Selective growth of SiC nanowires in interlaminar matrix for improving in-plane strengths of laminated Carbon/Carbon composites

1. Introduction

Carbon/carbon (C/C) composites are attractive thermo-structural materials used in aerospace applications due to their low density and desirable high-temperature mechanical properties [[1], [2], [3], [4]]. However, due to the intrinsic anisotropy of pyrocarbon matrix and the low fiber-matrix interface bonding strength, C/C composites typically exhibit unsatisfactory out-of-plane properties [[5], [6], [7], [8]].

Growing one-dimensional (1D) nanomaterials (such as carbon nanotubes (CNTs) [7,9,10], carbon nanofibers (CNF) [11,12] and SiC nanowires (SiCNWs) [[13], [14], [15]]) on carbon fibers is promising in improving the out-of-plane strength of C/C composites. Extensive researches focused on this strategy have achieved remarkable improvements related to the out-of-plane performance of C/C composites. CNTs are used as fillers firstly and out-of-plane strength improvements in the range of 60%-275% can be obtained via this method [7,9,10]. CNFs and SiCNWs are also important nano-fillers and introducing CNFs/SiCNWs generally contributes to an increase of out-of-plane strength in the range of 30%-100% [11,14,15]. In our previous work, we have demonstrated a significant improvement in interlaminar strength of C/C composites by introducing SiCNWs on carbon fibers [15]. However, several negative effects of this strategy emerge as researches go further, which greatly limit their applications as advanced thermal-structural materials. Firstly, high temperature required for the growth of SiCNWs as well as CNTs on carbon fiber surface usually lead to the reduction of fiber strengths [16,17]. Secondly, introducing SiCNWs and CNTs can easily lead to overly enhancing of fiber-matrix interface, which decreases the in-plane fracture toughness and causes brittle fracture of the composites [7,18,19]. As C/C composites are primarily used in aerospace fields where reliability and safety should be specially considered, brittle fracture of C/C components should always be avoided.

Former researches have proved that pyrocarbon coating on carbon fiber is an efficient routine for balancing between fiber-matrix interface bonding and fracture toughness of fiber-reinforced composites [[20], [21], [22]]. Additionally, applying pyrocarbon coating can protect fibers against SiO and CO gas erosion during the growth of SiCNWs at high temperature [23,24]. Therefore, a pyrocarbon coating is essential to tailor the mechanical properties of C/C composites and preserve fiber strength against high temperature erosion. As for SiCNW reinforced C/C composites, how to optimize the thickness of pyrocarbon coating and what is the coupling effect of pyrocarbon coating with in-situ growth SiCNWs on the mechanical performance of C/C composites are two key questions that need to be answered.

Here in this work, we attempted to introduce SiC nanowires (SiCNWs) at interlaminar matrix rather than on fiber surface by applying a pyrocarbon coating on carbon fibers to avoid both fiber strength degradation and brittle fracture of the composites. A detail analysis has been made to investigate the relationship between in-situ fiber strengths and the thickness of the pyrocarbon coatings. Also, the effect of introducing SiCNWs has been characterized and the reinforcing and toughening mechanisms of SiCNWs have been further discussed. It is successfully demonstrated that introducing SiCNWs selectively in interlaminar matrix is a feasible routine to achieve proper reinforcement of fiber-matrix interfaces in C/C composites, which not only avoids the degradation of fiber strength caused by the growth of 1D nanomaterial but also transfers the fracture mode of the overly reinforced composites from brittle fracture to pseudo-plastic fracture.

