The combined effect of temperature and seawater on the compression properties of carbon fiber vinyl ester composites for sandwich structures

Abstract

For naval applications, composite sandwich structures are of significant interest, and are often manufactured using thin (2 to 4 mm) composite facings made from carbon or glass fiber reinforcement, attached to a thick (25 to 50 mm) section of PVC cellular foam or balsa wood-based core materials using a suitable polymeric resin. In the present study, we focus on the hygrothermal effect on the fiber-dominated compression properties of carbon fiber reinforced vinyl ester resin based polymeric composite (CF/VE), used as “skin” for a polymeric composite sandwich material. Hygrothermal conditioning is achieved by saturating samples in simulated seawater at 40°C. Compression properties are evaluated for coupons extracted along warp and fill undergoing- no conditioning, conditioning till saturation (up to 6 months), and long-term conditioning (2 years). Sea-water saturation yields in up to 12% drop in compression strength with a further 3–4% drop resulting from long-term conditioning. No statistically significant modulus degradation is noticed due to short or long-term hygrothermal exposure. The failure mechanism of the warp extracted coupon, which fails in a splitting failure mode originating due to the delamination between the 0/90 interface, or the fill extracted coupon, which fails due to the instability caused by tow micro-buckling, remains unchanged due to combined exposure (short or long-term) of seawater and temperature. The loss in strength is attributed to the degradation of the fiber-matrix interface, which is validated via conducting single fiber push-in tests with a nominal diameter of 7 micron for conditioned and unconditioned coupons.

Keywords

Carbon fiber vinyl ester vacuum assisted resin transfer molding hygrothermal effects compression strength interfacial shear strength

Introduction

Fiber-reinforced polymer composites have gained high interest in applications that require a high strength-to-weight ratio.^1–3 Primarily, aerospace,^4,5 automotive,³ and sports industries⁵ have actively tried to replace conventional materials with carbon fiber-based composites. For naval applications, composite sandwich structures are of significant interest, and are often manufactured using thin (2 to 4 mm) composite facings made from carbon or glass fiber reinforcement, attached to a thick (25 to 50 mm) section of PVC cellular foam or balsa wood-based core materials using a suitable polymeric resin. In the present study, we focus on the characterization of facings made from carbon fiber reinforced vinyl ester resin based polymeric composite (CF/VE), used as “skin” for a polymeric composite sandwich material. The effect of exposure to sea water is investigated by pre-soaking of specimens for target duration at controlled temperature in the water bath. For marine applications, the behavior of CF/VE laminate-based sandwich structures exposed to long term sea water and coupled temperature effects is of current interest. Sandwich structures are often studied for their mechanical response in flexure using three-or four-point loading fixtures and an important mode of failure is the laminate subjected to compression. The effect of seawater on the tensile,⁶ fatigue,⁷ and interfacial delamination⁸ of CF/VE composites has been studied in the literature. Directly studying the behavior of these CF/VE laminates in compression is the topic of interest for this study.

The compressive properties of the composite are vulnerable to the test setup and specimen preparation. For a test to be valid, the stress field must be uniform in the gage area with minimal influence of the free edges. The test setup should have less sensitivity to imperfections and a valid failure mode that should happen within the gage region. Compression properties have been shown to be the function of means of load introduction.⁹ The effect of the means of load introduction on the test results is evaluated by examining the results from the Combined loading Compression (CLC) and Illinois Institute of Technology Research Institute (IITRI) fixtures.

An investigation of microscopic failure initiation and its progression is necessary to understand the resulting macroscopic compression properties. Various failure modes have been reported in the literature for compression failure of polymer composites- delamination,^10,11 kinking,^10–12 fiber micro-buckling,^{10,11,13–15} and in a few cases, fiber failure.^10,11 These failure modes initiate at a microscopic or mesoscopic level, resulting in macroscopic failure. Therefore, the macroscopic failure modes function as an indicator of underlying mesoscopic/microscopic failure mode. For example, a splitting failure mode usually implies delamination as the cause of failure initiation. The failure initiation modes are a function of constituent material properties and the microstructure of the composite under investigation. The mechanical properties’ dependency on the microstructure substantiates the difficulties in universal strength-based modeling of composites.

