Study on Fatigue Behaviors of Porous T300/924 Carbon Fiber Reinforced Polymer Unidirectional Laminates.
Light weighting, as one of the most urgent targets for vehicle and airplane manufacturing industries in meeting customer demands and government regulations for energy efficiency, can be achieved by the implementation of new materials with relatively low density as well as high stiffness and strength. Carbon fiber reinforced polymer (CFRP) composites, which can offer some unique advantages in terms of their high strength-to-density ratio compared to traditional materials, have become a promising material for future application.
Although CFRP composites have shown excellent fatigue resistance in previous applications, the fatigue failure of composite components can be very crucial. As some of the structural components (e.g., turbine blades) require good durability when subjected to complex loading conditions, which is detrimental to the components, fatigue life evaluation must be done during the components' design process.
Different from fatigue failure of metal, fatigue failure of CFRP composites is not dominated by a single main crack and its propagation, but is rather a combination of a variety of failure modes (e.g., fiber breakage, matrix micro-cracks, delamination, and fiber/matrix interfacial debonding) that act synergistically and simultaneously to cause damage to accumulate in the material .
Much research has been dedicated to the study of the fatigue behavior of CFRP composites, and some factors have proved to have impacts on the fatigue behavior: (1) mechanical properties of the material components, such as fiber, matrix, and fiber/matrix interface ; (2) fiber orientation and fiber volume fraction [3,4]; (3) loading conditions, including load frequency and load ratio [5, 6, 7]; (4) manufacturing defects and aging effect of the material, such as fiber defects  and voids in matrix [9, 10, 11] or mechanical degradation due to hygrothermal aging  or ultraviolet radiation .
Porosity is one of the most common defects in CFRP composites and is inevitable in fabrication of the material. Many efforts have been made to study the voids' effects on the mechanical properties of CFRP composites, the results of these studies show that many mechanical properties, including transverse tension/compression stiffness and strength, bending strength, interlaminar shear strength, and fatigue resistance, will degrade with voids present in the matrix [14, 15, 16]. Sanjay Sisodia investigated the effects of voids on CFRP laminates under tension-tension cyclic loading and quasi-static tensile loading, and the results showed that the fatigue life of composites is more sensitive to the void content in comparison with the static properties, including stiffness and strength; besides, the sensitivity to voids was found to be more significant for transverse cracking in the 90[degrees] plies under cyclic loading . Based on the test results, Sisodia suggested that quasi-static testing of the composites alone is not adequate for quality control of composite materials that may be used under cyclic loading. Despite all of these efforts, the understanding of voids' effects on fatigue of CFRP composites is still insufficient; the degradation of fatigue resistance of the material with voids present needs to be evaluated by an appropriate model in components design.
For the purpose of getting a better understanding of the voids' effects on the fatigue properties of CFRP unidirectional laminates, experimental research was carried out in this study by conducting tension-tension fatigue tests for unidirectional T300/924 CFRP laminates at different void levels under load control. The 12-ply and 16-ply 0[degrees] unidirectional laminates with void content at 0.8%, 2.2%, and 4% were fabricated by applying different compression pressures during the compression molding procedure. Dynamic stiffness degradations of the materials in the loading direction due to cyclic loading were also obtained by monitoring the axial displacement changes with respect to the number of cycles. For the assessment of the void content, a digital microscopy (DM) image analysis technique was implemented. On top of that, a scanning electron microscope (SEM) was also utilized to identify the underlying failure mechanisms. In an attempt to evaluate the voids' effects on fatigue behaviors of CFRP composite laminates as well as to predict the fatigue life, a residual stiffness model for porous composite laminates was presented by introducing the void content [v.sub.v] as a variable into the model of H. Mao. This model is able to reflect the voids' effects on the stiffness degradation of CFRP laminates and can be used to predict the fatigue life for composites at different porosity levels. The prediction results showed a good correlation with the test data.
The composite laminate plaques were fabricated by the compression molding approach. The reinforcement was T300 carbon fiber produced by Toray, and the matrix was HexPly 924 provided by Hexcel. For the purpose of obtaining plaques at different levels of porosity, 0.5MPa, 0.3MPa, and 0.1MPa were selected as the compression pressures during the compression molding procedure. The unidirectional laminate plaques have a size of 500mm x600mm with lay-up [[0[degrees]].sub.12] and [[0[degrees]].sub.16]. The resulting void contents were determined with the implementation of a digital microscopy (DM) image analysis technique. The fiber volume content of the laminates ranged from 47.9% to 49.5%, which were also obtained by the aforementioned DM approach. Details of the geometric and manufacturing parameters can be found in Table 1.