2. Experimental

2.1. Preparation of SiCNWs reinforced C/C (SiCNW-C/C) composites

The preparation of SiCNW-C/C composites is similar to the method described in our previous work [15] except that a protective pyrocarbon layer is prepared on carbon fiber before growing SiCNWs. The whole process for fabricating SiCNW-C/C composites is illustrated in Fig. 1(a). Firstly, a protective pyrocarbon coating was prepared on PAN-based plain-weave carbon fiber cloth (T300, 1 K, obtained from Toray Inc.) by chemical vapor deposition (CVD) method. Stacked pieces of carbon cloth were placed into a vertical isothermal reactor. The deposition process was carried out at ambient pressure and 1000 °C. CH₄ with a flow rate of 40 L/h and partial pressure of 30 kPa was kept for 5-40 h while N₂ was used as dilution and protective gas during the whole process. After the preparation of pyrocarbon coating, the stacked pieces of carbon cloth were suspended in a graphite crucible. A mixture of SiO₂, Si, and graphite powders (SiO₂: ˜50 wt%, Si: ˜20 wt%, graphite: ˜30 wt%) was placed below them. Then the crucible was placed into a heating furnace and was heated to 1773-1973 K in argon atmosphere and kept for 1.5 h to grow SiC nanowires at the interlaminar space of the stacked pieces. Details about SiCNWs synthesis in this method can be found in Ref [25]. Subsequently, these stacked pieces were placed in the vertical isothermal reactor again to accomplish the densification process. The deposition conditions remained the same as described above and the whole process lasted for 140 h. Additionally, SiCNW-C/C composites without pre-deposited pyrocarbon coating are prepared, which is marked as SiCNW-0-C/C, for the comparison analysis of mechanical behaviors.

3D model of the as-prepared composites (created by TexGen software) is shown in Fig. 1(b) for defining the viewpoints in the following contents. As for comparison, C/C composites without introducing SiCNWs were also prepared under the same infiltration conditions.

2.2. Mechanical tests

Interlaminar shear strengths and in-plane flexural strengths of SiCNW-C/C composites were measured using a CMT5304-30KN testing machine.

For the interlaminar shear tests, the samples were machined into the size of 4 mm × 6 mm × 3 mm. The experiments were carried out using the interlaminar shear device (ISD) designed by Rossell and Santare [26]. Fig. 1(c) shows a brief illustration of the interlaminar shear test device. Shear tests were carried out at a loading rate of 0.5 mm/min. The interlaminar shear strength (ILSS) was calculated based on the following formula:

(1)$τ=\frac{P}{A_{0}}$

In this formula, τ is the ILSS of the sample, P is the maximum failure load and A0 is the total area of the shear fracture. At least five samples were tested for each type of composites.

Flexural properties of the C/C composites specimens were measured by three-point bending test. The tests were performed under standard ASTM C1341-13. The samples were all machined to strips of 30 mm × 5 mm × 2 mm. Fig. 1(d) gives a brief illustration of the flexural test process. Three-point bending tests were carried out at a loading rate of 0.5 mm/min with a span of 20 mm. The flexural strength was calculated based on the following formula:

(2)$δ=\frac{3PL}{2bh^{2}}$

Here, δ is the flexural strength of the material; P is the maximum load; L is the span during the testing and b, h are the width and thickness of the specimens, respectively.

2.3. Characterization methods

Surface and fracture morphology images of the samples were obtained using a FEI NANOSEM450 field emission scanning electron microscope. Transmission electronic microscopy (TEM) images and selected area electron diffraction (SAED) of the SiCNWs were obtained under a Tecanai F30 G² high-resolution TEM. Cross section Polarized optical photos were applied using a Leica DMLP Polarized optical microscopy (PLM) to characterize the texture of pyrocarbon matrix. Raman Spectrum of the matrix was obtained using an inVia Reflex Raman spectroscopy. Porosity of the specimens was measured with a mercury porosimeter (AutoPore IV 9500).

3. Results and discussion

3.1. Deposition of pyrocarbon and growth of SiCNWs on carbon cloth

To realize moderate enhancement of fiber-matrix interface, it is essential to disperse SiCNWs in the matrix, which avoids the direct bonding between SiCNWs and the fibers. One of the common and efficient routine to solve this problem is to pre-deposit coatings on fiber surface, among which pyrocarbon coating should be a preferential choice for its high compatibility with carbon fibers at the whole temperature range. Fig. 2(a) shows the surface SEM image of carbon fibers with pyrocarbon coating. From the cross-section SEM image (insert map in Fig. 2(a)) the thickness of pyrocarbon coating can be estimated, which varies from ˜500 to ˜4000 nm.