The CF/VE investigated in this study are manufactured using a non-crimp fabric (NCF). Non-crimp fabric-based composites are of interest because of the lower manufacturing and storage cost than the prepreg-based composites. Due to the additional support provided by stitching, the delamination resistance between the stitched interface (0/90 in this study) of the NCF composite is typically higher.^16,17 On the other hand, NCF adds other geometric and material nonlinearity mechanisms in the microstructure of the manufactured composites, such as resin pockets, tow misorientations, stitching, and localized fiber fracture resulting from the stitching process. These sources of nonlinearities can change the failure modes and thus the resulting in-plane mechanical properties. The effect of stitching on mechanical properties has been studied in literature, with most studies finding a significant reduction (up to 20%) of in-plane mechanical properties.^16–19

Water transport for composite materials immersed in water is usually caused by diffusion and absorption. The absorption happens through the voids, micro-cracks, and gaps between the fiber/matrix interphase.²⁰ A connected fiber/matrix imperfect interphase acts as capillaries and increases the permeability of the composites to a variety of fluids. Moisture saturation can induce matrix swelling and plasticization- this often occurs in matrices of relatively high hydrophilicity.²¹ Swelling introduces negative pressure on the matrix caused by being bounded by the almost undeformed fibers, which results in residual radial stresses that can lead to interfacial de-cohesion. Plasticization reduces the glass transition temperature of the matrix, Tg, often resulting in the softening.²¹ For matrices with low to moderate hydrophilicity, swelling and plasticization have minor effects, but hydrolysis can induce polymer embrittlement.²¹ Moisture saturation has been shown to cause degradation in in-plane mechanical properties.²² Understanding the compression behavior requires understanding mechanisms of failure that depend upon the micromechanical properties and microstructure. Interfacial shear strength (IFSS), a measurement of the bonding between the fiber and matrix, has been shown to affect the tensile,^22,23 compression^22,24 and impact properties of the composites. Moreover, seawater saturation causes a significant reduction in the interlaminar shear strength resulting from additional routes for water uptake developed due to microcracking and de-bonding.^25,26 The hygrothermal effects on the Fiber/Matrix interfacial strength is of prime importance in understanding the hygrothermal effects on the compression properties.^22,27–29 In this work, the authors investigated the combined impact of seawater and temperature upon the fiber-matrix interface and their effect on failure modes and, thus, the resulting compression properties.

Materials and method

Materials

The reinforcement investigated in this study is Devold AMT AS stitched bonded fabric designated LT650-C10-R2VE. This is an equibiaxial Non-Crimp Fabric (NCF) produced using Toray’s T700 12 K carbon fiber (diameter ∼7 μm) tow with a vinyl ester compatible sizing. The matrix used is Dow Chemical’s DERAKANE 520A-40, a brominated vinyl ester formulated for the Vacuum Assisted Resin Transfer Molding (VARTM) process. The composite is manufactured using the VARTM process, for which the setup and cure time-temperature profile is described by Shivakumar.³⁰ The panel is cured for 72 h at 25.6°C and then ramped at 0.11°C/min for 7 h to 71.1°C where it is post-cured for another 20 h before cooling down at 0.56 C/min. Five 1” by 0.5” sections are extracted in warp and fill. Burn-off experiments were performed in a vacuum (to prevent ablation mass loss of carbon fiber) for the extracted sections. Table 1 shows the results of burn-off tests carried out in an inert atmosphere. No significant differences in volume fraction are obtained between the warp and fill coupons, with a standard deviation of less than 1% obtained between samples. The volume fraction obtained is higher than the one reported by Shivakumar et al.³⁰ using the areal density measurements. High-magnification optical images showed that there were no dry fiber regions or voids. Vacuum Assisted Resin Transfer Molding is an attractive process because it offers a cost-effective methodology for manufacturing large complex parts and thus is attractive for naval applications. The fabric is stitched with 14 g/sqm polyester knitting. Stitching has been shown to substantially improve the interlaminar fracture toughness in modes I and II and damage tolerance but with significant degradation in the in-plane properties.^16,17 The composite panels are fabricated using a 4-ply symmetric fabric layup, resulting in the layup of [0/90]_2s and [90/0]_2s in warp and fill, respectively. The total thickness of the panels obtained varies by around ± 5%, with an average thickness of 2.7 mm. On the other hand, the variation between the average thickness obtained between coupons varies by around ±1.75%, with an average thickness of 2.7 mm. The details of the fabric and the fabricated composite panel is shown in Figure 1(a) and (b). The consequence of the VARTM process is that the air side (the surface exposed to the vacuum bag) shows a much higher surface roughness than the Mold side. The average surface roughness (S_a) obtained using Keyence VK-X250k laser confocal microscope for the side A and B (Figure 1(b)) of the composite panel under investigation is 11.3 μm and 63 μm, respectively. The higher surface roughness value obtained for the air face is representative of the variation in curvature because of vacuum bagging, i.e., the flatness.

Table 1.