The samples for porosity characterization were cut by water jet from the laminate plaques with dimensions of 10mm x 10mm. Before DM analysis, the samples were first embedded into epoxy resin and cured for 24 hours; the cross sections were then prepared with #400, #600, #1200, and #2000 abrasive papers, and the final polishing was completed by using a synthetic diamond compound of 0.5um, 0.3um, and 0.1um. To accomplish the porosity characterization, a PC-based optical microscope with a CMOS camera was used in this study to capture micro-images of the cross-sectional area, with 100X and 200X magnification selected to provide adequate resolution of the fibers/matrix as well as a large enough representative area. For the purpose of providing a statistically representative data set, 5 samples for each plaque were prepared and 3 images per sample were examined. The captured images were analyzed by a software module based on MATLAB's image toolbox to obtain the morphological parameters of the voids in the composites, such as the voids' area content, denoted as [v.sub.v]; aspect ratio; and equivalent size of the micro-voids. Some typical micro-images of the samples cured at different compression pressures are shown in Figure 1.
Quasi-Static Tensile Test
The quasi-static tensile tests were carried out first on an MTS 647 hydraulic servo dynamic material test machine. The tests were performed under displacement control using a displacement rate of 1mm/min at room temperature (22[+ or -]0.5[degrees]C). For strain measurements, a 2-inch MTS E 634.31F-24 extensometer was used with its blade glued to the sample surface to avoid unexpected gliding during the monotonic loading procedure. Rectangular samples equipped with aluminum/glass fiber reinforced polymer (GFRP) composite tabs were used according to ASTM D3039/3039M to provide a reduced cross-sectional testing area as well as to protect the gripping section. In the tensile tests, 5 samples were tested for each laminate of the same porosity level. The designs of the specimens are shown in Figure 2.
The fatigue tests were also conducted on the MTS 647 hydraulic servo dynamic material test machine. The specimens were subjected to cyclic loading at a frequency of 7Hz at room temperature. The tests were performed under load control with the load ratio R = 0.1. Specimens with the same geometry as in the tensile tests were used according to ASTM D3479. Specimens remaining intact after application of 1,000,000 cycles were treated as run-outs in this study, and the S-N curves for unidirectional laminates at different void levels in the longitudinal direction were then obtained by plotting the maximum stress and fatigue life in a semi-logarithmic diagram. The dynamic stiffness degradation of composite laminates due to fatigue loading were also analyzed by monitoring the maximum/minimum displacements for each cycle; the residual Young's modulus for samples after the application of a certain number of load cycles can then be calculated based on the difference between the maximum and minimum stress/strain.
SEM Image Analysis
Quasi-static/fatigue fractography observation was also done in this study using a scanning electron microscope (SEM) to get a better understanding of the underlying failure mechanisms. A Denton vacuum coating machine was used before observation, given the poor electrical conductivity of the epoxy matrix. By using a carbon coating technique, a nanoscale carbon coat was formed on the fracture surfaces, thereby allowing electrical charge phenomena to be avoided in SEM observation. The fracture surfaces of samples for longitudinal tensile testing and longitudinal fatigue testing were compared, and the differences between the failure mechanisms were determined.
Results and Discussions
Samples from plaques cured at different compression pressures were analyzed in this section to determine the void content. Fifteen samples cut from 16-ply plaque cured at 0.5MPa and 12-ply plaque cured at 0.3MPa and 0.1MPa (denoted as plaques A, B, and C, respectively) were mounted and polished for observation by the optical microscope; three images per sample were examined in this study. The results from the void content measurement (Figure 3 (a)) show that the samples cured at lower compression pressure have obvious high void content. For the laminates cured at 0.1MPa, the void content is determined as 4.0%, whereas the laminates cured at 0.5MPa have a void content of only 0.8%. By plotting the aspect ratio and equivalent diameter of each void in a diagram (Figure3 (b)), the relationship between size and shape of voids can be investigated. As can be seen in the figure, most of the points lie on the right side of the red line, which suggests that smaller voids tend to be closer to spherical (aspect ratio close to 1). This result is consistent to what can be observed in the cross-sectional photos (Figure 1.), the interlaminar voids are usually larger in size and have a larger average aspect ratio compared to intralaminar voids. As the interlaminar strength is depend on the strength of matrix between fiber tows, the presence of the interlaminar voids may has an impact on the interlaminar shear strength (ILSS) and fatigue strength, which can be investigated in future study.
