Effect of hygrothermal aging on moisture diffusion and tensile behavior of CFRP composite laminates

Yong DU, Yu’e MA,*, Weno SUN, Zhenhi WANG

a School of Aeronautics, Northwestern Polytechnical University, Xi’an 710072, China

b School of Mathematics and Statistics, Northwestern Polytechnical University, Xi’an 710129, China

KEYWORDS CFRP;Hygrothermal aging;Moisture diffusion;Stacking sequence;Tensile behavior

Abstract Carbon Fiber Reinforced Polymer (CFRP)composites are widely used in aircraft structures,because of their superior mechanical and lightweight properties.CFRP composites are often exposed to hygrothermal environments in service.Temperature and moisture can affect the material properties of composites.In order to make clear the moisture diffusion behavior and the properties degradation of composites, the TG800/E207 composite laminates with four stacking sequences[0]16, [90]16, [±45]4s, and [(+45/0/0/–45)s]s are designed and manufactured.Moisture absorption tests are carried out at 80 ℃,90%RH.It is shown that the moisture absorption curves of composite laminates present a three-stage.A modified Fickian model was proposed to capture the diffusion behavior of TG800/E207 composite laminates.The relationships among the non-Fickian parameters, the environmental parameters and the stacking sequences of CFRP were correlated and compared.Results showed that the modified Fickian curve is sensitive to the diffusivity of Stage I and Stage II.Compared with unaged specimens,the maximum tensile stress for[0]16,[90]16,[±45]4s,and[(+45/0/0/–45)s]s decreased by 14.94%,28.15%,11.96%,and 26.36%,respectively.The strains at failure for[0]16,[90]16,[±45]4s,and[(+45/0/0/–45)s]s decreased by 55.38%,62.65%,46.41%,and 31.71%,respectively.The elastic modulus for[0]16,[90]16,[±45]4s,and[(+45/0/0/–45)s]s increased by 90.93%,94.57%,49.22%,and 8.22%,respectively.[90]16 sample has the minimum saturated moisture content and the maximum strength degeneration.

1.Introduction

Carbon Fiber Reinforced Polymer (CFRP) composites have been widely used in aerospace and aeronautic fields due to their high strength-to-weight ratio, high stiffness, weight reduction and long fatigue life.1–3The complex environments such as high temperature, moisture, chemical etching and ultraviolet radiation can cause irreversible physical and chemical changes in CFRP,which have significantly influence on the structural integrity and lifetime performance.4–6Temperature and moisture are the main environmental factors and can lead to performance degradation of composite material, and even material failure.7Because there is no moisture absorption behavior of carbon fiber, the moisture absorption only occurs in the resin matrix and causes the local density of the material to change.This leads to stress redistribution and delamination between matrix and carbon fiber.The moisture absorption and hydrolysis of the resin can lead to the decrease of their molecular weight.8,9Moreover, the mismatch of hygrothermal expansion coefficients of the carbon fiber and the resin matrix will induce the difference of swelling deformation and generate the hygrothermal residual stress.The interfacial properties between fiber and matrix may be weakened and even induce microcracks and delamination.10–12It is essential to assess the diffusion behavior and the performance degradation of CFRP subjected to hygrothermal environment.

Previous studies have indicated that moisture can diffuse into the composites when exposed to hygrothermal environment, and high temperature accelerates the diffusion process of moisture.13–15The process of moisture diffusion of composites can be divided into three steps.16,17The first step is that the water molecules diffuse through matrix and collect in pore and micro-crack.The second step is that the presence of water in polymer structure increases the molecular structure spacing of the resin and forces the material expansion.And the last step is that the cracks and other damages appear inside matrix due to the diffusion of water molecules, which eventually increase the hygroscopic capacity.Many moisture diffusion models including the Fickian model,18–20the non-Fick model,21the Langmuir-type model22and the time-varying diffusion coefficient model23can be used to describe the diffusion behaviors.Fickian model is the most commonly used model and assumes that the diffusivity of material is constant in the hygrothermal environment.However, non-Fickian behavior was observed in some composites.24–26For instance, Bao et al.21found that the moisture diffusion behavior is dependent on resin type and proposed a two-stage diffusion considering the relaxation of bismaleimide induced by absorbed moisture.Wong et al.27analyzed the relationship between the non-Fickian parameters and the thickness, and developed a thickness-dependent moisture absorption model to describe moisture absorption behavior of the epoxy-based molding compounds.Placette et al.28tested the absorption and desorption of six epoxy mold compounds,and developed a dual-stage diffusion model including Fickian and non-Fickian processes.It is shown that saturated moisture concentration is independent of temperature, and non-Fickian behavior plays a large role in the moisture absorption with the increasing time.For the polymer composite, the moisture diffusion is affected by material thickness, experimental parameters, fiber arrangement and type of fiber and resin.29–31