Fiber volume fraction from burn off tests.

	Fiber volume fraction
	Warp (%)	Fill (%)
Sample 1	64.8	63.2
Sample 2	64.5	62.7
Sample 3	65.9	63.2
Sample 4	65.9	64.4
Sample 5	66.3	64.0
Average	0.655 ± 0.0076	0.635 ± 0.0069

The coupons are cut along warp and fill. Henceforth these coupons are referred to as warp and fill extracted coupons, respectively. The sea-water saturated coupons are prepared by soaking the samples in simulated seawater at 40 C. Simulated seawater is prepared by mixing coarse sea salt (NaCl) with deionized water until the specific gravity of 1.022 of the resulting solution is achieved. The Combined Effect of temperature and seawater is then evaluated upon the compression properties of CF/VE composites. It has been shown³¹ that the CF/VE studied in this work saturates in the controlled temperature and seawater environment after 2 months. Example weight gain versus square root time curves are shown in Figure 2. More details about the kinetics of Fickian diffusion of seawater³¹ and diffusion coupled with wicking within the vinyl ester matrix can be found elsewhere.²⁰ Simulated seawater is reintroduced every 2–4 weeks to prevent excessive change in the system’s density, i.e., the salt concentration in water. Ten warp and fill extracted coupons are soaked for 6 months, while five warp and fill cut coupons are soaked for 2 years. Henceforth, these coupons are short-term and long-term soaked coupons, respectively. Additionally, five 1” by 10” CF/VE coupons (manufactured using the same processing parameters) that have been soaked for 15 years (same aging conditions as the short-term and long-term soaked coupons, i.e., simulated seawater at 40°C) are extracted from a different panel. These coupons are cut into 10 mm *10 mm sections and polished along the warp and fill face to examine the local microstructural changes due to extreme long-term exposure to seawater.

Methods

Compression testing

Compression test setup and procedure are done per ASTM D6641³² using the CLC fixture. The load is transferred to the sample using end loading combined with the shear loading to ensure efficient stress transfer and failure in gage length. MTS 810 test system with a 100 kN load frame setup is used for compression testing. The details of the test setup and sample dimensions are shown in Figure 1(c). The samples are precision cut using a grinding wheel equipped with a diamond blade. Alignment of samples within the fixture is achieved using a laser level. The surface in contact with the platen is critical to effectively transfer the load when the sample is subjected to compression. Sample ends are grinded using an automatic polisher with 320 grit to achieve optimal contact with the platen.

Digital image correlation validation

Compression testing requires strain measurement on both faces of the coupons to eliminate the effect of bending that may arise due to misalignment in the test setup.³² Two Digital Image correlation (DIC) systems- GOM Aramis with 12 MP cameras and Correlated solutions VIC 3D with 1.4 MP cameras- are used to evaluate the strain on either face of the coupon. Digital Image correlation offers spatial strain resolution and thus the zones of strain concentration which can be a valuable tool in determining the zones of failure initiation and consequently the failure propagation. However, the accuracy of the measured strain via DIC is a function of the DIC setup calibration. Therefore, the measured strain requires validation. Two sets of experiments are conducted to validate the strain obtained via DIC - (1) Compression testing using DIC (VIC 3D) and strain gauge, (2) Compression testing using VIC 3D and GOM Aramis- measuring strain on the side B (Figure 1(b)) during compression testing. Side B is chosen for validation because of the higher surface roughness and thus a higher probability of poor correlation for DIC setup. Figure 3(a) shows an excellent agreement between the strain obtained via strain gauges and the VIC 3D setup. Figure 3(b) shows the results obtained for the two DIC setups measuring strain on the same face. The results in Figure 3(b) are calculated by defining a representative strain gauge area and an effective extensometer for both setups. The strain measured via a representative strain gauge is an average strain for each sub-volume within the defined area. In contrast, the strain is calculated by measuring the change in length of the effective extensometer, which is only a function of endpoint displacements. It was observed that the choice of location and length of the extensometer significantly changes the strain obtained, which is a direct consequence of the high strain gradient on the face of the sample. Thus, the strain measurements in this study use the representative strain gauge.

