Study on out-of-plane tensile strength of angle-plied reinforced hybrid CFRP laminates using thin-ply

Abstract Thin-plies are generally defined as composites with ply thicknesses below 100 μm. These materials are rapidly gaining interest for high-performance applications, for example, the aerospace sector. Many practical techniques have been proposed to prevent delamination and improve the strength of composite laminates. A recent study has shown that the delamination could be postponed by replacing layers of CFRP with thin-ply in a unidirectional composite laminate, a configuration known as hybrid laminates reinforced with thin-plies. Since fiber orientation is known to be one of the most important parameters in composite laminate design, this study investigates the effect of oriented layers of thin-ply or both thin-ply and conventional CFRP in a hybrid laminate under out-of-plane tensile loading. A numerical Representative Volume Element (RVE) model for CFRP and thin-ply was generated, considering the unidirectional [0], cross-ply [45/−45], and [0/90] in order to better understand the effect of angle-plied hybrid composite laminates. Experimental results show that angle-plied composite laminates present higher failure load under out-of-plane tensile loading compared to the unidirectional ones. This can be attributed to the fact that an initiated crack is faced with a significantly more complex crack path in an angle-plied laminate to advance in the through-the-thickness direction.


Introduction
The use of carbon fiber reinforced polymer (CFRP) materials in different industries is continuously increasing [1][2][3][4].Thin-ply laminates are defined as laminates composed using plies with a thickness of less than 100 lm [5].These layer thicknesses are available through the spread-tow process [6] which produces flat and straight plies with a dry ply thickness as low as 0.02 mm [7].With an improved spreading process, small layer thicknesses and more homogeneous fiber are achieved [8].
By reducing the thickness of the single layer, the quantity of the individual layers and therefore the degrees of freedom in the orientation are increased [9].This results in a larger number of interfaces in thin-ply laminates, lowering the shear stresses [9,10].Moreover, thin-ply laminates are known for their ability to delay the onset of the matrix damage mechanisms and suppress out-of-plane microcracking [5] and free edge delamination [9,11] for static, fatigue, and impact loadings.Due to their superior damage and delamination resistance properties, thin-ply laminates could exhibit higher interlaminar shear properties [12] and strain energy [13] compared to conventional plies.Thin plies can also represent a favorable approach to improve the performance of adhesively bonded CFRP due to their ability to enhance the off-axis performance of composites and postpone delamination [13].Moreover, due to the in situ effect [14] the location of composite failure could be changed from the ply interface toward the mid-thickness of the composite adherend.
It has been mentioned that, in the case of single lap joints, failure load seems to be directly correlated to the distance of the ply at 0 to the adhesive layer.Since the 0 layer is known to be the location of the final failure, it can be assumed that its distance from the adhesive layer hinders the crack propagation process and thus increases the joint strength [31].However, in the case of double lap joints under impact loading [32] joints having substrates with fibers oriented parallel to the loading direction (unidirectional 0 ) were found to present the highest shear strength.
A previous study by the authors [33] has shown that replacing layers of conventional composite in a unidirectional laminate with layers of thin-ply can postpone the delamination and increase the strength under out-of-plane tensile loading.In the first stage of the study, the effect of the thin-ply thickness in a specific hybrid laminate was studied.It was found that the use of 25% thin-ply per total thickness of the laminate presents the highest laminate strength under out-of-plane tensile loading (representing the optimum amount).In a second stage, the effect of distributing the optimum amount through the thickness was investigated.Figure 1a shows the studied configuration for unidirectional reference CFRP, thin-ply, and the optimum hybrid laminate configuration.Figure 1b illustrates the experimental results obtained.
The current article investigates the effect of using angleplied layers in hybrid composite laminates under out-of-plane tensile loading, both numerically and experimentally.The HS 160 T700 by CIT and TP415 by NTPT are used as conventional CFRP and thin-ply, respectively.An experimental study was performed by testing reference angle-plied CFRP and thin-ply configurations.Afterwards, a RVE model was carried out to better understand the effect of angle-plied layers in both CFRP and thin-ply configurations.The experimental results showed that angle-plied composite laminates present higher performance load under out-of-plane tensile loading when compared to the unidirectional configurations.