Moisture absorption characterization and its effects on mechanical behaviors of CFRP have been carried out by many researchers.With the same temperature, the maximum moisture absorption content of CFRP increases with the increasing relative humidity.32With the same relative humidity,the maximum moisture absorption content of CFRP increases with the increasing temperature.33Hygrothermal aging has a significant influence on the mechanical properties of CFRP.34In most of the cases, tension, compression, interfacial shear properties of CFRP decreased significantly after hygrothermal aging.35–42However,Gao et al.3investigated the tensile strength degradation of T700/HT280 composite in 80 ℃/85 %RH (Relative Humidity)and found that the tensile strength increases at first,and then decreases with the aging time increases.So, the detailed mechanism of hygrothermal aging is still inconclusive.Thus,the performance degradation in composites is needed for further exploration.

For CFRP,most of the work available in the literature only deals with composite laminates with one stacking sequence.Few studies in literature have been developed to describe the diffusion behavior and the hygrothermal aging effect of composite laminates with different stacking sequences.In this study,CFRP [0]16, [90]16, [±45]4s, and [(+45/0/0/–45)s]slaminates were exposed to hygrothermal environment of 80 ℃and 90%RH for accelerated aging.Damage mechanisms during hydrothermal aging were observed and analyzed.A Modified Fickian model was developed to further explore the moisture diffusion behaviors of CFRP composites.The tensile behavior of sample after hygrothermal aging was investigated to evaluate their performance degradation.The changes in the maximum tensile stress, the strain at failure and the elastic modulus due to hygrothermal aging were analyzed and discussed.

2.Experimental study

2.1.Material and specimen preparation

Domestic TG800 carbon fibers and E207 thermosetting epoxy resin were used as raw materials to manufacture CFRP composite laminates.Firstly,TG800 carbon fibers and E207 epoxy resin were stacked and fabricated into a prepreg of 0.125 mm using hot compression technique.The volume fraction of epoxy resin of TG800/E207 prepreg is (33 ± 2)%.Secondly,2.0 mm thick TG800/E207 composite laminates of the sixteen layers were laid and cured at 150 ℃for 4 h in the vacuum bags.The main parameters of carbon fiber and resin are shown in Table 1.In order to study the stacking sequence effects on the moisture absorption and mechanical behavior of TG800/E207 composite laminates, [0]16, [90]16, [±45]4s, and[(+45/0/0/–45)s]slaminates were selected.Six specimens per stacking sequence were prepared.The dimensions of the laminated composites are given in Fig.1.

2.2.Moisture absorption tests

Table 1 Material properties of TG800 carbon fiber and E207 epoxy matrix.

Fig.1 Specimen geometry for tensile test.

2.3.Tensile tests

Tensile tests of unaged and aged specimens were carried out using MTS 810 tester with a 100 kN load capability.Three replicate specimens were tested at a constant crosshead speed of 1 mm/min according to ASTM D3039.44This assessment was performed by comparing tensile properties of aged specimens with unaged specimens.The maximum stress, the strain at failure, and the elastic modulus for each sample were reported and compared.Failure modes and fracture surfaces were analyzed and compared.