The longitudinal axial stiffness and tensile strength were measured for unidirectional composite laminates with voids at 0.8%, 2.2%, and 4.0%. The experimental results for the quasi-static tensile tests are presented in Table 2. Five specimens were tested for each lay-up and void content, with the stiffness and strength calculated by averaging the test data. According to the experimental results, the porosity seems to have no impact on the longitudinal stiffness of the composite, while the static strength of the laminates is affected by the presence of voids: the tensile strength drops by around 5% as the void content increases from 0.8% to 4.0%.
Longitudinal tension-tension fatigue tests for 16-ply 0[degrees] UD laminate with 0.8% voids and 12-ply laminates with 2.2%/4% voids were conducted in this study. The original S-N diagram and the S-N diagram with the maximum stress normalized by static tensile strength are presented in Figure 4 (a) and (b), respectively. According to the test results, the relationship between applied stress and fatigue life of the samples can be expressed as linear functions within the applied load range. The longest fatigue life for a certain load level can be found, as can be expected, for the laminates with void content at 0.8%, while the laminates with voids at 4% show the worst fatigue resistance. Moreover, a difference of about 100MPa can be observed between the fatigue strength ([N.sub.f]=[10.sup.6]) of specimens with void content at 0.8% and 4%, and the fatigue life of specimens with void content at 2.2% lies between that of the other two groups. It can be concluded that the presence of voids does promote the micro-cracks propagation rate in the material, thus accelerating the fatigue damage evolution process; however, the S-N curves of laminates at all three porosity levels seem to close toward each other as the load level decreases from 85% to 60%, which may mean that the voids' effects on fatigue of composite laminates are smaller for lower load levels in comparison to high levels.
Figure 5 shows the hysteresis loops for the 100th, 1000th, and 3000th cycles in the longitudinal fatigue test at an axial load of 16.7 KN. It can be easily noticed that in the load control test, a translational displacement as well as a continuous changing in the slopes of the hysteresis loops can be observed as the number of cycles increases; this is mainly due to the stiffness degradation in the loading direction as a result of the micro-cracks propagation in the matrix as well as sliding in the grips area. In this study, the dynamic modulus for each cycle was also recorded by calculating the differences between the maximum and minimum axial loading/displacements for the cycle. Figure 6 demonstrates the dynamic stiffness degradation for longitudinal fatigue testing of unidirectional laminates with void content at 0.8% at axial stresses of 82%, 75%, and 68% of the ultimate tensile strength. As can be seen in the figure, the axial modulus reduction rate of the material is very sensitive to the load level and number of cycles. Compared with the other cases, the sample tested at a stress level of 82% shows the most rapid decrease in the axial modulus with respect to the number of cycles until failure occurs, while a much slower stiffness degradation process can be observed for the sample tested at a stress level of 75%. Different from that of the test at u=82%, after a rapid stiffness degradation stage through the first hundreds of cycles, the modulus reduction rate of the samples tested at u=75% show an obvious deceleration in the medium stage before accelerating again at the very end of the test; a similar trend can also be seen in the stiffness degradation curve of the specimen tested at 68% of the tensile strength, which has the lowest stiffness degradation rate among the three specimens. This result is similar to the previous research on the fatigue failure of fibrous composites, and the failure mechanisms have been explained by a three-stage model. During the first stage of fatigue, a multitude of microscopic transverse cracks initiate either from (1) single filament breakage randomly distributed in the material or from (2) defects such as a weak fiber/matrix interface and voids, which directly lead to a rapid axial modulus reduction; in the second stage, the dominant failure modes are delamination and fiber/matrix interface debonding, and the stiffness reduces slowly and progressively before the final stage, where more severe failure modes occur (such as unstable delamination and fiber breakage), and a sudden drop in axial modulus can be observed.
