_{s s} = B σ^{n} t

(1)

A linear fit was applied to this data, as seen in Figure 5, to compute the parameters $B$ and $n$ as follows:

{\dot{ε}}_{s s} = B σ^{n}

(2)

\log ({\dot{ε}}_{s s}) = \log (B) + n \log (σ)

(3)

In equations (1)–(3), $ε_{s s}$ and ${\dot{ε}}_{s s}$ represent the secondary creep strain and creep strain rate, respectively; $σ$ is the applied stress level; $\log ()$ denotes the natural logarithm; and $B$ and $n$ are fitting parameters, found to be 6.0567 × 10⁻¹¹ s⁻¹ MPa^-n and 0.4371, respectively. The secondary creep strain rate ( ${\dot{ε}}_{s s}$ ) is on the order of 10⁻¹⁰ s⁻¹, which is within the expected range for SiC/SiC composites.^4,44,45 Note that the creep parameter values may deviate from those of the constituents (i.e., SiC fiber, SiNC matrix, and BN interphase) due to the complex interactions between the constituent phases and the presence of high localized stresses, which can be significantly greater than the globally applied stresses. Similar to the Quasi-static tests subsection, the scatter in ${\dot{ε}}_{s s}$ , shown in Figure 5, is attributed to manufacturing-induced defects and variability, as well as possible experimental factors such as residual bending moments.

The residual strength results are presented in Figure 6. The modulus remains relatively unchanged across varying applied stress levels. Upon completion of creep-fatigue testing, residual compressive stresses are present in the sample. During creep-fatigue testing, the SiNC matrix experiences a higher creep rate, causing it to shed load to the fibers during the holding stage.^46,47 Upon unloading, the matrix retains compressive residual stresses, while the fibers experience residual tensile stresses.^16,44 Since higher stress is required to overcome these residual compressive stresses, the yield point of the samples after creep-fatigue tests is proportional to the applied stress level, with a 16% improvement observed when increasing the applied stress level from 20% to 50% of the UTS. Nevertheless, the second curve at an applied stress level of 50% UTS, showing a lower yield point, is treated as an outlier because the compressive stresses—and thus the yield stress of the matrix—are expected to be proportional to the applied stress level. Note that due to manufacturing-induced defects, the as-received SiC/SiNC does not exhibit a clear yield point. A summary of the yield stresses obtained from residual strength testing is presented in Table 4. The outlier curve at 50% was excluded from both the results in Table 4 and the analysis. Furthermore, a significant reduction in UTS (up to 37%) and strain-to-failure (up to 41%) was observed due to fiber embrittlement by oxidation, load transfer from the matrix to the fibers,⁴⁸ and borosilicate formation, as discussed in the Microstructural analysis section.

Table 4.

Residual strength test results for SiC/SiNC CMC.

Stress level (%)	Ultimate tensile strength (MPa)	Elastic modulus (GPa)	Strain-to-failure (%)	Yield stress (MPa)
20	319.1	111.9	0.468	249.2
30	305.2	116.6	0.340	263.1
40	351.7	120.8	0.358	273.4
50	340.3	114.3	0.319	289.0

C/SiNC

The creep-fatigue test results of C/SiNC at various temperatures are shown in Figure 7. The consistent failure of C/SiNC samples, as compared to the SiC/SiNC, is attributed to the higher matrix microcracks density, which results in increased levels of diffusion-limited oxidation.⁴⁹ The matrix is the primary load-bearing constituent and sustains the most damage.

Under creep-fatigue loading, the applied stress significantly reduces the microcrack closure effects mentioned in the Quasi-static tests subsection, and the oxidation regime is similar to the oxidation of T300 carbon fiber alone.³ Additionally, due to the difference in oxidation activation energies, the growth of silica scale resulting from matrix oxidation is slower than the volatilization of carbon fibers,^3,50 causing the C/SiNC CMCs to perform relatively poorly in creep-fatigue testing compared to the SiC/SiNC CMCs. The effects of varying hold-times on material behavior were also studied. A shorter hold-time resulted in a slight increase in elongation and an extended lifetime of the composite sample. Finally, the level of strain recovery was observed to be proportional to the applied stress level, as shown in Figure 7. The elastic strain, developed primarily during the loading phase, with a small fraction accumulating during the holding time due to mismatch in constituent elastic properties, is recovered upon unloading.

