Hydrothermal aging of carbon reinforced epoxy laminates with nanofibrous mats as toughening interlayers

Hydrothermal aging of carbon reinforced epoxy laminates with nanofibrous mats as toughening interlayers

Polymer Degradation and Stability 126 (2016) 188e195 Contents lists available at ScienceDirect Polymer Degradation and Stability journal homepage: w...

2MB Sizes 0 Downloads 15 Views

Polymer Degradation and Stability 126 (2016) 188e195

Contents lists available at ScienceDirect

Polymer Degradation and Stability journal homepage: www.elsevier.com/locate/polydegstab

Hydrothermal aging of carbon reinforced epoxy laminates with nanofibrous mats as toughening interlayers Maria Di Filippo a, Sabina Alessi a, Giuseppe Pitarresi a, Maria Antonietta Sabatino a, Andrea Zucchelli b, Clelia Dispenza a, c, *  degli Studi di Palermo, Viale delle Scienze 6, 90128 Palermo, Dipartimento di Ingegneria Chimica, Gestionale, Informatica, Meccanica (DICGIM), Universita Italy b Department of Industrial Engineering (DIN) and Advanced Mechanics and Materials e Interdepartmental Center for Industrial Research (AMM ICIR), University of Bologna, Viale Risorgimento 2, 40136 Bologna, Italy c School of Chemical Science and Engineering, Department of Fiber and Polymer TechnologyeRoyal Institute of Technology (KTH), SE e 100 44 Stockholm, Sweden a

a r t i c l e i n f o

a b s t r a c t

Article history: Received 19 November 2015 Received in revised form 11 February 2016 Accepted 17 February 2016 Available online 18 February 2016

Electrospun mats have been applied as toughening interlayers in high performance carbon fiber epoxy composites. While the toughening mechanism exerted by the mat at the interface is the subject of several recent studies, no investigations are reported on the aging behaviour of laminates comprising these nanostructured elements. This work investigates the influence of the combined effect of water and temperature (90  C) on laminates with Nylon 6,6 electrospun membranes placed either at the middle plane only or at each interlayer. The water-uptake behaviour is modelled by a two-stage diffusion model and compared with the behaviour of the neat resin and of the laminate without mats. Interestingly, a lower water uptake is observed for the laminates with mat-modified interfaces and this is possibly due to a significantly reduced porosity. The effect of hydrothermal aging on the thermal (Tg) and mechanical properties (transverse flexural modulus and interlaminar shear strength) of the various laminates is also investigated. © 2016 Elsevier Ltd. All rights reserved.

Keywords: Carbon reinforced epoxy composites Electrospun nanofibers Hydrothermal aging Two-stage water uptake model Thermo-mechanical properties Interlaminar fracture toughness

1. Introduction Carbon fiber reinforced epoxies are widely used as structural materials in aeronautical and advanced automotive applications, due to the combination of lightweight, high specific mechanical stiffness and strength, excellent chemical and corrosion resistance and flexibility in the manufacturing of complex shapes. They have some limitations too. One of the most common failure mechanisms of these materials is delamination under cyclic or impact loading. Over the years, several approaches have been followed in order to improve the delamination resistance, such as the modification of the epoxy matrix with the introduction of elastomers [1e5] or thermoplastic polymers [6e8] as a dispersed or co-continuous heteroface, or by the addition of various nanofillers [5,9,10].

* Corresponding author. Dipartimento di Ingegneria Chimica, Gestionale, Infor degli Studi di Palermo, Viale delle Scienze matica, Meccanica (DICGIM), Universita 6, 90128 Palermo, Italy. E-mail address: [email protected] (C. Dispenza). http://dx.doi.org/10.1016/j.polymdegradstab.2016.02.011 0141-3910/© 2016 Elsevier Ltd. All rights reserved.

More recently, several research groups have focused their attention on an innovative toughening strategy based on the use of nanofibrous polymeric mats produced by electrospinning. The nanofibrous mats are introduced in composite laminates in between the fabric plies, with the purpose of creating an interlock between adjoined plies. The toughening mechanism is called “Velcro effect”, since it resembles the mechanical interlock between fibers at the basis of the adhesive strength of Velcro. For nanofibrous mats this effect is essentially based on synchronized multiple weak interactions: VdW forces and hydrogen bonds which establish over a large interfacial area generated by the nanofibers that stretch out in the resin, perpendicularly to the ply [11,12]. The nanofibrous mats evaluated for this application can differ for the chemical structure of polymers used, Nylon 6,6, polysulfone, poly(ε-caprolactone) among the most common ones, and for the geometrical characteristics of the nanofibers and their membranes (nanofiber diameter, preferential vs. random alignment, mat thickness, mat density, etc.) [13e15]. When Nylon 6,6 nanofibrous mats are used, significant improvements in mechanical properties