The voids' effect on stiffness degradation of CFRP composite laminates was evaluated by comparing the modulus reduction at specified load levels with respect to load cycles for specimens with different void contents. The dynamic modulus reduction curves at load level of 0.7 for laminates with void contents at 2.2% and 4% are presented in Figure 7 (a), and Figure 7 (b) demonstrates the stiffness degradation at load level of 0.8 for laminates with void contents at 0.8% and 2.2%. As expected, the presence of voids is proved to accelerate the stiffness degradation process in both cases. As void content increases, the axial modulus decreases more rapidly until reaching the fracture modulus, thus leading to a shorter fatigue life. Since the microscopic voids can act as stress concentrators in the polymer matrix, this phenomenon can be explained by the promotion effect of the voids on the initiation and propagation of transverse cracks; therefore, the crack density is able to reach saturation in a smaller number of cycles for laminates at higher porosity levels.
Finally, it needs to be noted that although the stiffness degradation is very sensitive to the void content, the modulus at rupture seems not to be affected by the presence of voids, as shown in Figure 7 (b).
Failure Mechanism Analysis
In this study, the fractographic results for the longitudinal static tensile tests and fatigue tests of the samples were compared and discussed based on SEM images of the fracture surfaces. The images of fracture surfaces of the specimens that failed under longitudinal monotonic loading and cyclic loading are presented in Figure 8 (a) and (b), respectively. A relatively "flat" fracture surface can be observed in the sample that failed under monotonic loading, with most of the fibers breaking near the primary crack surface, whereas the sample that failed under cyclic loading shows a more rough fracture surface, with large quantities of fiber pull-outs and fiber/matrix interface debondings visible. The most probable reason for this is that the transverse micro-cracks initiated in the first stage of the fatigue test are more likely to deflect into the longitudinal direction, which will form interfacial debondings, or go across the fiber overhead to form fiber-bridging cracks. Both of these cases will lead to the fiber pull-outs at the final failure.
Residual Stiffness Model
The dynamic stiffness of composites, as the preceding test results show, degrades continuously and progressively as the number of cycles increases. Therefore, the fatigue damage can be described by the initial and residual stiffness of the material from the macro perspective. The most common method is to define the fatigue damage index D by the initial modulus, the residual modulus, and the final modulus at a certain load level (see equation (1)).
D = [[E.sub.0] - E(N)]/[[E.sub.0] - E([N.sub.f])] (1)
Where [E.sub.0] indicates the initial modulus without damage, E(N) indicates the residual modulus of the Nth cycle, [N.sub.f] is the fatigue life.
In order to describe the aforementioned stiffness degradation features due to fatigue loading, H. Mao proposed a typical residual stiffness model for CFRP composites that can well reflect the three-stage characteristics of the stiffness reduction trend in the composites fatigue process. R. Qiu  modified Mao's model  by introducing load level u into the equation. For the purpose of reflecting the voids' effect on the stiffness reduction, a residual stiffness model is established in this paper by introducing void content into the preceding modified model, as shown in equation (2).
[mathematical expression not reproducible] (2)
Where A is a constant relating to the properties of fiber/matrix as well as fiber volume content. u refers to the load level, which is defined by the applied load range normalized by static tensile strength. [v.sub.v] stands for the void content. [delta], [[phi].sub.1], and [[phi].sub.2] are functions of load level u and void content [v.sub.v]; the expression of the functions can be obtained through curve fitting of the stiffness degradation data from the tests. Quadratic polynomial equations are used in this paper, and the charts of the functions [delta], [[phi].sub.1] and [[phi].sub.2] are shown in Figure 9.
The fatigue life [N.sub.f] in equation (2) can be predicted according to the model proposed by Shokrich [19, 20] and Harris . As shown in the S-N diagram, translational displacement and rotation can be observed in the S-N curve for laminates at different porosity levels. For the purpose of taking the voids' effect into consideration, the constant in the conventional model is replaced by the function of void content. The modified expression is as shown in equation (3).