After the pre-depositing of pyrocarbon coating, SiCNWs were synthesized on the fiber cloth via chemical vapor deposition as described in Section 2.1. The volumetric content of SiCNWs in SiCNW-C/C composite is approximately 0.65%. Fig. 2(b) shows the morphology of as-synthesized SiCNWs. SiCNWs are found randomly distributed on the carbon cloth. The diameters of SiC nanowires vary from ˜100 nm to ˜300 nm. Compared with SiCNWs grafting on carbon fiber by the sizing method, in-situ growing SiCNWs shows higher density but with larger diameter [14]. No catalyst was applied in the whole growth process, so it is inferred that the growth of the SiCNWs was dominated by vapor-solid (VS) mechanism [25]. TEM image of the SiCNWs (Fig. 2(c)) reveals that there are some dentations at the surface of nanowires. Higher resolution TEM (HRTEM) images (Fig. 2(d)) along with the SEAD patterns further indicates the existence of stacking faults perpendicular to axial direction ([111] direction of β-SiC, Pattern Reference Code: 00-002-1050) of the SiCNWs [27]. It can be inferred that dislocations of these stacking faults cause the dentations on the surface. Such surface dentations are beneficial for the nanoscale mechanical interlocking between the nanowires and matrix, which will improves the load transfer efficiency from matrix to nanowires [28].

3.2. Optimization of pyrocarbon coating thickness

The thickness of pyrocarbon coating is a key parameter to tailor the fiber-matrix bonding states, which greatly influences the mechanical properties of C/C composites. As is known to us, thicker pyrocarbon coating contributes to weaker fiber-matrix bonding. A desirable fiber-matrix interface should be both effective in transferring the load from matrix to fibers and efficient in allowing fiber-matrix debonding [29,30]. Specially, for SiCNW-C/C composites, the protection of fiber against strength degradation during the growth of SiCNWs is another important aspect that should be considered for determine the thickness of pyrocarbon coating. To determine the optimized coating thickness, it is essential to investigate the in-situ fiber strengths after the preparation of coatings and the growth of SiCNWs, which can be obtained via the fracture mirror method [31,32]. Fig. 3 shows the in-situ fiber strengths with different coating thickness. The R² values (the coefficient of determination, which is a statistic parameter to measure the goodness-of-fit) for the fitting lines are 0.9053, 0.7104, 0.9255, 0.8543 and 0.8265, respectively. This means that these straight lines fit the data quite well and these data follows the Weibull distribution. The shape (k) and scale (λ) parameters of the strength distributions can be obtained from these Weibull diagrams (Fig. 3).

The mean fiber strength δf can be obtained via the follow equation:

(3)$δ_{f}=λ*Г(1+\frac{1}{k})$

The influence of pyrocarbon thickness on mean fiber strength are shown in Fig. 4. The fibers without coating protection show the lowest in-situ fiber strength and as the thickness of pyrocarbon increases, the in-situ fiber strengths increase slowly. The in-situ fiber strength increases by 6.9% when a 0.5 μm coating is applied. When the coating thickness increases from 0.5 μm to 2.0 μm, the in-situ fiber strengths show slight increase. Another significant increase of fiber strength (20%) can be observed when the coating thickness increased to 4.0 μm.

For choosing the optimized thickness of pyrocarbon coating for C/C composites, it should be noticed that the increase of coating thickness results in the increase of interlaminar gap and decrease of fiber-matrix interface bonding and interlaminar shearing resistance for laminated C/C composites. Therefore, although a coating 4 μm in thickness shows the best protection effect, it is not proper for applying in laminated C/C composites. When the coating thickness varies from 0.5 to 2.0 μm, the in-situ fiber strengths show slight improvement. Besides, as the fiber-matrix interface in C/C composites are proved to be weak, thinner pyrocarbon coating will contributes to better interface enhancement besides less time and commercial costs. Therefore, a coating thickness of 0.5 μm is preferred here. Subsequently, the sample with pyrocarbon coating 0.5 μm in thickness and the growth of SiCNWs is densified via the CVI process together with the referenced C/C composites.

3.3. Field emission property

The final density of C/C composites is 1.67 ± 0.2 g/cm³ and the density of SiCNW-C/C composites is 1.69 ± 0.2 g/cm³. The porosity of C/C composites is 12.069% and the porosity of SiCNW-C/C composites is 6.6423% according to mercury porosimetry measurement. The microstructure of pyrocarbon matrix plays a critical role in the mechanical performances of C/C composites [[33], [34], [35]]. Optical observations via polarized light is the regular method to characterize pyrocarbon matrix. After introducing SiCNWs on the carbon fiber, the gaps between the neighboring layers increased by 23% statistically. Due to the anisotropy of pyrocarbon, the orthogonal branches of polarized light extinct when rotating the cross section of the pyrocarbon [36,37]. For C/C composites, the pyrocarbon at the interlaminar space is relatively dense and smooth under polarized light as observed in Fig. 5(a).