Microstructural analysis

The test samples were characterized to understand the effects of manufacturing-induced defects on CMC response. Imaging techniques were used to investigate the correlation of the crack propagation with the microstructural features observed on the fracture surface of the sample. Figure 8 presents SEM images of a SiC/SiNC sample after quasi-static testing at 800°C, illustrating several defects, including intertow porosity, which occurs between fiber bundles (Figure 8(b)), as well as matrix microcracks and intratow voids within individual fiber bundles (Figure 8(c)). These microstructural imperfections compromise the material’s integrity by providing pathways for crack initiation and growth, ultimately reducing mechanical performance. Understanding the nature and distribution of these defects provides valuable insights into the damage and failure mechanisms of the composite material.

Fracture surface characterization of SiC/SiNC

A confocal laser scanning microscope was used to capture micrographs of the fracture surfaces of the room temperature and 800°C quasi-static samples. The micrographs of the room temperature and 800°C samples are compared in Figure 9. It is evident from Figure 9(a) that there is a significant amount of fiber pullout on the fracture surface of the room-temperature sample. Fiber pullout indicates a pseudo-ductile failure, as crack deflection is more prominent when a propagating crack encounters the BN coating. This suggests that the fracture behavior of the 800°C sample is much more brittle than that of the room temperature sample shown in Figure 9(b), which is expected due to the effects of high temperatures on the Hi-Nicalon^TM fiber strength, as discussed in the Results section.

Quasi-static testing resulted in microcracks propagating from the matrix to the fibers since the matrix is the primary load-bearing constituent. Consequently, fiber fractures were observed in both the 0° (Figure 10(a),(c)–(e)) and 90° orientations (Figure 10(b)), indicating that the failure of transversely loaded fibers leads to stress redistribution to the longitudinally loaded fibers. At elevated temperatures, the BN interphase exhibited a brittle fracture mechanism. This is evidenced by the fragments of interphase left behind on the fiber surface after fracture, as illustrated in Figure 10. The brittle nature of the interphase fracture at high temperatures suggests that its ability to absorb energy and deform plastically is significantly reduced under elevated temperature conditions. Understanding these fracture behaviors is critical for predicting the performance and reliability of the composite material in high-temperature applications.

After quasi-static testing, the Hi-Nicalon^TM fiber fracture surfaces are almost flat, strongly indicating a brittle fracture mechanism. These flat fracture surfaces are referred to as mirror regions, characterized by their smooth and semi-reflective appearance^34,51 as depicted in Figures 10(d) and (e). In contrast, the flat fiber fracture surfaces that exhibit some raised points and slight unevenness are known as hackle regions³⁶ as seen in Figure 10(a)–(c). These distinctions between mirror and hackle regions provide valuable insights into which sample areas experienced stable crack growth versus those that underwent a fast fracture.⁵¹

After the residual strength testing, a detailed fracture surface analysis was performed. It was observed that the combination of the BN interphase and the SiNC matrix underwent oxidation, transforming into glassy phases of borosilicate and silica, which effectively fused the fibers, leading to embrittlement as seen in Figure 11. Embrittlement is a result of the strong bonding between the fibers and the borosilicate phase, which restricts the fibers’ ability to deflect a propagating crack by debonding from the interphase, thereby reducing the overall toughness of the composite.⁵² Additionally, a closer examination of the fiber fracture surfaces, as shown in Figure 12, revealed the presence of BN fragments and fragments of the glassy silicate phase. These observations indicate that the oxidation process not only affects the matrix but also significantly impacts the integrity of the fiber/matrix interface and the BN interphase, contributing to the degradation of the composite’s mechanical properties.

A visual comparison between the BN coating and SiNC matrix after quasi-static testing at room temperature and the borosilicate phase formed due to oxidation after creep-fatigue testing at 800°C is presented in Figure 12. Under room temperature conditions, oxidation does not occur, and the fractured segment of the BN coating cleanly debonds from the fibers and displays a matte appearance, as seen in Figure 12(a). In contrast, the borosilicate phase that forms is bonded to the fibers and exhibits a glassy appearance, as shown in Figure 12(b), after exposure to high temperatures and oxidative conditions during creep-fatigue testing. Severe matrix cracking due to creep damage is also depicted in Figure 12(b).⁵¹ This difference in appearance and bonding behavior highlights the significant impact of oxidative environments on the damage mechanisms of the composite materials.