Effect of torque on test results

Test setup can play a significant role in efficient load transfer and thus the resulting compression properties. Care must be taken to ensure that the failure happens within the gage region with efficient load transfer. The torque used to constrain the sample for the CLC fixture can have a constraining effect on the sample and thus the resulting compression properties. Three dry samples extracted in warp and fill for each torque condition (20lb-in-35lb-in) are tested using the CLC fixture to understand the torque effect. Side faces were not tabbed for any of the torque conditions. The resulting stress-strain plots are shown in Figure 4(a) and (b), while the results are reported in Table 2. All but one warp cut coupon corresponding to a 35 lb-in. Torque (Figure 4(a)) failed within the gage area. The torque for the fill extracted coupon is varied between 20b-in. And 30lb-in. No correlation in strength, modulus, or the resulting stress-strain curve is observed as a function of torque. 30lb-in. Torque showed the maximum average strength and was chosen for further testing. Furthermore, since no failure was observed at the ends, the side faces were not tabbed for the further tested CLC coupons.

Table 2.

Effect of torque on fill and warp extracted coupons.

	Warp		Fill
Torque (lb-in)	Strength (MPa)	Modulus (GPa)	Strength (MPa)	Modulus (GPa)
20	413	51	338.2	53
25	404	54	312.8	55
30	409	52	349.1	52
35	393	53

Role of fixture on test results

The stress field developed during compression testing is sensitive to the loading mechanism. Various fixtures have been proposed to calculate the compression properties of polymer composites. These fixtures can be broadly characterized into three categories – End-loaded, Shear-loaded, and Combined-loaded fixtures. End-loaded fixtures have decreased in popularity because of the higher probability of failure within the fixture. Illinois Institute of Technology Research Institute fixture is a compression testing fixture that transfers the load via shear loading. An advantage of using the IITRI fixture is that the flatness at the end of coupons has little effect on the resulting compression properties as the load is transferred through the side faces. Side faces are tabbed using Garolite (G10) and bonded using epoxy resin with hardener (5 to 1 by weight) and glass filler to prevent stress concentration while the side faces of CLC coupons are not tabbed because of the gage failure observed for un-tabbed coupons while evaluating the effect of torque. A displacement rate of 1.5 mm/s is chosen for IITRI testing as per.³³ Five dry coupons cut in warp and fill tested for IITRI and CLC fixture. All coupons failed in the gage region. Table 3 shows the resulting mechanical properties for both fixtures. The compressive strength obtained via the CLC fixture is higher than that obtained via the IITRI fixture. This result indicates the inefficient transfer of load from the faces through the sample for the IITRI fixture, which might be due to the difference in surface roughness of the faces or due to the inherent material un-symmetry present because of the local differences in microstructure of the fabric. Henceforth, the compression properties are reported using the CLC fixture.

Table 3.

Comparison of mechanical properties obtained via CLC and IITRI fixture.

	Warp		Fill
Fixture	Strength (MPa)	Modulus (GPa)	Strength (MPa)	Modulus (GPa)
CLC	453	52.0	373	48.6
IITRI	431	52.5	352	48.3

Results

Fifteen dry coupons, ten short-term soaked coupons, and five long-term soaked coupons, extracted in warp and fill, are tested using the CLC fixture. As previously stated, the coupons are tested with 30lb-in. Torque. Typical stress-strain curves for dry and seawater-soaked warp and fill cut coupons are shown in Figure 5. Face 1 and 2 correspond to the specimen’s strain on the opposite faces (Mold and Air face). The average of the two strains is the desired value since the amount of bending due to sample misalignment does not affect the average strain.

Thin laminates under compression are prone to buckling. The minimum thickness required to avoid Euler buckling can be estimated by the equation (1).³³ Taking a conservative approach and assuming a modulus of 50 GPa with compressive strength of 600 MPa gives a minimum thickness of 1.7 mm for a gage length of 13 mm, which is smaller than the thickness of CF/VE being investigated. Moreover, the stress-strain curve validates the absence of global buckling as there is no evidence of instability. Even though the global buckling is absent, there is significant bending in fill coupons. The presence of constant bending in the fill coupon implies the nonlinearity arising from material imperfections rather than a systematic misalignment in the test setup. A closer examination via optical microscopy revealed a higher degree of misorientation in fill coupons (Figure 5), perhaps indicating bending due to geometric nonlinearity.

h \geq \frac{1}{0.9069 \sqrt ((1 - \frac{1 \cdot 2 F^{C U}}{G_{x}}) (\frac{E^{l}}{F^{C}}))}

(1)

The summary of compression strength and compression modulus obtained for dry aged and sea-water saturated coupons is shown in Figure 6 and Table 4. Compression modulus decreases between 1.35–2.1% and 0.9–1.3% for warp and fill extracted coupons, respectively. The higher value of compression modulus for the long-term soaked coupons compared to the short-term soaked coupons perhaps indicates the insignificance of the two data sets. A student t-test comparison between the short-term and long-term soaked warp samples results in a Z value of 0.47, indicating a less than 50% confidence interval of the difference between the two data sets. On the other hand, the difference between dry aged and saturated warp coupons shows a greater than 80% confidence interval. Thus, the authors conclude a 1.35–2.2% decrease in compression modulus of warp coupons due to combined exposure to seawater and temperature. There is no significant change in modulus between the long-term and the short-term saturated warp coupons. Fill coupons show a relatively lower decrease with a higher standard deviation in modulus. Thus, it is difficult to statistically state any reduction in fill coupons compression modulus due to long-term hygrothermal conditioning.