For SiCNW-C/C composites as shown in Fig. 5(b), a transition layer about 20 μm in thickness exists, which is neighboring with the fiber cloth. Many discrete spots or strips can be found in this transition layer. Chemical states of the original fiber surface cannot be preserved after the growth of this transition layer. Therefore, surface defects caused by SiCNWs growth should not be the nucleus for the growth of these spots and strips for the interlaminar pyrocarbon matrix. So it can be speculated that these spots and strips originate from the nanowires tips and most of these cones are in irregular shapes compared with the growth cones reported in previous literatures [36,38]. Based on these PL and SEM observations, it can be found that the crystallite orientations of pyrocarbon matrix change greatly by introducing SiCNWs into the matrix (Fig. 5(d)).

Raman spectra have been collected to further analyze the as-synthesized pyrocarbon. For pyrocarbon, the main peaks in the Raman spectra are the G (at ˜1560 cm^-1) and D peaks (at ˜1360 cm^-1) respectively under visible excitation [39]. The relative intensity ratio of the D-band to the G-band (I_D/I_G) is inversely proportional to the crystallite size of pyrocarbon. In this case, I_D/I_G is 1.41 for SiCNW-C/C composites and 1.00 for C/C composites as shown in Fig. 5(e). As is known to us, the I_D/I_G is inverse proportion to L_a of the grapheme layers, where L_a refers to the crystallite size of each grapheme layer in pyrocarbon [40]. Therefore, it can be inferred that after introducing SiCNWs in the matrix, the crystalline size of pyrocarbon decreased. It should be noticed that results reported by Chen et al. [14] provides a contrary conclusion compared with the results here. Their work shows that introducing SiCNWs induces larger L_a of the pyrocarbon. The difference between the present work and their work is that the density of SiCNWs is much higher by the in-situ CVD method than the sizing method. Higher density of SiCNWs contributes to more nucleate sites and helps with the nucleating process, which restrains the growth of pyrocarbon crystallite. This phenomenon is consistent with our previous work that higher density of CNTs can decrease both the crystallite and anisotropy of pyrocarbon [41]. During the CVI process, CH₄ molecules react with each other and form into polycyclic aromatic hydrocarbon, which subsequently nucleates at defects on the substrate. The introducing of SiCNWs provides additional nucleation sites for the growth of pyrocarbon. Also, as pyrocarbon grows radially on surface of fibers and SiCNWs, the orientation of the SiCNWs determines the orientation of the surrounded pyrocarbon (Fig. 5(b & d)). Combined with the PLM results, it can be concluded that the incorporation of SiCNWs in the matrix not only changes the crystallite orientation but also refines the crystallite size of the matrix.

After introducing SiC nanowires, the fiber volume fraction decreased from ˜33% to ˜30% and the porosity decreased from 12.069% to 6.6423% measured by the mercury porosimetry. Meanwhile, it can be observed from the porosity distribution diagram (Fig. 5(f)) that after introducing SiC nanowires, the fraction of pores with diameter ranging from 10 μm to 100 μm decreased obviously while pores with diameter ranging from 2 μm to 6 μm increased. Therefore, introducing SiC nanowires helps decrease the presence of large pores but results in more small pores.

3.4. Mechanical testing results

To investigate the effect of SiCNWs on the interlaminar strength and shear bending strength of C/C composites, both in-plane shear tests and out-of-plane bending tests have been performed. The interlaminar shear test results of SiCNW-C/C and C/C composites are shown in Fig. 6 and the out-of-plane bending results are shown in Fig. 7. Under the testing conditions shown in Section 2.2, average shear and bending strengths of SiCNW-C/C were 10.06-1.27+1.08 MPa and 162.44-16.07+12.65MPa, respectively. In comparison, the average shear and bending strengths of C/C were 3.90-0.80+1.72 MPa and 103.24-16.11+13.77 MPa , respectively. After introducing SiCNWs, average shear strength and bending strength increased by 158% and 57%. Additionally, it should be noticed that the flexural strength of SiCNW-0-C/C (92.40-26.21+20.65 MPa) decreased slightly compared with the original C/C composites, which is caused by the degradation of fiber strength by the in-situ growth of SiCNWs on the fiber and subsequent overly enhanced fiber-matrix interfaces. Compared the mechanical performance of SiCNW-0-C/C and SiCNW-C/C, it is obvious that the introducing of a proper pyrocarbon coating is efficient in relieving the side effects of in-situ growth of SiCNWs.