[mathematical expression not reproducible] (3)
Where a = ([[sigma].sub.max] - [[sigma].sub.min])/2[[[sigma].sub.t]], q = ([[sigma].sub.max] + [[sigma].sub.min])/2[[[sigma].sub.t]], c = [[[sigma].sub.c]/[[sigma].sub.t], [[[sigma].sub.t]] is tensile strength, [[[sigma].sub.c]] is compression strength. f is a constant. [gamma]([v.sub.v]) and k([v.sub.v]) are functions of void content [v.sub.v]. [gamma]([v.sub.v]) is the intercept of the S-N curve on the vertical axis, which describes the translational displacement due to porosity, while k([v.sub.v]) is the slope of S-N curve, which indicates the rotation of the S-N curve as a result of the presence of voids.
By combining equations (1) and (2), the normalized residual stiffness at the Nth cycle can be predicted by equation (4). Figure 10 shows the comparison between the experimental data on the longitudinal stiffness degradation for the laminates at the porosity levels of 0.8% and 4% and the prediction results of the present residual stiffness model. A good correlation can be observed between the test data and the prediction results for samples with 0.8% void content under both 75% and 80% of the ultimate tensile strength. As for samples with 4% voids, the predicted residual stiffness shows a more rapid degradation with respect to number of cycles, this may be caused by the scatter in voids distribution or scatter in fatigue life of the material.
[mathematical expression not reproducible] (4)
The voids' effect on the mechanical properties of CFRP laminates was investigated in this paper. The following conclusions can be made:
1. The porosity level of CFRP composite laminates with different lay-ups fabricated through compression molding was found to be related to the compression pressure; low compression pressure would induce more voids. The results of morphological analysis of the voids suggested that smaller voids tend to be closer to spherical.
2. Tension-tension fatigue tests for composite laminates at different void levels indicated that the fatigue life of CFRP laminates is very sensitive to the porosity level in both the longitudinal and transverse directions of the UD laminates, and the transverse properties, including fatigue life and static strength, were found to be more sensitive to void content in comparison with the longitudinal properties. The SEM image analysis results of the fatigue fracture surfaces showed that fiber/matrix interfacial debonding and delamination were more likely to occur under cyclic loading.
3. A residual stiffness model for porous CFRP laminates was established that is able to evaluate the voids' effect on the fatigue behaviors of the material. The prediction results on the stiffness degradation for UD laminates at different porosity levels showed a good correlation with the experimental data.
(1.) Vassilopoulos, A., Keller, T., "Fatigue of Fiber-reinforced Composites," Springer-Verlag London, Limited, 2011.
(2.) Bathias, C., "An engineering point of view about fatigue of polymer matrix composite materials," International Journal of Fatigue, 28: 1094-1099, 2006, doi:10.1016/j.ijfatigue.2006.02.008.
(3.) Song, J., Wen, W., et al, "Effects of temperature and fiber volume fraction on mechanical properties of T300/QY8911-IV composites," Journal of Reinforced Plastics and Composites, 34(2): 157-172, 2014, doi:10.1177/0731684414565939.
(4.) Allah, M., Abdin, E., et al. "Effect of fibre volume fraction on the fatigue behaviour of GRP pultruded rod composites," Composites Science and Technology, 56(1): 23-29, 1996, doi:10.1016/0266-3538(95)00125-5.
(5.) Zhang, W., Zhou, Z., et al, "A fatigue damage meso-model for fiber-reinforced composites with stress ratio effect," Materials and Design, 107: 212-220, 2016, doi:10.1016/j.matdes.2016.06.040.
(6.) Shuler, S., Wu, X., "Influence of Loading Frequency on the Room Temperature Fatigue of a Carbon Fiber/SiC Matrix Composite," Journal of the American Ceramic Society, 76(9): 2327-2336, 1993, doi:10.1111/j.1151-2916.1993.tb07772.x.
(7.) Xiao, X., "Modeling of Load Frequency Effect on Fatigue Life of Thermalplastic Composites," Journal of Composite Materials, 33(12): 1141-1158, 1999.
(8.) Huang, X., "Fabrication and Properties of Carbon Fibers," Materials, 2: 2369-2403, 2009, doi: 10.3390/ma2042369.
(9.) Chambers, A., Earl, J., et al. "The effect of voids on the flexural fatigue performance of unidirectional carbon fibre composites developed for wind turbine applications," International Journal of Fatigue, 28: 1389-1398, 2006, doi:10.1016/j.ijfatigue.2006.02.033.