M. Di Filippo et al. / Polymer Degradation and Stability 126 (2016) 188e195

of laminates are documented [12,15e18]. Indeed, increases in mode I and mode II delamination strain energy release rate, GIc and GIIc, and up to a 60% increase in the threshold impact force are reported [18]. No detrimental effects on the thermal properties (glass transition temperature) of the material have been observed, in consideration of the low weight fraction of the thermoplastic material in the system [17]. In evaluating the performance of composite materials, in addition to mechanical and thermal resistance, also durability is of a paramount importance [19,20]. In both aeronautical and automotive applications, temperature cycling and exposure to water are the most common environmental conditions which large portions of structural components experience during their service life [19]. Hydrothermal aging is, therefore, of great significance and it has been already investigated to a great extent [19e32]. Conversely, at the best of our knowledge, no studies are reported on the hydrothermal aging of carbon fiber reinforced composites comprising nanofibrous mats as interlayers. It is well known that water penetrates in polymeric composites following three main pathways: diffusion through the matrix; capillary flow along the imperfect fiber-resin interface; and flow in micro-cracks or voids present in the matrix as a consequence of poor consolidation [21,22]. It is interesting to establish if the presence of a nanofiber-reinforced interface works as a barrier for water transport across the material or it actually favors water absorption and diffusion, hence increasing the susceptibility of the laminate to hydrothermal aging. Indirect effects resulting from changes in the quality of consolidation cannot be ruled out. This work studies the effects of an accelerated hydrothermal aging treatment (at higher temperature for a shorter time) on both thermal and mechanical behaviour of high performance carbon fiber/epoxy composite systems toughened with electrospun Nylon 6,6 nanofibrous membranes. In particular, two lay-ups are considered, one with a single mat placed in the middle plan of a [0]10 reinforced laminate and the other with mats placed at each interlayer. Reference systems are the panel of neat resin and the laminate without mats. Morphological analysis of laminate crosssections is carried out by Scanning electron Microscopy (SEM). ImageJ software is used to estimate the porosity degree of the laminates from SEM micrographs. Water absorption data are fitted by a two-stage diffusion model [29e32] which introduces two empirical parameters, the “apparent” diffusion coefficient and the quasi-equilibrium water-uptake value. The thermal and mechanical properties of the aged composite laminates are investigated by dynamical mechanical thermal analysis (DMTA), short beam shear (SBS) and three point bending (TPB) mechanical tests, at the various stages of conditioning: before immersion in hot water, at saturation, after the first drying step at room temperature and after the second drying step at high temperature.

without further purification. Electrospun non-woven mats are fabricated by using an electrospinning apparatus (Spinbow s.r.l, Bologna, Italy) already described in a previous work [17]. Highly porous mats with 30 mm thickness have been used. The random orientation of nanofibers (fiber diameter distribution 170 ± 30 nm) and micron-size porosity has been confirmed by SEM analysis (Fig. 1). The reinforcement is a unidirectional carbon fiber fabric, UNIC CUT 300/10 HM-U659 10 HM, purchased by Dalla Betta Group s.r.l. The fabric is characterized by a surface density of 300 g/m2 and Young's modulus of 398 GPa. The carbon fiber bundles are held together by polyamide stitching threads. 2.2. Preparation of neat epoxy resin panel and of epoxy resin/Nylon mat/carbon fiber laminates The preparation of the neat epoxy resin panel has been already described in Ref. [17]. Carbon fiber/epoxy [0]10 laminates are fabricated by hand lay up. Three different systems are manufactured: the [0]10 reference system without mats, a system with only one mat in the middle plane (CFN1), and a system with nine mats placed at each interlayer (CFN9). The curing process both for the neat epoxy resin panel and for the all laminates is carried out in a heated hydraulic press, equipped with a temperature control system. The temperature cycle is made of a heating step up to 180  C at 2  C/min, an isothermal plateau at 180  C for 2 h, a slow cooling step to 100  C at 2  C/min and a fast cooling step to room temperature without temperature control. Carbon fiber weight fractions WCF of the [0]10, CFN1 and CFN9 systems, calculated from the known weight of carbon fibers and mats (when present) in the panel and the measured weight of the panel after cure, are reported in Table 1. Average thicknesses of the composite systems is 2.7 ± 0.3, 2.9 ± 0.2 and 3.0 ± 0.1 mm, for [0]10, CFN1 and CFN9, respectively. From each panel, specimens are cut to the required dimensions for the different characterisations and subjected to a post-curing treatment at 200  C for 2 h before use. 2.3. Hydrothermal aging treatment Hydrothermal aging is carried out in three steps on samples with DMTA-type (both neat resin and composites) and SBS-type

2. Materials and experimental procedures 2.1. Materials The epoxy resin used in this work is a blend of a tetrafunctional epoxy monomer, 4,40 -methylenebis(N,N-diglycidylaniline) (TGDDM), and a difunctional epoxy monomer, bisphenol A diglycidyl ether (DGEBA). The curing agent is 4,40 -Diaminodiphenyl sulfone (DDS). All reagents are provided by Sigma Aldrich (Italy). The weight ratio of two monomers TGDDM:DGEBA is 84:16, while the amount of the hardener is 26 phr (per hundred of resin by weight). Nylon 6,6 Zytel® E53 NC010 is used for producing electrospun fibrous mats and it is kindly provided by DuPont. Formic acid (FAc) and chloroform (CLF) are purchased by Sigma Aldrich and are used


Fig. 1. SEM image of electrospun Nylon 6,6 mat.