The shear fracture behaviors of SiCNW-C/C and C/C composites are similar with each other. Typical examples of load-displacement curves are shown in Fig. 6(a). These curves indicate the brittle fracture mode of both SiCNW-C/C and C/C composites under interlaminar shear stress.

However, the bending fracture behaviors of SiCNW-C/C and C/C composites show some significant differences, as demonstrated in Fig. 7(a). For both SiCNW-C/C and C/C composites, their flexural fracture can be ascribed to pseudoplastic fracture [42]. Both their fracture curves can be divided into three periods: elastic deformation, yielding and final failure. However, the yield stages of SiCNW-C/C and C/C composites are quite different observed from Fig. 7(a). For SiCNW-C/C composites, the load decreases when the curve reaches the yield point, followed by a fluctuation stage without further increase of the load. Meanwhile, for C/C composites, the load continues to increase after reaching the yield point until the final fracture happens. From this phenomenon, it can be inferred that for these two types of composites, the fracture of load-bearing fibers occurs at different time. For SiCNW-C/C composites, no further increase of load after the yield point indicates the fracture of fibers at yield point. Meanwhile, further increase of load after yield point for C/C composites indicates that fibers did not fracture at the yield point until the final fracture of the sample. Thus, there should be two different energy dissipation approaches during the yield stage for SiCNW-C/C and C/C composites. If we define the point where the load reaches its maximum point as final failure, it can be inferred from Fig. 7(a) that: (1) For C/C composites, delamination and fiber-matrix debonding are two dominated energy dissipation aspects; (2) For SiCNW-C/C composites, the nonlinear part of the load-displacement curve in Fig. 7(a) originates from the fracture of the fibers and subsequent pulling-out of the fiber, indicating that fibers fracture and frictional sliding of the fibers are the primary aspects for energy dissipation.

To better understand the fracture processes, fracture observations are performed, and the detail results and fracture mechanism discussions are shown below.

3.5. Shearing fracture comparison of C/C and SiCNW-C/C composites

Due to the intrinsic chemical stability of carbon fiber and pyrocarbon, the bonding between carbon fiber and pyrocarbon matrix is relatively weak. Besides, the intrinsic anisotropy of pyrocarbon results in the failure of matrix under low levels of shear stress. Both factors mentioned above lead to the relatively low interlaminar strength of C/C composites.

Fig. 8 shows the typical interlaminar shear fracture of C/C composites. From Fig. 8 it can be seen that some debris from the neighbor layer attaches on the fracture surface and the interlaminar bonding seems to be quite inefficient. A closer investigation of the debris provides more information of the shear fracture process (Fig. 8(b)). Two main reasons are responsible for the failure of interlaminar bonding in C/C composites here: the failure of fiber-matrix bonding (area 1) and the inefficient bonding of pyrocarbon in neighbor layers, referred to as the interlaminar interface (area 2). The first reason has been mentioned frequently in previous researches [43] and is the key problem that need to be solved. However, there has been no special attention paid for the second reason so far. Different from carbon fiber reinforced polymer composites, when preparing C/C composites, pyrocarbon matrix grows around the fibers layer by layer to form a shell structure similar to multi-walled carbon nanotubes. The bonding of each layers and the interlaminar interface are predominated by Van der Waals force. The shear fracture (Fig. 8(c)) indicates that the bonding of pyrocarbon layers are quite weak as no efficient chemical bonding exists. Thus, to improving the shear resistance ability of C/C composites, both the fiber-matrix interface and the interlaminar interface of neighbor layers should be enhanced simultaneously.