EDS analysis was conducted on an untested, undamaged sample and on a sample that had undergone creep-fatigue at 800°C, followed by residual strength testing, as shown in Figure 13. EDS allowed for the determination of fracture surfaces’ chemical composition. The weight percentages of the undamaged SiNC matrix were as follows: Si (43.7%), C (27.7%), B (19.4%), N (6.7%), and O (2.5%). In contrast, the oxidized SiNC matrix exhibited the following weight percentages: Si (37.0%), O (33.0%), C (15.6%), and N (14.4%). These results reveal a slight decrease in silicon content and a significant increase in oxygen content, indicating the formation of silica. Additionally, the volatilization of carbon at elevated temperatures during matrix oxidation resulted in a slight reduction in carbon content.

Fracture surface characterization of C/SiNC

Following the quasi-static testing of the C/SiNC samples at 1200°C, fractography was conducted. The results of this analysis are shown in Figure 14. Fiber bridging is observed in Figure 14(a)–(c), indicating load-sharing between the constituents. The voids resulting from fiber pullout and fiber oxidation are observed in Figures 14(c) and (d). The fibers that were pulled out and exposed to the oxidative environment are observed in Figures 14(e) and (f), indicating a more ductile fracture mechanism. This is attributed to the fact that oxidation of the matrix does not occur during the short duration of quasi-static tests, preventing fiber fusion and subsequent embrittlement.

Following the creep-fatigue testing at 900°C, the fracture surfaces of the samples were analyzed. Similar to the SiC/SiNC samples, silica formation was observed in the C/SiNC samples. However, as shown in Figure 15, fiber oxidation was observed in the post-creep fatigue C/SiNC sample. This is indicative of the volatilization of carbon,⁵³ resulting in the formation of voids on the surface of the fibers, as depicted in Figure 15(a). These voids significantly compromised the structural integrity of the fibers, facilitating crack propagation, as seen in Figure 15(b), and contributing to the overall mechanical property degradation of the C/SiNC composite under creep-fatigue conditions. Additionally, the lateral surface areas of the exposed carbon fibers are oxidized first, resulting in a conical shape, as seen in both Figures 14 and 15, which agrees with the literature.⁵⁴

Conclusion

This study presented a detailed analysis of the thermomechanical response and damage mechanisms of C/SiNC and SiC/SiNC CMCs, manufactured through PIP, under quasi-static and cyclic loading with varying hold times at intermediate and high temperatures. Quasi-static tests were performed to determine the tensile behavior and mechanical properties, and creep-fatigue tests were conducted at intermediate and high temperatures to simulate typical operating conditions. Residual tensile strength tests were then performed to further assess material degradation, and lastly, fracture surface characterization using imaging techniques such as SEM and EDS was carried out to analyze the failure mechanisms in detail. Following are the important observations from this study:

• Quasi-static tests from room temperature to 1200°C revealed an approximately 30% reduction in tensile strength at elevated temperatures for both CMC systems, with strain-to-failure decreasing by 17% in SiC/SiNC and 31% in C/SiNC.

• Failure mechanisms in quasi-static tests at elevated temperatures are governed by chemical reactions of non-stoichiometric phases in SiC/SiNC, as well as thermal property mismatch and PyC interphase volatilization in C/SiNC.

• Manufacturing-induced defects, such as intertow porosity, intratow voids, and matrix microcracks, were identified as critical factors influencing mechanical properties. These defects act as stress concentrators, initiating crack propagation and compromising material integrity, while also providing oxygen ingress channels into the microstructure.

• Residual strength testing of SiC/SiNC after 100 hours of creep-fatigue loading in air showed a reduction in UTS of up to 37% and a decrease in strain-to-failure of up to 41% due to oxidation-induced damage. Additionally, yield strength increased by 16% when the applied stress level was raised from 20% to 50% of the UTS, attributed to compressive stresses retained in the matrix.

• Passive oxidation in SiC/SiNC CMCs at intermediate and high temperatures leads to matrix silica formation, followed by BN interphase oxidation and borosilicate formation, inducing fiber/matrix fusion and subsequent embrittlement.