3.Results and discussion

3.1.Moisture absorption behavior

The moisture absorption behaviors of TG800/E207 composite laminate samples with four different stacking sequences are shown in Fig.2.The relations between the moisture absorption content and the square root of absorbing time are shown in Fig.2(a).The moisture absorption content of all specimens increased with an increasing absorbing time.The saturated moisture content is reached when the time is 1608 h.It can be seen that the moisture absorption curves of TG800/E207 composite laminates present a three-stage.In the first stage (Stage I)—initial linear part (0–192 h),moisture content increased with increasing absorbing time.In the second stage (Stage II)—rapid linear part (216–696 h), rapid moisture absorption was observed.In the final stage (Stage III)—saturation part (720–1608 h), moisture absorbing rate decreased gradually, and finally the moisture absorption reached saturation.As shown in Fig.2(b), the saturated moisture content of [0]16, [90]16, [±45]4s, and[(+45/0/0/–45)s]sare 0.870 %, 0.806 %, 0.843 %, and 0.876 % respectively.Compared with [0]16, the differences of saturated moisture content of [90]16, [±45]4s, and[(+45/0/0/–45)s]sare –7.36 %, –3.10 %, and + 0.69 %,respectively.

3.2.Moisture diffusion model

3.2.1.Fickian model

Based on the Fickian law, the diffusivity D, can be obtained from the initial linear part of the Fickian diffusion curve(Mt/M∞≈0.6).The mathematical equations of D can be calculated as45

Fig.2 Moisture absorption behavior of TG800/E207 composite laminates with different stacking sequences.

where M∞is the saturated moisture content and is obtained directly from Fig.2(b); h is the half thickness of laminated composites, M1and M2are the moisture concentration of the two points at time t1and t2, respectively.

Both Stage I and Stage II are linear in Fig.2(a).DIand DIIare defined as the diffusivities calculated by Stage I and Stage II, respectively.The diffusivities of Stage I and Stage II in moisture absorption curves of TG800/E207 composite laminates are listed in Table 2.It can be found that diffusivities are increasing with absorbing time for TG800/E207 specimens regardless of stacking sequence.

According to Fick’s second law, the moisture content of composite is a function of time and can be calculated as46

Table 2 Diffusivities of Stage I and Stage II in moisture absorption curves of composite laminates.

Fickian diffusion curves calculated by the diffusivities of Stage I and Stage II from Table 2 were presented in Fig.3.It can be seen that Fickian diffusion curve calculated by DIand DIIhas a significant difference with experimental data.For the Fickian curve based on DI, there exists a nearly perfect fit with the experimental data for TG800/E207 specimens in the initial linear part of the Fickian diffusion curve (Mt/M∞<0.3, Stage I).When the moisture absorption rate reaches about 70 % of the equilibrium moisture absorption rate (0.3 ≤Mt/M∞<0.7, Stage II), experimental data are significantly higher than that of Fickian diffusion curves.This is because that the water molecules enter the material quickly through the voids and cracks in the early stage of moisture absorption.With the increasing absorbing time,the polymer in the material will undergo chemical changes,and the material solidifies again to produce a large number of hydrophilic groups.47However, Fickian diffusion curves of [(+45/0/0/–45)s]sare slightly higher than experimental data at Stage II shown in Fig.3(d), which may be due to the effects of stacking sequence.Finally, the difference between curves and experimental data is even greater and the diffusivity of moisture absorption further increases.In addition, large differences are also observed for Fickian diffusion curves based on DIIcompared with experimental data.Therefore, the moisture absorption of TG800/E207 composite laminates cannot be described only by the diffusivities of Stage I or Stage II (DIor DII) in the hygrothermal environment.

Late in the autumn, when the weather was rough, windy, and wet,and the cold penetrated through the thickest clothing, especially atsea, a wretched boat went out to sea with only two men on board, or,more correctly, a man and a half, for it was the skipper and hisboy

Fig.3 Moisture absorption curves of composite laminates with different stacking sequences.