For SiCNW-C/C composites, the shear fracture shows some distinctive features compared with that of C/C composites (Fig. 9). Most of the fracture debris from the neighbor layer still attaches to the fracture after shear fracture (Fig. 9(a)). Further explorations indicate that the debris is the matrix pyrocarbon and the pre-deposited coating (Fig. 9(b)). No exposed fibers are observed (Fig. 9(c) & (d)), which indicates that the fracture originates at the coating-matrix interface but not at the matrix or the fiber-matrix interface, which is different with SiCNW-C/C without pyrocarbon coating [14,15]. Besides, some roots of the SiCNWs and debris of pyrocarbon can be observed on the fracture surface while the others still sticks on the neighboring layer, indicating efficient chemical bindings are realized at the coating-matrix interface. Compared with the shear fracture of C/C composites, fiber-matrix debonding and interlaminar debonding are both diminished. Therefore, the fiber-matrix bonding and interlaminar bonding have been enhanced by introducing SiCNWs, which is the main reason for the improved interlaminar shear strength.

3.6. Flexural fracture comparison of C/C and SiCNW-C/C composites

For laminated C/C composites, both interlaminar shear stress and in-plane tensile stress exist under flexural strength. Thus, they fail either by delamination or in-plane fracture of fibers. The flexural fracture of C/C composites shows some distinctive delamination features as shown in Fig. 10. Some gaps parallel to the interlaminar plane appear after three-point bending tests. The fiber bundles with the surrounding pyrocarbon fracture simultaneously. These gaps originate from the relatively weak bonding sections described in Fig. 8(c). Besides, the fracture of individual fiber bundles shows some brittle fracture features, which means that the fracture of individual bundles should not be the reason for the pseudoplastic fracture behavior of the C/C composites as shown in Fig. 7(a). Based on this observation, it can be speculated that the pseudoplastic fracture is the result of interlaminar pyrocarbon debonding observed in Fig. 10(a) & (c). Therefore, the whole fracture process should include the following process: (1) when the load reaches its critical level (the yield point), pyrocarbon in neighbor layers debonds with each other firstly; (2) as the displacement further increases, final fracture of the sample happens. In stage (1), as the debonding happens, the whole sample transfers into several individual laminates that bear the loads. As no fracture of individual layer occurs, these layers can still bear the load efficiently. So, a further increase of load can be observed in the load-displacement curves. The fracture energy produced in this stage is consumed by the interlaminar debonding process.

For SiCNW-C/C composites, the fracture shows no obvious interlaminar cracks after bending tests (Fig. 11(a)). Additionally, large amount of pulled-out fibers are observed at the fracture (Fig. 11(b)), which is different from C/C composites and is the reason for its pseudoplastic fracture behavior. As no obvious delamination is observed, it can be speculated that the yield stage of SiCNW-C/C composites originates from the pulling-out of fibers [42]. The stress where yield point for SiCNW-C/C composites happened is much higher than that of C/C composites. The cross section of pyrocarbon fracture is quite smooth and the boundary of interlaminar interface diminishes (Fig. 11(c) & (d)). Some pre-deposited coating still sticks on the pulled-out fiber surface while the other parts of the fiber surface are exposed (Fig. 11(e)), indicating that the fracture are resulted from the failure of fiber-matrix interface and the relative weak bonding area of the coating and matrix. This phenomenon is quite different from the fracture process of C/C composites under flexural stress. After the yield point, cracks propagate mainly along the fiber-matrix interface for SiCNW-C/C composites and along the interlaminar interface for C/C composites. Therefore, SiCNW-C/C composites show enhanced interlaminar strength under flexural stress and this is the dominant reason for its improved flexural performance compared with C/C composites.

Higher magnification images of the fracture surface (Fig. 11(f-h)) conform that introducing SiCNWs plays the critical role in enhancing the interlaminar strength. Two different types of fracture can be detected: (1) matrix that contains sufficient SiCNWs and (2) matrix that contains less SiCNWs.