Conventional CFRP
The materials used in the studied configurations were chosen to be as representative as possible of a final application in  the aerospace sector.Thus, a unidirectional prepreg carbonepoxy composite with ply thickness of 0.15 mm was selected, with the commercial reference Texipreg HS 160 T700 (Seal Spa, Legnano, Italy).This is an orthotropic material, whose mechanical properties are presented in Table 1.
The elastic mechanical properties of the CFRP correspond to the orientation of a 0 CFRP ply (1 and 2 are defined as fiber and out-of-plane direction)

Thin-ply
Unidirectional 0 oriented carbon-epoxy prepreg composite with ply thickness of 0.07 mm was selected, with the commercial reference NTPT-TP415.The elastic orthotropic properties for this thin-ply were characterized using a servohydraulic testing machine (Instron 8801), with a load cell of 100 kN and following appropriate testing standards.

Tensile test 0˚and 90T
he tensile test in order to obtain E 1 and E 2 is performed using ASTM D 3039/D 3039 M standard.Figure 2 presents the schematic design for specimens used for (a) Longitudinal and (b) Out-of-plane tensile test.Mentioned t in Figure 2 represents the thickness of the specimen.A quasi-static test (2 mm/min) was performed according to the mentioned standard.
Digital image correlation was performed in order to obtain the strain using an open-source Moire software in order to measure contour of deformation, strain on the composite material.The specimens were painted with white spray and speckled with white ink dots (see Figure 3).This is necessary to support the DIC analysis, based on the correlation of a set of neighboring points considered in pictures captured before and after loading [20].Figure 4 shows the representative stress-strain curve obtained for the specimens with 0 and 90 fiber direction.

Shear test
Different standards have been used to perform the shear test for composite unidirectional materials.In this case, the    ASTM D 3518/D 3518M standard was used in order to obtain G 12 : A plate with lay up of [45/À45] 6s was manufactured.Figure 5 shows a schematic design for the manufactured specimen.A quasi-static test (2 mm/min) was performed according to the mentioned standard.
Digital image correlation was performed in order to obtain the strain using an open-source Moire software for data processing.Figure 6 shows the stress-strain curve obtained for the specimens.Figure 7 shows the samples after failure.

DCB test
In this case, the D 5528-01 standard was used to obtain mode I fracture energy (GIC).Figure 8 shows the schematic design of the specimens manufactured for DCB test.Steel blocks were attached to the DCB specimens.A Teflon layer was used to create a precrack in the DCB specimen, which is shown by as green line in Figure 8.
The specimens are attached to steel blocks, as shown in Figure 8, allowing testing in a universal testing machine.The surface of laminates was abrased smoothly using sandpaper and then cleaned with acetone, since release agent was used in the plate manufacturing step in order to prevent the adhesion between the plates and the mold.Moreover, plasma treatment was performed on the composite surfaces.Furthermore, it has to be mentioned that the steel blocks were sandblasted and then washed with acetone.
CBBM [35] method is used in order to obtain G IC and Figure 9 shows the representative R-curve obtained using the CBBM method.The DCB specimen testing and its failure morphology are shown in Figure 10.Table 2 shows a summary of mechanical property characterization for NTPT-TP415 thin-ply

Plate manufacturing
The manufacturing process of the reference CFRP or thinply starts with the layer-by-layer stacking of the CFRP prepreg, until the desired block thickness is attained.It should be mentioned that the overall thickness of the laminates is set to 3.2 mm.A mold is used to ensure the thickness of the manufactured plates and release agent is used to ensure easy separation of the plate and the mold after curing.Finally, the plates were cured in a hot plate press under 30 bar of pressure and 130 C for 2 h as recommended by the producer.
After curing, the plates were cut to the desired dimensions (25 Â 25 mm 2 ).Steel cubes (see Figure 11) were then   attached to blocks, using the PLEXUS MA422 adhesive, which cures in room temperature after 24 h.The adhesive overflow was removed from the outer surface of the blocks using sandpaper.

Configurations
Based on the results previously obtained by the authors [33], the CFRP þ 25%thin-ply/3 configuration was considered for this study.For the first stage of the study, in a hybrid block, the conventional CFRP was kept unidirectional (0 ) and thin-ply layers were oriented at [45/À45] n and [0/90] n seeking symmetry of the final laminate (see Figure 12a, b).Moreover, the reference angle-plied thin-ply laminates ([0/90] ns and [45/À45] ns ) were also examined.