Table 4.

Effect of seawater on compression properties of CF/VE.

	Warp		Fill
Conditioning (N)	Strength(MPa) (Stdev.)	Modulus(GPa) (Stdev.)	Strength(MPa) (Stdev.)	Modulus(GPa) (Stdev.)
Dry (20)	411 (23.4)	51.0 (2.7)	395 (29.9)	50.7 (3.3)
6 months soaked (10)	361 (27.4)	49.9 (2.2)	348 (46.9)	50 (3.0)
2 years soaked (5)	348 (21.5)	50.3 (1.3)	330 (68.4)	50 (2.5)

On the other hand, strength shows a significant decrease between the dry aged and saturated coupons. The strength decreases by 12.3 and 12% for warp and fill extracted short-term soaked coupons with a further reduction in strength by an additional 3.1 and 4.2% when conditioned for 2 years. Another interesting observation can be made from Table 4. The standard deviation corresponding to the compressive strength of fill coupons increases with long-term exposure to seawater, suggesting that some fill coupons are more prone to strength degradation than others, perhaps an indicator of change in failure mechanisms.

Discussion

Modes of failure for dry aged and saturated coupons

Compression failure is not only the function of the material properties of reinforcement, interface, and matrix but is also largely dependent upon the microstructure of the composite under investigation. Macroscopically, failure modes can indicate the fundamental cause of microscopic/mesoscopic failure initiation. Figure 7 shows the macroscopic failure modes observed for dry warp and fill extracted coupons. Splitting is observed as the macroscopic failure mode for all warp cut coupons with delamination between the 0/90 layers, suggesting failure governed by the fiber-matrix interface. Fill coupons show kink bands in the post-failure microscopy and a high degree of matrix cracking.

To investigate the underlying failure mechanisms, the edges of warp and fill extracted samples are polished using the Buehler MetaServ 250 auto polisher. In-situ imaging under compression is done to observe the evolving microstructure of warp and fill cut coupons. The coupons are tested till 90% failure load and imaged using Keyence VHX-7000 to monitor any zones of failure initiation. The misorientation (∅) is characterized by calculating the tangent inverse of equally spaced points distributed across the tows of samples (Figure 8(a)). The maximum misorientation (∅_max) calculated before loading and before failure for fill extracted coupons increases from 4.1° to 5.2°, respectively, as seen in Figure 8(a). The increase in misorientation is indicative of tow micro-buckling. The finding is consistent with the computational study done by Wisnom et al., showing that NCF compression strength is controlled by 0° tow geometric instability arising at mesoscale and accompanied by resin shear plastic flow.²⁴ The coupons exhibit no microstructural changes via optical microscopy until the failure load is achieved. At this point, there is a catastrophic failure arising from the overall shear of the coupon. The post-failure microscopy shows kinking, matrix-cracking, and delamination as the failure mode. Warp extracted coupons do not show significant changes in microstructure during compression testing. Though, optical microscopy (Figure 8(b)) of coupons tested to 90% failure load revealed delamination at the edges of the coupon providing further evidence of failure to be controlled by the fiber-matrix interface with delamination as the failure initiation mode.

Sea-water saturated samples are taken out of their controlled environment and are immediately speckled (for DIC measurements) and tested. Therefore, it is not possible to understand the failure evolution of saturated samples via polishing. However, DIC can be used to understand the failure initiation and evolution of these coupons. An instantaneous change in out-of-plane deformation could indicate failure initiation and subsequent propagation. Thus, out-of-plane velocity quantifying the change in out-of-plane deformation is defined as per equation (2). Macroscopically, saturated warp extracted coupons fail in a splitting failure mode with a reduction in strength, indicating a decrease in interfacial strength of CF/VE due to long-term exposure to seawater. Figure 9(a) and (b) show the failure evolution via DIC and the resulting stress-strain curve obtained for a saturated warp cut coupon. The failure initiates at the edges via delamination and then propagates longitudinally and transversely until failure.