In Section 1, two main features should be noticed. On one hand, although the fracture surface of matrix is relatively smooth, which indicates that cracks propagate directly to carbon fibers, pulling out of carbon fibers still happens. A fragmented pre-deposited pyrocarbon layer can be detected on the surface of pulled out fibers, indicating the moderate bonding strength of fiber-matrix interface. When the tips of crack propagate to the fibers, this moderate bonding interface debonds and prevents the direct failure of fibers against the crack tip stress (Fig. 11(d)). On the other hand, efficient amount of SiCNWs embedding in the matrix and the pyrocarbon around them are in amorphous state. The fracture is relatively smooth while large amounts of SiCNWs pulling out can be observed. Introducing SiCNWs benefits the improvement of critical bending stress that cause the fracture of pyrocarbon matrix. Besides, the pulling-out of SiCNWs and the fracture of SiCNWs themselves also consume additional energy during the fracture process. Similar phenomenon has been reported for CNT reinforced pyrocarbon matrix in prior related works: when CNTs are sufficient the predicted crack deflection function of CNTs is inhibited and the CNTs mainly reinforcing the matrix [41,44]. Thus, for the case that SiCNWs are sufficient in matrix, SiCNWs helps reinforcing the matrix and increasing the critical failure stress of the matrix while the pre-deposited pyrocarbon interface helps with the deflection of cracks.

Compared with Section 1, the effect of SiCNWs in Section 2 is quite different. As the number of SiCNWs is less than that in Section 1, the reinforcing effect of SiCNWs is quite limited according to the rule of mixture [45]. However, one should notice that SiCNWs change the texture orientation of pyrocarbon matrix from “along the circumference of carbon fibers as shown in pure C/C” to “a wavy texture” that is formed by the nucleation and growth of pyrocarbon around each single SiCNW (Fig. 11(g) & (h)). This results in the formation of cone-shaped pyrocarbon and lots of defects (mainly including the boundaries between the neighbored cone-shaped pyrocarbon) inside the matrix. When cracks propagate to section (2), these defects can easily deflect the cracks from propagating towards the fibers. One can observe that large amounts of matrix still attaches to the pulling-out fibers and no crack can propagate directly the fibers (Fig. 11(g)). This is similar with the stepwise cracks observed by previous works [14,21,44], which is beneficial for consuming more energy during the fracture. Therefore, SiCNWs in this section mainly helps toughening the matrix rather than reinforcing the matrix.

For C/C composites and similar materials, once the local stress around defects is above the critical level, cracks will breed from these defects and propagate along the path that consumes the least energy. As demonstrated in 3.3, introducing SiCNWs refines the pyrocarbon crystallite, which subsequently weaken the stress concentration at defects inside the pyrocarbon matrix [46]. Therefore, it helps improving the flexural stress threshold above which destructive cracks will generate inside the matrix.

When cracks generate at the defect nucleation under flexural stress, they propagate along the weakest interfaces (fiber-matrix interface and interlaminar interface) before the final failure under flexural stress. Several previous studies attempt to introduce 1D nanomaterials to enhance the fiber-matrix interfaces, which actually cause the excessive enhance of the interface and provide additional paths for cracks propagate towards the fibers [7,18,19]. Therefore, fibers break easily when crack tips reach the fibers where the fiber-matrix interface is too strong, which causes the brittle fracture and decrease of in-plane strength. To overcome this side-effect, interfaces should be enhanced moderately while the crack deflection function of the interfaces can still be preserved [47,48]. By comparing the fracture behaviors of SiCNW-C/C and C/C composites, we can find that interlaminar interfaces are enhanced due to the introducing of SiCNWs while the cracks can still be deflected from propagating to the fibers for SiCNW-C/C composites. Thus, the moderate enhancement of interface can be realized by introducing SiCNWs in interlaminar matrix as shown above.

4. Conclusion

SiCNWs were successfully introduced in the interlaminar matrix of C/C composites by chemical vapor deposition. SiCNWs in the matrix not only change the crystallite orientation but also refine the crystallite size of pyrocarbon. Moderate enhancement of interfaces in C/C composites has been realized by this method. Compared with C/C composites, the interlaminar shear strength and three-point flexural strength of SiCNW-C/C composites increase by 158% and 57%, respectively. Introducing SiCNWs is efficient in overcoming the delamination of C/C composites under flexural stress. When SiCNWs are sufficient in the matrix, they help reinforcing and improving the critical failure stress of the matrix. When their content is relatively small, SiCNWs help toughening the matrix. The brittle fracture caused by the introducing of nanomaterials on carbon fiber surface has been overcome by this method, which is beneficial for improving the reliability and safety of C/C components for aerospace applications.

Acknowledgement

This work has been supported by the National Natural Science Foundation of China under Grant Nos. 51502242, 51432008, U1435202, and the Fundamental Research Funds for the Central Universities (3102016ZY009).

---