Scanning electron microscope studies
The cross-section of blocks with polished surfaces was observed using a Scanning Electron Microscope (SEM).In specimens manufactured with CFRP prepreg, resin-rich, and fiber-rich area is apparent (marked by red rectangle in Figure 13a).On the other hand, the fibers are well distributed in specimens manufactured using thin-ply prepregs.Therefore, less resin-rich and fiber-rich area is observed, when compared to the conventional CFRP (see Figure 13) which is in line with the literature review [8].

Surface treatment
As mentioned in Section 2.3, composite blocks have to be attached to steel blocks, (as shown in Figure 11) in order to    allow for assembly in the testing machine.Initially, the surface of the CFRP blocks was abrased using sandpaper.Afterwards, the blocks were cleaned with acetone, removing any release agent remaining from the plate manufacturing step (see Section 2.3).Finally, a plasma treatment was performed on composite surfaces.The steel blocks were sandblasted followed by cleaning with acetone.

Testing condition
The SLJs were tested using an Instron 8801 servo-hydraulic testing machine with a load cell of 100 kN, at a constant crosshead speed of 1 mm/min (quasi-static).All tests were performed under laboratory ambient conditions (room temperature of 24 C, relative humidity of 55%).A minimum of three repetitions were made for each configuration tested.

CFRP
The manufactured blocks were tested as described in Section 2.7. Figure 14a shows the representative load-displacement curve obtained for unidirectional and angle-plied CFRP laminates, tested under out-of-plane tensile loading.Experimental results illustrate that angle-plied laminates present higher strength compared to the unidirectional ones.According to the literature [31], higher strength in the angle-plied blocks may be due to the more complex through the thickness crack path these materials exhibit, especially when compared to the crack paths of the unidirectional configurations.Overall, this results in a higher failure load (see Figure 14b) and this was clearly observed in the experimentally obtained failure mechanism.As seen in Figure 15, an initiated crack in the laminate can easily propagate through the thickness in the unidirectional laminate.The crack path is expected to be more complex in angle-plied layers if the crack is moving through the thickness, which requires higher fracture energy.Therefore, as seen in Figure 14b, the crack will actually be forced to propagate between two plies in the angle-plied layers.

Thin-ply
Unidirectional and angle-plied thin-ply laminates were also manufactured and tested.The representative load-displacement curves registered under out-of-plane tensile loads are presented in Figure 16a.According to the experimental results, the angle-plied laminates present higher strength compared to the unidirectional ones.The explanation presented in Section 3.1 is also applicable.Moreover,  presents the failure mechanism for the unidirectional and angleplied thin-ply laminates.As seen in Figure 16b, in a unidirectional laminate the crack can easily propagate through the thickness of the laminate while in an angle-plied configuration the crack path progression through the thickness is more restricted, leading to its propagation between two layers.

Hybrid laminate with unidirectional CFRP and oriented thin-ply
The hybrid laminate configurations presented in Section 2.4 were tested and the resulting load-displacement curves are shown in Figure 17a.Generally, hybrid configurations with the unidirectional CFRP and angle-plied thin-ply layers were found to present higher failure loads when compared to the unidirectional laminate (CFRP[0] þ 25%thin-ply/3[0]).This is acceptable due to the increase in the failure load in angleplied thin-ply laminate loaded under out-of-plane tension loading (see Section 3.2). Figure 17b presents the failure mechanism for the mentioned unidirectional and angle-plied hybrid laminate.As seen in Figure 17b, the crack could easily propagate through the thickness in unidirectional hybrid laminate.In contrast, the crack progression through the thickness is more complex in an angle-plied laminate, which forces failure to occur between plies.

Hybrid laminate with oriented CFRP and thin-ply
Figure 18a shows the representative load-displacement curves obtained for hybrid configurations with oriented CFRP and thin-ply layers.As presented in Sections 3.1 and 3.2, the reference angle-plied CFRP and thin-ply laminates present higher failure load than the corresponding unidirectional reference laminates.This explains the increase in the failure load in a hybrid laminate with angle-plied CFRP and thin-ply.Figure 18b presents the failure mechanism for the hybrid blocks with unidirectional and angleplied CFRP and thin-ply.As seen in Figure 18b, a crack could easily propagate through the thickness in unidirectional hybrid laminate.