• C/SiNC exhibited rapid fracture in under 20 hours of creep-fatigue testing due to higher matrix crack density and more pronounced diffusion-limited oxidation compared to SiC/SiNC.

• Strain recovery was found to be proportional to the applied stress level, with lower hold times during creep-fatigue testing contributing to greater elongation and longer life of C/SiNC.

• C/SiNC failure under creep-fatigue is driven by the volatilization of the PyC interphase and carbon fiber, resulting in void formation on fiber surfaces, thereby compromising load-sharing capabilities.

While exact changes in thermomechanical behavior cannot be precisely predicted due to inherent material variability and the limited availability of CMC specimens for academic research, this study provides an overall understanding of damage mechanisms in SiC/SiNC and C/SiNC composites under various testing conditions, which is crucial for their application in high-temperature environments. These findings clarify the relationships between microstructural mechanisms and mechanical properties, offering valuable insights for the design and development of CMC components. To enhance understanding of CMC behavior under different environmental conditions, future work could focus on further analyzing the oxidation behavior of these materials, including the chemical and physical changes that may affect their mechanical behavior under creep-fatigue loading.

Footnotes

Acknowledgments

This research was sponsored by the Army Research Office under grant number W911NF2220237, with program manager Denise Ford, and the Department of Energy under grant number DEFE0031759, with program manager Matthew Adams. This manuscript was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Army Research Office (grant number W911NF2220237) and the Department of Energy (grant number DEFE0031759).

ORCID iD

Mohamed H Hamza

Data Availability Statement

Data will be made available on request.*

References

Padture

. Advanced structural ceramics in aerospace propulsion. Nat Mater 2016; 15: 804–809.

Binner

Porter

Baker

, et al. Selection, processing, properties and applications of ultra-high temperature ceramic matrix composites, UHTCMCs – a review. Int Mater Rev 2020; 65: 389–444.

Halbig

McGuffin‐Cawley

Eckel

, et al. Oxidation kinetics and stress effects for the oxidation of continuous carbon fibers within a microcracked C/SiC ceramic matrix composite. J Am Ceram Soc 2008; 91: 519–526.

Bhatt

Kiser

. Creep behavior and failure mechanisms of CVI and PIP SiC/SiC composites at temperatures to 1650 °C in air. J Eur Ceram Soc 2021; 41: 6196–6206.

Rubio

Ramanujam

Binner

. Ultra-high temperature ceramic composite. Adv Appl Ceram 2018; 117: s56–s61.

Arai

Inoue

Goto

, et al. Carbon fiber reinforced ultra-high temperature ceramic matrix composites: a review. Ceram Int 2019; 45: 14481–14489.

Cheng

Zhang

, et al. Advances in ultra-high temperature ceramics, composites, and coatings. J Adv Ceram 2022; 11: 1–56.

Christin

. Design, fabrication, and application of thermostructural composites (TSC) like C/C, C/SiC, and SiC/SiC composites. Adv Eng Mater 2002; 4: 903–912.

Pryce

Smith

. Matrix cracking in crossply ceramic matrix composites under quasi-static and cyclic loading. Acta Metall Mater 1994; 42: 861–870.

10.

Almansour

Maillet

Ramasamy

, et al. Effect of fiber content on single tow SiC minicomposite mechanical and damage properties using acoustic emission. J Eur Ceram Soc 2015; 35: 3389–3399.

11.

Almansour

Morscher

. Tensile creep behavior of SiCf/SiC ceramic matrix minicomposites. J Eur Ceram Soc 2020; 40: 5132–5146.

12.

Khafagy

Datta

Chattopadhyay

. Multiscale characterization and representation of variability in ceramic matrix composites. J Compos Mater 2021; 55: 2431–2441.

13.

Hamza

Chattopadhyay

. Multi deep learning-based stochastic microstructure reconstruction and high-fidelity micromechanics simulation of time-dependent ceramic matrix composite response. Compos Struct 2024; 345: 118360.

14.

Sullivan

. Time-dependent stress rupture strength of Hi-Nicalon fiber-reinforced silicon carbide composites at intermediate temperatures. J Eur Ceram Soc 2016; 36: 1885–1892.

15.

Morscher

GN.