(10.) Sisodia, S., Gamstedt, E., et al. "Effects of voids on quasi-static and tension fatigue behaviour of carbon-fibre composite laminates," Journal of Composite Materials, 49(17): 2137-2148, 2015, doi: 10.1177/0021998314541993.
(11.) Zhang, A., Lu, H., et al. "Effects of voids on residual tensile strength after impact of hygrothermal conditioned CFRP laminates," Composite Structures, 95: 322-327, 2013, doi:10.1016/j.compstruct.2012.08.001.
(12.) Hu, Y., Lang, A., et al. "Hygrothermal aging effects on fatigue of glass fiber/polydicyclopentadiene composites," Polymer Degradation and Stability, 110: 464-472, 2014, doi:10.1016/j.polymdegradstab.2014.10.018.
(13.) Zhang, J., Sun, C., et al. "Experiment Research of Environment Effects on Fatigue Life of Carbon/Bismaleimide Composite Laminates with Central Hole," Key Engineering Materials, 452-453: 525-528, 2011, doi: 10.4028/www.scientific.net/KEM.452-453.525.
(14.) Huang, H., Talreja, R.. "Effects of void geometry on elastic properties of unidirectional fiber reinforced composites," Composites Science and Technology, 65: 1964-1981, 2005, doi:10.1016/j.compscitech.2005.02.019.
(15.) Zhu, H., Wu, B., et al. "Influence of Voids on the Tensile Performance of Carbon/epoxy Fabric Laminates," Journal of Materials Science & Technology, 27(1): 69-73, 2011, doi: 10.1016/S1005-0302(11)60028-5.
(16.) Wisnom, M., Reynolds, T., et al. "Reduction in Interlaminar Shear Strength by Discrete and Distributed Voids," Composites Science and Technology, 56: 93-101, 1996, doi: 10.1016/0266-3538(95)00128-X.
(17.) Rui, Q., "Research on Fatigue Life Prediction and Damage Analysis of 2.5D Woven Composites," Nanjing University of Aeronautics and Astronautics, Nanjing, China, 2013.
(18.) Mao, H., Mahadevan, S., "Fatigue damage modeling of composite materials," Composite Structures, 58: 405-410, 2002.
(19.) Shokrich, M., Lessard, L., "Progressive Fatigue Damage Modeling of Composite Materials, Part I: Modeling," Journal of Composite Materials, 34: 1056-1080, 2000.
(20.) Shokrich, M., Lessard, L., "Progressive Fatigue Damage Modeling of Composite Materials, Part II: Material Characterization and Model Verification," Journal of Composite Materials, 34: 1081-1116, 2000.
(21.) Harris, B., Gathercole, N., et al. "Life-prediction for constant-stress fatigue in carbon-fibre composites," Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 355: 1259-1294, 1997, doi:10.1098/rsta.1997.0055.
This work was supported by University Research Program funded by Ford Motor Company (grant number 2013-5097R).
Haolong Liu and Weidong Wen
Nanjing University of Aeronautics and Astronautics
Xuming Su and Carlos Engler-Pinto
Ford Motor Company
University of Michigan
Table 1. Manufacturing parameters of the CFRP plaques. Types Lay-up Thickness Compression pressure /mm /MPa UD laminates [[0[degrees]].sub.12] 1.53 0.1/0.3 [[0[degrees]].sub.16] 2.03 0.5 Table 2. Quasi-static tensile results for composite laminates at different porosity levels. Stiffness /GPa Tensile strength /MPa Lay-up Void Average Standard Average Standard content modulus deviation UTS deviation [.sub.16] 0.8% 114.7 0.89 1402 53.9 2.2% 114.9 1.57 1377 79.1 [.sub.12] 4.0% 114.3 1.69 1332 66.3
|Printer friendly Cite/link Email Feedback|
|Author:||Liu, Haolong; Wen, Weidong; Su, Xuming; Engler-Pinto, Carlos; Kang, HongTae|
|Publication:||SAE International Journal of Materials and Manufacturing|
|Date:||May 1, 2017|
|Previous Article:||Re-Design for Automotive Window Seal Considering High Speed Fluid-Structure Interaction.|
|Next Article:||Compression Behavior of DIN C10C and SAE-AISI 1010 Steels During Riveting of Clutch Disc Spacer Bolts: Experimental and Computational Analysis.|