Numerical study
A representative volume element (RVE) model is employed to better understand the advantages associated with angleplied laminates in micro-scale.The RVE was modeled using the ABAQUS commercial software.Following the result of the SEM micrographs presented in Section 2.5, the model was designed to account for the difference in fiber distribution between the CFRP and the thin-ply (fiber clustering in CFRP RVE and relatively uniform fiber distribution for the thin-ply).Moreover, the number of the fiber was calculated using Eq. ( 1) in which N and D are the number and the diameter of the fibers, respectively, and L, W, and H are the length, width and the depth of the RVE, respectively.Moreover, V f is the volume fraction.The properties of the used fiber and matrix for the CFRP and thin-ply presented in Table 3 are provided by the manufacturer.

CFRP
A 3D elastoplastic RVE model with dimensions of 100Â100Â100 ðlmÞ 3 was studied to partially model two plies of CFRP.Therefore, fiber directions of [0], [45/À45],    and [0/90] were considered (see Figure 19).According to Table 3, and Eq. ( 1) the number of fibers was calculated considering the before mentioned RVEs dimensions.Moreover, fiber clustering is also considered for CFRP RVE models (according to the SEM micrograph data shown in Section 2.5).The boundary condition applied is shown in Figure 20.An out-of-plane displacement of 1 lm was applied to plane number 1 (in y direction) to fulfill the outof-plane tensile loading and the other plane perpendicular to the y-axis (plane number 6), was fixed in the mentioned direction.Both planes perpendicular to the xand z-axis (in this case plane numbers 4 and 5) were fixed in related perpendicular directions (x and z, respectively).The boundary conditions were equally applied for all configurations under study.The critical plane in each RVE should be determined, which is expected to be a plane transverse to the displacement direction.The planes were evaluated for each configuration and the critical plane was determined as plane number 5 for all configurations.Figure 20 shows the Von Mises equivalent stress distribution for CFRP RVE with the ply orientation of [0], [45/À45], and [0/90] on plane number 5. The color bar was restricted to 148 MPa which corresponds to the CFRP matrix maximum strength.Therefore, the elements with stress value higher than 148 MPa (shown by grey color) are known as failed elements.An open-source software (IC Measure) was used in order to obtain the area of failed elements.The level of failure is the defined as the area of failed elements divided by the total area of the RVE. Figure 21 presents the level of failure described above.According to Figure 21, since the unidirectional RVE shows a lower level of failure, it is expected for the angle-plied laminates to present higher strength compared to the unidirectional ones, with the [0/90] configuration expected to present the highest strength (the [0/90] configuration shows the lowest level of failure).

Thin-ply
The same approach mentioned in Section 4.1 was used for RVEs with the thin-ply configuration, as shown in Figure 22.Considering Figures 19 and 22, an attempt was made to implement the difference in fiber distribution between CFRP and thin-ply (considering relatively uniform fiber distribution in thin-ply RVE models compared to the models presented in Section 4.1).Consequently, RVEs with the fiber direction of [0], [45/À45], and [0/90] were considered (see Figure 22).The boundary condition is the same as mentioned in Section 4.1.
The critical plane in each RVE should be determined, which is expected to be a plane transverse to the displacement direction.Planes were evaluated for each configuration and the critical plane was determined to be plane number 5    The color bar was restricted to 138 MPa which corresponds to the maximum strength of the thin-ply matrix.Therefore, the elements with stress value higher than 138 MPa (shown in grey color) are known as failed elements.The level of failure as defined in Section 4.1 was then obtained and presented in Figure 24.According to Figure 24, as the unidirectional RVE is presenting higher level of failure.It is expectable for laminates with angle-ply layers to present higher strength compared to the unidirectional ones, with the [0/90] configuration expected to present the highest strength (the [0/90] configuration shows the lowest level of failure).