\frac{d w}{d t} = (w_{i + 1} - w_{i}) * η

(2)

Saturated Fill cut coupons show high variation in strength. Thus, the effect of saturation upon fill extracted coupons cannot be captured by just understanding the failure evolution of one coupon. Out-of-plane velocity is used to understand the failure evolution of fill cut coupons (Figure 10). Like the warp extracted coupons, the failure initiates at the edge. Unlike the warp cut samples, the failure is localized, and the out-of-plane velocity evolution is radial until the catastrophic failure. Resembling the failure sequence of dry fill cut coupons, this trend suggests the cause of failure localization to be tow micro-buckling.

Figure 11 shows the failure evolution of a saturated fill coupon captured via DIC images. The failure initiates as a combination of tow micro-buckling and delamination. The crack then propagates both transversely and through-thickness, eventually causing failure. The dry and saturated fill cut coupons undergo tow micro-buckling; thus, the mechanical properties in the vicinity of localization and the microstructure dominate the failure. The above evidence of failure evolution alludes to the cause of higher variation in failure strength observed in saturated fill extracted coupons. A lower fiber-matrix interfacial strength in the vicinity of micro-buckling will result in splitting as the failure mode. In comparison, a relatively higher interfacial strength will result in a catastrophic failure, perhaps due to the geometric nonlinearity caused by kinking. The above rationale is augmented by Waas et al., who found kinking and splitting to be competitive failure modes that depend upon mode II fracture toughness and mode-II interfacial strength.¹²

Effect of seawater upon the fiber-matrix IFSS

As discussed in the previous section, the decrease in compressive strength for both warp and fill extracted coupons insinuate a decrease in fiber-matrix interfacial properties. Thus, the authors evaluate the differences in interfacial strength between the dry aged and saturated coupons using the single fiber Push-in method,^34,35,36 corresponding to the nominal fiber diameter of 7 μm. First, the fibers are statically loaded using the Nanomechanics inc. iMicro Nanoindenter with a displacement rate of 50 nm/s using a truncated cone with a 5 μm initial diameter and a cone angle of 60°. The resulting load-depth curve is then used to quantify the interfacial strength as per equation (3).

T_{c}^{S L} = \frac{S_{0} P_{c}}{2 π^{2} r^{3} E_{11}^{f}}

(3)

Where S₀ is the stiffness of the linear portion of the load-displacement curve, P_c is the debonding load, E₁₁ is the longitudinal fiber modulus (230 GPa), and r is the Fiber radius. The initial non-linear portion of the load-depth curve represents the tip contacting the fiber. Once the contact is established, the deformation is elastic until the onset of nonlinearity, expressing a break in the fiber-matrix interface. For the coupons tested, it was found that the slope of the load-depth curve consistently remains linear between 15 mN and 25 mN. Thus, S₀ is calculated from the slope of the load-depth curve corresponding to the load ranging between 15 to 25 mN. Since there is no apparent debonding load (P_c) in the load-depth curve, P_c is identified using the intersection of the load-depth curve with the 5% offset line of slope S₀. Therefore, the definition of debonding load is abstract. Thus, values reported in this study must only act as the qualitative differences between the data sets.

Interfacial strength represents the bonding between the fiber and matrix and thus likely differs from one location to another. Hence, 25 tests are conducted stochastically for each fiber tow that lies within the cross-section of the coupon to capture the range of values of interfacial strength across the cross-section. Each location is then optically imaged to identify the tests for which the tip-maintained contact with the fiber throughout the test. Five fill and warp cut coupons per environmental conditioning are evaluated. Figure 12(a) and (b) show the optical image of pushed-in fibers revealing interfacial cracking and an example of the obtained load depth curve.

The histogram expressing the range of measured values of fiber-matrix IFSS for dry aged and saturated coupons is shown in Figure 13. As expected, a peak shift in the normalized probability curve for both warp and fill coupons indicate a loss of interfacial strength due to long-term combined exposure to seawater and temperature. Furthermore, there is a reduction in the distribution width, reducing the standard deviation of the normalized curve, which is perhaps an indication of proportionality between the strength and reduction in interfacial strength.

Degradation in interface due to long-term combined exposure to seawater and temperature is a consequence of chemical degradation and the continual hygrothermal stresses developed due to the disparate coefficient of thermal and moisture expansion between the resin and fiber. Thus, the degradation would only increase with time.³⁷ Therefore, few coupons conditioned for 15 years are inspected using a Scanning Electron Microscope (Figure 14). Based on Figure 14, cracking occurs along the tow circumference and along with the stitching. It should be stated that only two out of the various coupons inspected showed cracking of the interface. Fibers within the tow have similar misorientation. Thus, each tow acts like a sub-composite under steady hygrothermal stresses, resulting in the highest residual stresses developing at the tow-matrix interface. Similarly, the degradation due to negative pressure introduced due to swelling of the matrix will be maximized at the tow-matrix interface. Therefore, the fiber-matrix interfacial strength degradation is maximum for fibers at the tow matrix interface. A similar rationale can be followed for the degradation in stitching-matrix interfacial strength.