Stress-environmental effects on fiber-reinforced SiC-based composites. Ceramic matrix composites: materials, modeling and technology. Hoboken, NJ: John Wiley & Sons, Ltd, 2014, pp. 334–352.

16.

Morscher

Ojard

Miller

, et al. Tensile creep and fatigue of Sylramic-iBN melt-infiltrated SiC matrix composites: retained properties, damage development, and failure mechanisms. Compos Sci Technol 2008; 68: 3305–3313.

17.

Gowayed

Ojard

Chen

, et al. Accumulation of time-dependent strain during dwell-fatigue experiments of iBN-Sylramic melt infiltrated SiC/SiC composites with and without holes. Compos Part A Appl Sci Manuf 2011; 42: 2020–2027.

18.

Nasiri

Patra

, et al. Oxidation behaviour of SiC/SiC ceramic matrix composites in air. J Eur Ceram Soc 2016; 36: 3293–3302.

19.

Morscher

. Tensile creep and rupture of 2D-woven SiC/SiC composites for high temperature applications. J Eur Ceram Soc 2010; 30: 2209–2221.

20.

Teng

Zhu

Jia

, et al. Research on fatigue failure of SiCf/SiC composites at high temperature based on X-ray computed tomography. Ceram Int 2023; 49: 40541–40550.

21.

Tejero-Martin

Bennett

Hussain

. A review on environmental barrier coatings: history, current state of the art and future developments. J Eur Ceram Soc 2021; 41: 1747–1768.

22.

Eaton

Linsey

. Accelerated oxidation of SiC CMC’s by water vapor and protection via environmental barrier coating approach. J Eur Ceram Soc 2002; 22: 2741–2747.

23.

Lee

KN.

Environmental barrier coatings for SiCf/SiC. Ceramic matrix composites: materials, modeling and technology. Hoboken: John Wiley & Sons, Ltd, 2014, pp. 430–451.

24.

Yang

Chen

, et al. Life assessment of thermomechanical fatigue in a woven SiC/SiC ceramic matrix composite with an environmental barrier coating at elevated temperature. Int J Fatig 2023; 172: 107584.

25.

Ruggles-Wrenn

Sharma

. Effects of steam environment on fatigue behavior of two SiC/[SiC+Si 3N4] ceramic composites at 1300°C. Appl Compos Mater 2011; 18: 385–396.

26.

Sheikh

Hayhurst

Taylor

, et al. Experimental investigation of the effect of mechanical loading on thermal transport in ceramic matrix composites. J Multiscale Model (JMM) 2009; 01: 403–431.

27.

Yang

Zhang

Wang

, et al. Tensile behavior of 2D-C/SiC composites at elevated temperatures: experiment and modeling. J Eur Ceram Soc 2017; 37: 1281–1290.

28.

Sauder

Brusson

Lamon

. Influence of interface characteristics on the mechanical properties of HI-Nicalon type-S or tyranno-SA3 fiber-reinforced SiC/SiC minicomposites. Int J Appl Ceram Technol 2010; 7: 291–303.

29.

Chateau

Gélébart

Bornert

, et al. Modeling of damage in unidirectional ceramic matrix composites and multi-scale experimental validation on third generation SiC/SiC minicomposites. J Mech Phys Solid 2014; 63: 298–319.

30.

Bhatt

Garg

. Thermal stability of Hi-Nicalon SiC fiber in nitrogen and silicon environments . NASA TM-106987.

31.

Skinner

Chattopadhyay

. Multiscale temperature-dependent ceramic matrix composite damage model with thermal residual stresses and manufacturing-induced damage. Compos Struct 2021; 268: 114006.

32.

Artz

. Modeling lifetime performance of ceramic matrix composites with reduced order homogenization multiscale methods. New York, NY: Columbia University, 2022.

33.

Guo

Kagawa

. Tensile fracture behavior of continuous SiC fiber-reinforced SiC matrix composites at elevated temperatures and correlation to in situ constituent properties. J Eur Ceram Soc 2002; 22: 2349–2356.

34.

Bodet

Bourrat

Lamon

, et al. Tensile creep behaviour of a silicon carbide-based fibre with a low oxygen content. J Mater Sci 1995; 30: 661–677.

35.