Discussion
Figure 25 presents a summary of experimentally obtained failure loads for the unidirectional and angle-plied laminates.Generally, the experimental study showed that hybrid laminates present higher strength compared to the reference conventional CFRP and composite.Moreover, angle-plied conventional CFRP and thin-ply laminates present higher strength compared to the unidirectional ones for both [45/À45] ns and [0/90] ns configurations.Therefore, an angleplied hybrid composite laminate (unidirectional CFRP with angle-plied thin-ply or angle-plied CFRP and thin-ply) presents higher failure load than unidirectional hybrid laminates (see Figure 25).A comparison between Figure 25a and b shows that the failure load obtained for angle-plied hybrid laminates with oriented CFRP and thin-ply is slightly higher than that exhibited by configurations with unidirectional CFRP and angle-plied thin-ply.This is mainly due to a more complex through the thickness crack path.Experimental results also show that the angle-plied laminates with [0/90] ns stacking sequence present higher strength for  all configurations.A microscale numerical study shows that in the case of [0/90] ns stacking sequence, the matrix failure is delayed compared to the [45/À45] ns configuration.

Conclusion
This study investigated the effect of introducing angled ply layers in a composite laminate, approaching the problem from numerical and experimental perspectives.Two stacking sequences of [45/À45] ns and [0/90] ns were considered for the manufacture of reference CFRP and thin-ply-based laminates.A hybrid laminate configuration was also studied experimentally.A numerical elastoplastic RVE model was generated for the CFRP and thin-ply with unidirectional and angle-plied fibers.Based on the obtained results, it can be concluded that: 1.A numerical analysis indicated that angle-plied CFRP and thin-ply laminates experience a lower level of failure under out-of-plane tensile loading, and therefore the matrix failure process is delayed.Accordingly, angle-plied laminates are expected to present higher strength compared to the unidirectional ones.2. According to the same numerical study, the stacking sequence of [0/90] presents a lower level of failure compared to [45/À45] for both CFRP and thin-ply configurations.3. Experimental results confirm that using angled plied laminates in the reference or hybrid laminate increases the failure load under out-of-plane tensile loading.This is because the through the thickness crack path is more complex in angle-plied laminate than unidirectional laminates.This is a line with the literature review.4. According to experimental result, hybrid laminates reinforced with thin-plies and with the stacking sequence of [0/90] ns present the highest strength under out-of-plane tensile loads.

Figure 1 .
Figure 1.(a) Schematic design for conventional CFRP, thin-ply, and hybrid laminate and (b) summary of the experimental results for unidirectional reference CFRP, thin-ply, and hybrid laminate.Adopted from [33].

Figure 3 .
Figure 3. Figure captured for DIC process from 90 specimen while testing.

Figure 4 .
Figure 4. Representative stress-strain curve obtained from the 90 and 0 tensile test.

Figure 5 .
Figure 5. Schematic design of specimens for shear test.

Figure 6 .
Figure 6.Representative stress-strain curve obtained from shear test.

Figure 8 .
Figure 8. Schematic design of the specimens manufactured for DCB test.

Figure 9 .
Figure 9. Representative R-curve obtained for DCB specimen using CBBM method.

Figure 11 .
Figure 11.Schematic design of steel blocks attached to conventional CFRP laminates and loading condition.

Figure 12 .
Figure 12.Schematic design of the studied hybrid laminates with oriented plies.

Figure 14 .
Figure 14.(a) Representative load-displacement curve and (b) failure mechanism for reference unidirectional and angle-plied CFRP laminates.

Figure 15 .
Figure 15.Expected through-the-thickness crack path in unidirectional and angle-plied laminates.

Figure 16 .
Figure 16.(a) Representative load-displacement curve and (b) failure mechanism for reference unidirectional and angle-plied thin-ply laminates.

Figure 17 .
Figure 17.(a) Representative load-displacement curve and (b) failure mechanism for hybrid laminate with unidirectional CFRP and oriented thin-ply.

Figure 18 .
Figure 18.(a) Representative load-displacement curve and (b) failure mechanism for hybrid laminate with oriented CFRP and thin-ply.

Figure 20 .
Figure 20.Von Mises stress distribution for unidirectional and angle-plied CFRP RVE on plane number 5.

Figure 21 .
Figure 21.Level of failure obtained for unidirectional and angle-plied CFRP RVE on plane number 5.
for all configurations.

Figure 23 .
Figure 23.Von Mises stress distribution for unidirectional and angle-plied thin-ply RVE on plane number 5.

Figure 24 .
Figure 24.Level of failure obtained for unidirectional and angle-plied thin-ply RVE on plane number 5.

Table 3 .
Geometrical and mechanical properties for the fiber and matrix.