Conclusions

In this experimental study, the authors demonstrated the combined effect of long-term temperature and seawater upon the compression behavior of NCF-based CF/VE composites. Compression testing is done using the CLC fixture while the effect of role of fixture in determining the resulting compression properties is evaluated using an IITRI fixture. DIC is used to measure the strain. The following conclusions are made in the study:

1. Coupons extracted along warp and fill shows ∼12% decrease in compression strength due to sea-water saturartion. Long-term (2 years) exposure causes a further 3% and 5% drop in compression strength of coupons extracted along warp and fill, respectively.

2. Compression modulus is minimally effected for coupons extracted along warp and fill due to short or long-term exposure to sea water.

3. Coupon extracted along warp macroscopically fails in splitting failure mode initiated by edge delamination which is observed via optical microscopy of 90% loaded sample. There is no change in the failure initiation mode for the sea-water saturated coupons, and the measured lower compressive strength is attributed to the loss in interfacial strength.

4. An increase in misorientation is observed just before failure in coupons extracted along fill during compression testing, which indicates tow micro-buckling. The evidence of tow micro-buckling is obtained for the saturated fill coupons via a radial increase of out-of-plane velocity localized at the sample edge before failure. The interfacial strength in the vicinity of the localized deformation becomes a crucial micromechanical property upon which the failure mode and the compression strength are dependent. A higher variation in compression strength of soaked coupons is thus attributed to the variation in interfacial strength in the vicinity of localized deformation.

5. Single fiber Push-in tests for fibers with nominal diameter of 7 μm are conducted to quantify the effect of long-term exposure to seawater and temperature upon the IFSS. The results verified a significant drop in IFSS. Interfacial degradation is maximum for fibers corresponding to the tow-fiber interface.

Footnotes

Acknowledgements

The authors would like to acknowledge Dr. Stephen Young, Dr. Siriruk Akawut, Mr. Zach Arwood, and Mr. Josh Crabtree at the University of Tennessee for their assistance with the experimental setup and data analysis. In addition, Dr. Penumadu would like to acknowledge the support of the US Office of Naval Research from the Solid Mechanics Program under Program Managers Dr. Yapa Rajapakse and Dr. Paul Hess. He would also like to acknowledge the support of the US Department of Energy through the award to support IACMI.

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 research is partially supported by the general infrastructure grant of the DOE-Nuclear Energy University Program (DE-NE0000693) and the Nuclear Engineering University Program (NEUP) Award Number 12-3528.

ORCID iD

Vivek Chawla

References

Savage

. Carbon-carbon composites. London, UK: Chapman & Hall, 1993.

Loud

. Commercial and industrial applications of composites. In: Handbook of composites, 1998, p. 941.

Bajpai

. Update on carbon fibre, 2013.

Awasthi

Wood

. Carbon/carbon composite materials for aircraft. Brakes 2008; 560: 553–559.

Fitzer

Manocha

. Carbon reinforcements and carbon/carbon composites. New York, NY, 1998, p. 73.

Siriruk

Penumadu

Sharma

. Effects of seawater and low temperatures on polymeric foam core material. Exp Mech 2012; 52: 25–36.

Siriruk

Penumadu

. Degradation in fatigue behavior of carbon fiber-vinyl ester based composites due to sea environment. Compos B Eng 2014; 61: 94–98.

Siriruk

Penumadu

Jack Weitsman

. Effect of sea environment on interfacial delamination behavior of polymeric sandwich structures. Compos Sci Technol 2009; 69: 821–828.

Chaterjee

Adams

Oplinger

. Test methods for composites a status report volume II. Compression test methods. Report No. DOT/FAA/CT-93/17. 1993; 231.

10.

Schultheisz

Waas

. Compressive failure of composites, part I: testing and micromechanical theories. Prog Aerosp Sci 1996; 32: 1–42.

11.

Waas

Schultheisz

. Compressive failure of composites, part II: experimental studies. Prog Aerosp Sci 1996; 32: 43–78.

12.

Prabhakar

Waas

. Interaction between kinking and splitting in the compressive failure of unidirectional fiber reinforced laminated composites. Compos Struct 2013; 98: 85–92.

13.