Yun

Dicarlo

. Thermomechanical behavior of advanced SiC fiber multifilament tows. In: Proceedings of the 20th Annual Conference on Composites, Advanced Ceramics, Materials, and Structures—B: Ceramic Engineering and Science Proceedings. Hoboken, NJ, USA: John Wiley & Sons, Inc., 1996, pp. 61–67.

36.

Pysher

Goretta

Hodder

, et al. Strengths of ceramic fibers at elevated temperatures. J Am Ceram Soc 1989; 72: 284–288.

37.

Lamouroux

Camus

Thébault

. Kinetics and mechanisms of oxidation of 2D woven C/SiC composites: I, experimental approach. J Am Ceram Soc 1994; 77: 2049–2057.

38.

Richards

Young

De Francqueville

, et al. Response of ytterbium disilicate–silicon environmental barrier coatings to thermal cycling in water vapor. Acta Mater 2016; 106: 1–14.

39.

Villeneuve

Naslain

Fourmeaux

, et al. Longitudinal/radial thermal expansion and Poisson ratio of some ceramic fibres as measured by transmission electron microscopy. Compos Sci Technol 1993; 49: 89–103.

40.

Menessier

Dument

Geutte

, et al. Axial and Radial Coefficients of Thermal Expansion of Carbon Fibers in the 20°–430° C Temperature Range as Derived from the Thermal Expansion of 1‐D‐C‐SiO2 (B2O3) Composites. In A Collection of Papers Presented at the 13th Annual Conference on Composites and Advanced Ceramic Materials: Ceramic Engineering and Science Proceedings. Hoboken, NJ, USA: John Wiley & Sons, Inc, 2008, pp. 1426–1439.

41.

Cheng

Zhang

, et al. Oxidation behavior and mechanical properties of C/SiC composites with Si-MoSi2 oxidation protection coating. J Mater Sci 1999; 34: 6009–6014.

42.

Zheng

G-

Sano

Uchiyama

, et al. Effect of boron addition on oxidation resistance of carbon fiber polycarbosilane-derived SiC composites. J Mater Sci Lett 1998; 17: 2047–2049.

43.

Sauder

Lamon

. Tensile creep behavior of SiC-based fibers with a low oxygen content. J Am Ceram Soc 2007; 90: 1146–1156.

44.

Morscher

Pujar

. Creep and stress–strain behavior after creep for SiC fiber reinforced, melt-infiltrated SiC matrix composites. J Am Ceram Soc 2006; 89: 1652–1658.

45.

Zhu

Mizuno

Kagawa

, et al. Creep and fatigue behavior of SiC fiber reinforced SiC composite at high temperatures. Materials Science and Engineering: A 1997; 225: 69–77.

46.

Rugg

Tressler

Bakis

, et al. Creep of SiC–SiC microcomposites. J Eur Ceram Soc 1999; 19: 2285–2296.

47.

Khafagy

Chattopadhyay

. High-fidelity micromechanical modeling of the effects of defects on damage and creep behavior in single tow ceramic matrix composite. J Compos Mater 2023; 57: 377–392.

48.

Zhu

Mizuno

Nagano

, et al. Tensile creep behavior of a SiC-fiber/SiC composite at elevated temperatures. Compos Sci Technol 1998; 57: 1629–1637.

49.

Huang

, et al. Creep behavior of C/SiC composite in hot oxidizing atmosphere and its mechanism. Ceram Int 2017; 43: 9355–9362.

50.

Ogbuji

Opila

. A comparison of the oxidation kinetics of SiC and Si3 N 4. J Electrochem Soc 1995; 142: 925–930.

51.

Simon

Bunsell

. Mechanical and structural characterization of the Nicalon silicon carbide fibre. J Mater Sci 1984; 19: 3649–3657.

52.

Pandey

. Quasi-static and creep behavior of enhanced SIC/SIC ceramic matrix composites. Blacksburg, VA: Virginia Tech, 2000.

53.

Verrilli

Kantzos

Telesman

. Characterization of damage accumulation in a carbon fiber-reinforced silicon carbide ceramic matrix composite (C/SiC) subjected to mechanical loadings at intermediate temperature. West Conshohocken: ASTM Special Technical Publication, 2000, pp. 245–261.

54.

Goulmy

Caty

Rebillat

. Characterization of the oxidation of C/C/SiC composites by Xray micro-tomography. J Eur Ceram Soc 2020; 40: 5120–5131.