3.2.2.Modified Fickian model

As shown in Fig.3,the moisture absorption curves of TG800/E207 composite laminates at the 80 ℃, 90 %RH show three stages which are unrelated to stacking sequence.When Mt/M∞<0.3, the absorption curves agree with Fickian diffusion.When 0.3 ≤Mt/M∞<0.7,the moisture absorption of TG800/E207 composite laminates increases linearly and deviates gradually from the Fickian behavior.When 0.7 ≤Mt/M∞, the moisture absorption curve increases to equilibrium and reaches the maximum moisture absorption.Based on the results in Fig.3, the modified diffusion model was proposed to better describe the diffusion behavior of TG800/E207 composite laminates, as

where M(t) is the moisture content at time t; the subscripts I,II, and III are Stage I, Stage II and Stage III, respectively.MI(t), MII(t) and MIII(t) are the moisture contents over time in Stage I,Stage II and Stage III, respectively, and can be calculated as

where a and k are unknown parameters to be determined experimentally; T is temperature; Troomis room temperature.In this model, the concentration-gradient plays a central role in Stage I, and then Stage II is dominated by relaxation of the polymer, and finally Stage III is primarily influenced by aging temperatures.However, the modified Fickian model is also influenced by stacking sequence of composites.Moisture diffusion parameters of TG800/E207 composite laminates were listed in Table 3.The modified Fickian model curves and the experimental data were shown in Fig.4.It can be seen that the modified Fickian model shows a nearly perfect fit with the uptake experimental data of TG800/E207 composite laminates.The modified Fickian model can effectively describe the three-stage diffusion process of TG800/E207 composite laminates with different stacking sequences.

3.3.Damage mechanisms of moisture absorption

Fig.4 Modified Fickian model curves and experimental data of composite laminates with different stacking sequences.

Table 3 Moisture diffusion parameters of TG800/E207 composite laminates.

Hygrothermal aging damage of [0]16, [90]16, [±45]4s, and[(+45/0/0/–45)s]ssamples at different aging stages are shown in Figs.5(a)–(d), (e)–(h), (i)–(l), and (m)–(p) respectively.At the unaged stage shown in Fig.5(a), there are no obvious delamination, cracks or other damages on the surface of [0]16samples.In Fig.5(b), the surface of [0]16sample is similar to that of the unaged sample,and the surface is relatively smooth.In Fig.5(c), the micro-cracking occurs and spreads in the x-direction.In Fig.5(d), the micro-cavities occur successively after the micro-cracking damage.For [90]16sample, there are no obvious damage in Figs.5(e) and (f).Micro-cracking and micro-cavities are observed in Fig.5(g) while micro-cracking spreads in the x-direction.In Fig.5(h), micro-cracking and micro-cavities occur more frequently,and the damage number of which increases gradually in the aging time.For [±45]4ssample in Fig.5(k), micro-cracking first occurs and spreads in the inclined direction, which is different from [0]16samples in Fig.5(c) and [90]16samples in Fig.5(g).At the later stage of aging shown in Fig.5(l), the inclined micro-cracking expands and the new micro-cracking appears in the x-direction.The main hygrothermal aging damage of[(+45/0/0/–45)s]ssample is micro-cracking and microcavities.Micro-cracking damage occurs the earliest and produces the most damage in Figs.5(o) and (p).Micro-cavity damage emerges late and damage is less in Fig.5(p).

Fig.5 Hygrothermal aging damage of TG800/E207 samples at different aged stages.

In Stage I, the resin matrix only absorbs moisture to increase the flexibility of the molecular chain, leading to the plasticization of the resin.At this time,the hygrothermal aging effect on the mechanical properties of the composite material is reversible.If the samples saturated were dried, the mechanical properties of the samples saturated will return to the original state.In Stage II, micro-cracks and micro-holes appeared on the surface of the sample.The new damage induced by hygrothermal aging will cause further the increase of moisture absorption rate in Stage II.This is also the reason why the diffusivity of Stage II is greater than that of Stage I in Table 2.The damage is irreversible at this time.47With the further progress of moisture absorption in Stage III,the damage increases gradually and expands on the surface of the sample.