Lee

Soutis

. A study on the compressive strength of thick carbon fibre-epoxy laminates. Compos Sci Technol 2007; 67: 2015–2026. DOI: 10.1016/j.compscitech.2006.12.001

14.

Basu

Waas

Ambur

. Compressive failure of fiber composites under multi-axial loading. J Mech Phys Solids 2006; 54: 611–634.

15.

Lee

. Measuring the notched compressive strength of composite laminates: specimen size effects. Compos Sci Technol 2008; 68: 2359–2366.

16.

Mouritz

Cox

. A mechanistic approach to the properties of stitched laminates. Compos A Appl Sci Manuf 2000; 31: 1–27.

17.

Dransfield

Baillie

Mai

. Improving the delamination resistance of CFRP by stitching-a review. Compos Sci Technol 1994; 50: 305–317.

18.

Yudhanto

Watanabe

Iwahori

, et al. Compression properties and damage mechanisms of stitched carbon/epoxy composites. Compos Sci Technol 2013; 86: 52–60.

19.

Mouritz

Leong

Herszberg

. A review of the effect of stitching on the in-plane mechanical properties of fibre-reinforced polymer composites. Compos A Appl Sci Manuf 1997; 28: 979–991.

20.

Fichera

Carlsson

. Moisture transport in unidirectional carbon/vinylester panels with imperfect fiber/matrix interface. J Compos Mater 2016; 50: 751–760.

21.

Colin

Verdu

. Humid ageing of organic matrix composites. In: Durability of composites in a Marine enviornment, 2014, pp. 47–114.

22.

Figliolini

Carlsson

. Mechanical properties of carbon fiber/vinylester composites exposed to marine environments. Polym Compos 2014; 35: 1559. DOI: 10.1002/pc.22809

23.

Pukánszky

. Influence of interface interaction on the ultimate tensile properties of polymer composites. Composites 1990; 21: 255–262.

24.

Drapier

Wisnom

. Finite-element investigation of the compressive strength of non-crimp-fabric-based composites. Compos Sci Technol 1999; 59: 1287–1297.

25.

Sun

Zhao

Luo

, et al. Effect of temperature and cyclic hygrothermal aging on the interlaminar shear strength of carbon fiber/bismaleimide (BMI) composite. Mater Des 2011; 32: 4341–4347.

26.

Pavlidou

Papaspyrides

. The effect of hygrothermal history on water sorption and interlaminar shear strength of glass/polyester composites with different interfacial strength. Compos A Appl Sci Manuf 2003; 34: 1117–1124.

27.

Chen

Miao

Ding

. Influence of moisture absorption on the interfacial strength of bamboo/vinyl ester composites. Compos A Appl Sci Manuf 2009; 40: 2013–2019.

28.

Bradley

Grant

. The effect of the moisture absorption on the interfacial strength of polymeric matrix composites. J Mater Sci 1995; 30: 5537–5542.

29.

Pérez-Pacheco

Cauich-Cupul

Valadez-González

, et al. Effect of moisture absorption on the mechanical behavior of carbon fiber/epoxy matrix composites. J Mater Sci 2013; 48: 1873–1882.

30.

Shivakumar

Swaminathan

Sharpe

. Carbon/vinyl ester composites for enhanced performance in marine applications. J Reinf Plast Compos 2006; 25: 1101–1116.

31.

Summerscales

. Durability of composites in the marine environment. 2014. Epub ahead of print 2014. DOI: 10.1007/978-94-007-7417-9_1

32.

ASTM . Standard test method for compressive properties of polymer matrix composite materials using a combined loading compression (CLC). Annu B ASTM Stand 2012; i: 1–11.

33.

Method

. Standard test method for compressive properties of polymer matrix composite materials with unsupported gage section by shear. ASTM 2020; 08: 1–16.

34.

Rodríguez

Molina-Aldareguía

González

, et al. A methodology to measure the interface shear strength by means of the fiber push-in test. Compos Sci Technol 2012; 72: 1924–1932.

35.

Medinam

Molina-Aldareguía

González

, et al. Comparison of push-in and push-out tests for measuring interfacial shear strength in nano-reinforced composite materials. J Compos Mater 2016; 50: 1651–1659.

36.

Ureña

Rams

Escalera

, et al. Characterization of interfacial mechanical properties in carbon fiber/aluminium matrix composites by the nanoindentation technique. Compos Sci Technol 2005; 65: 2025–2038.

37.

Ray

. Temperature effect during humid ageing on interfaces of glass and carbon fibers reinforced epoxy composites. J Colloid Interface Sci 2006; 298: 111–117.