3.4.Mechanical properties

3.4.1.Stress–strain curves

The stress–strain curves of aged and unaged specimens with different stacking sequences are shown in Fig.6.In Fig.6(a),it is obvious that the stress–strain curves of [0]16samples increase linearly up to the point of specimen failure whether hygrothermal age or not.The maximum average stress of the unaged sample is 2240.23 MPa, which is 334.72 MPa higher than that of the aged sample.In Fig.6(b),the maximum average stress of unaged and aged [90]16samples are 61.37 MPa and 44.09 MPa, respectively.The maximum average stress of[90]16samples is reduced by 28.16 % after saturated moisture absorption.In Fig.6(c), all curves of [±45]4ssamples have almost linear behavior up to the strain of 0.04 %, followed by the onset and accumulation of damages with the increasing of loading.The maximum average stress of aged[±45]4ssamples is 28.85 MPa lower than that of unaged [±45]4ssamples.The strain of the aged samples at failure is about half that of the unaged samples.In Fig.6(d), the stress–strain curves of[(+45/0/0/–45)s]ssamples have the same tendency as that of [0]16and [90]16samples.The failure processes of[(+45/0/0/–45)s]ssamples are sudden brittle fractures.The maximum average stresses of the unaged [(+45/0/0/–45)s]ssamples and the aged [(+45/0/0/–45)s]ssamples are 1346.45 MPa and 991.54 MPa, respectively.The maximum average stress of [(+45/0/0/–45)s]ssamples is reduced by 26.36 % after the saturated moisture absorption.

Results from tensile tests of unaged and aged specimens are summarized in Table 4.Compared with the mechanical properties of unaged specimens, the decreases of the maximum stresses for [0]16, [90]16, [±45]4s, and [(+45/0/0/–45)s]sare 14.94 %, 28.15 %, 11.96 %, and 26.36 %, respectively.The decreases of the strains at failure for [0]16, [90]16, [±45]4s,and [(+45/0/0/–45)s]sare 55.38 %, 62.65 %, 46.41 %, and 31.71 %, respectively.The increases of the elastic moduli for[0]16, [90]16, [±45]4s, and [(+45/0/0/–45)s]sare 90.93 %,94.57 %, 49.22 %, and 8.22 %, respectively.Considering the results in Fig.2(b), we can find that [90]16has the minimum saturated moisture content and the maximum strength degeneration.Hygrothermal aging has a significant effect on the tensile mechanical properties of TG800/E207 composite laminates, which is also dependent on the stacking sequences of specimens.

Fig.6 Stress–strain curves of aged and unaged specimens.

Table 4 Summary of tensile mechanical properties of unaged and aged specimens.

In Table 4, it can be concluded that the maximum tensile stress and the strains at failure of all samples after hygrothermal aging decreased severely.This is mainly because that the chemical degradation of material occurs after hygrothermal aging.The chemical degradations including the hydrolysis of the resin and the molecular chain scission dominate the effect of hygrothermal aging on tensile strength.While the elastic modulus of all samples increased after hygrothermal aging.The increases of the elastic modulus of each sample from large to small were [90]16> [0]16> [±45]4s> [(+45/0/0/–45)s]safter hygrothermal aging.The elastic moduli of [0]16and[90]16samples increase more than 90 %.This is because that hygrothermal aging can cause the secondary curing and the post-curing of the resin, which can greatly improve the elastic modulus of the material.The influence of secondary curing and post-curing caused by hygrothermal aging on [0]16and[90]16samples is greater than that of [±45]4sand[(+45/0/0/–45)s]ssamples.

About the effect of hygrothermal aging on elastic moduli of composites, it is still inconclusive according to the published studies.The majority of literature believe that the elastic moduli of composites decrease after hygrothermal aging,2,37,48while Ma et al.38investigated the tensile behavior of T700/MTM46 laminates after hygrothermal aging and found that elastic moduli remain unchanged.Therefore, the effect of hygrothermal aging on elastic moduli of composites may be related to the raw material type and the stacking sequences of composites.

3.4.2.Failure mode

The dominant failure modes of unaged and aged TG800/E207 specimens are shown in Fig.7.It could be seen that there were three typical failure characteristics, namely, fiber breakage,fiber pull out and matrix crack.In Figs.7(a)and(b),the obvious fiber breakage and fiber pull out are observed.The damage of the aged specimens is more serious than that of the unaged specimens, and the fiber bundles of the aged specimens are completely dispersed.In Figs.7(c)and(d),matrix crack occurs on the near loading tab.In Figs.7(e) and (f), fiber breakage and fiber pull out are observed in the middle of the [±45]4ssamples.Fiber fractures present brush shape and the adjacent carbon fiber is not neat.In Figs.7(g)and(h),the failure modes of [(+45/0/0/–45)s]sspecimens are similar with [±45]4sspecimens.The main failure modes are also fiber breakage and fiber pull out.In summary, we can conclude that failure modes of unaged and aged specimens with the same lay-up are essentially the same, but the failure of aged specimens is more serious than unaged specimens.

Fig.7 Failure modes of unaged and aged specimens after tensile failure.

Fig.8 Fracture morphologies of unaged and aged specimens with different stacking sequences.

The fracture morphologies of unaged and aged specimens are shown in Fig.8.In Fig.8(a), most of the fibers were fractured at the same time and the fracture surface is relatively flat at unaged [0]16specimen.In Fig.8(b), the fiber bundles were pulled out and the delamination of fiber/matrix interface was observed at aged [0]16specimen.In Figs.8(c) and (d), matrix fracture is the main failure mode of unaged and aged [90]16specimen.There is a little adhesion-like fiber bundles in Fig.8(c).Compared with unaged specimens, the adhesionlike fiber bundles significantly increase and the micro-cracks appear on the surface of the laminate shown in Fig.8(d).For unaged[±45]4sspecimen shown in Fig.8(e),the main failure modes are delamination and fiber fracture.The fiber fractures are mainly concentrated in the fiber layers on the upper and lower surfaces,while the fiber layer in the middle is mainly delamination.In Fig.8(f), delamination, fiber fractures and matrix fractures were observed at each fiber layer.And burrlike fibers were observed on the upper and lower fiber layers at aged [±45]4sspecimen.As is shown in Figs.8(g) and (h),delamination and fiber fracture are the main failure modes of unaged and aged [(+45/0/0/–45)s]sspecimens.Compared with unaged specimen, multi-fiber fractures occurred on the fiber layer at aged specimen.

4.Conclusions

Effect of hygrothermal aging on moisture diffusion and tensile behavior of TG800/E207 composite laminates with four different stacking sequences ([0]16, [90]16, [±45]4s, and[(+45/0/0/–45)s]s) were investigated.The conclusions can be drawn as follows:

(1)The absorbed moisture content of TG800/E207 composite laminates in 80 ℃, 90 %RH increased with an increasing absorbed time and reached the saturation around 1608 h after hygrothermal aging.The saturated moisture contents of [0]16,[90]16, [±45]4s, and [(+45/0/0/–45)s]sare 0.870 %, 0.806 %,0.843 %, and 0.876 %, respectively.Compared with[0]16, the differences of saturated moisture contents of [90]16,[±45]4s, and [(+45/0/0/–45)s]sare –7.36 %, –3.10 %,and + 0.69 %, respectively.

(2) The moisture absorption behaviors of TG800/E207 composite laminates can be divided into three stages as follows: Stage I is the initial linear part; Stage II is the rapid linear part; Stage III is the saturation part.A modified Fickian model was developed according to the three-stage moisture absorption behavior.The modified Fickian model is dependent on the diffusivity of Stage I and Stage II, while the model can effectively describe the moisture absorption of TG800/E207 composites laminates with different stacking sequences.

(3) For [0]16, [90]16, [±45]4s, and [(+45/0/0/–45)s]ssamples compared with the mechanical properties for unaged specimens, the maximum tensile stress decreased by 14.94 %,28.15%,11.96%,and 26.36%,respectively;the strain at failure decreased by 55.38 %, 62.65 %, 46.41 %, and 31.71 %,respectively; the elastic modulus increased by 90.93 %,94.57 %, 49.22 %, and 8.22 %, respectively.The saturated moisture absorption of [90]16sample is the lowest, but its tensile strength degradation is the most serious.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This study was financially supported by the National Natural Science Foundation of China (Nos.91860128, 12032018, and 52061135101).