M. Di Filippo et al. / Polymer Degradation and Stability 126 (2016) 188e195

Table 1 Properties of the carbon reinforced epoxy laminates, with one (CFN1) and nine (CFN9) Nylon 6,6 mats as interlayers and with no mats ([0]10) as reference. System

Carbon fiber weight fraction WCF%

Porosity degree f%

Glass transition temperature Tg [ C]

Transverse flexural modulus Ef [GPa]

Interlaminar shear strength ILSS [MPa]

[0]10 CFN1 CFN9

69 ± 1.4 70 ± 1.5 65 ± 1.5

9.9 ± 0.4 10.8 ± 0.4 5.2 ± 0.2

265 263 263

5.82 ± 0.39 5.39 ± 0.19 5.51 ± 0.07

44.67 ± 0.96 46.05 ± 0.62 59.10 ± 4.21

(only composites) geometry. Detailed description of their dimensions and fiber alignment is provided in the following paragraph. The first step (absorption) is conducted by immersing dry samples (immediately after the post-cure treatment) in deionized water at 90  C, for 4 weeks. After the absorption step, two consecutive desorption steps are carried out: first, the samples are kept in a box with calcium chloride, at room temperature, for 4 weeks; then they are placed in oven at 150  C for 2 days when a constant weight is attained [23]. For the neat epoxy resin specimens the first desorption step is applied for much longer time (approx. 3 months). During the absorption/desorption steps samples are weighted at increasingly longer time intervals. Excess water is removed by blotting paper. The amount of absorbed water, Mt, is calculated according to the following equation:

Mt ¼

Wt  Wi  100 Wi


where, Wt is the weight at time t and Wi is the initial weight. Mt values are the average of six specimens for each system. A two-stage diffusion model is used to fit the data [29e32]. The model considers two stages: the first one is controlled by a concentration gradient, according to Fick's law; the second one is associated with the structural relaxation of the polymeric chains due to water uptake. The mathematical form of this model is given by the following equation:

Mt ¼ M∞0

( "  0:75 #) pffiffi Dt 1 þ k t 1  exp  7:3 2 h


where, Mt is the water uptake at the generic time t, M∞0 the quasiequilibrium uptake of the diffusion-dominated first stage, k a constant related to rate of relaxation of the polymeric structure in the second stage, D the apparent diffusion coefficient and h the specimen thickness. M∞0, D and k are determined both for the neat resin system and the composite laminates. 2.4. Characterizations Dynamic Mechanical Thermal Analysis (DMTA) is carried out on both neat epoxy resin samples (30  8.0  3.0 mm) and composite samples (30  8.0  thickness mm) cut to their length to be perpendicular to the fibers direction (transverse reinforcement samples), using a Rheometrics DMTA V instrument. A single cantilever bending setup is adopted. Tests are performed in a temperature swift mode between room temperature and 300  C, at a heating rate of 5  C/min, under nitrogen flow. The frequency and strain values are set at 1.8 Hz and 0.02%, respectively. The storage modulus E0 and the loss factor tand versus temperature are recorded and the glass transition temperature Tg is determined as the temperature corresponding to the main peak of tand curve. All the mechanical tests are conducted by an Instron 3367 electro-mechanical machine, equipped with either a 1 kN or 30 kN load cell depending on the type of test. Three point bending (TPB) test is carried out on transverse

reinforcement samples. The specimens were 60 mm long and 15 mm wide, according with the guidelines of the UNI EN ISO 1425 standard [33]. Short beam shear (SBS) test is performed on longitudinal reinforcement samples according to the ASTM D 2344 standard [34]. In particular, specimens are characterized by length to thickness ratio equal to 6 and width to thickness ratio equal to 2. For each composite system six samples are tested and the average interlaminar shear strength (ILSS) and standard deviation are calculated. A Philips 505 Quanta 200 field-emission scanning electron microscope (SEM) is used to evaluate the morphology of nanofibrous mat and cross-sections of composite systems. Prior to SEM observation, the composite specimens are etched in a super-acid solution for 5 min. All samples are gold sputtered before testing. The ImageJ software is used for image analysis [35], in order to estimate the porosity degree of the laminates. Image analysis converts the SEM image in a binary image, in which voids are black and filled space (with matrix and carbon fibers) is white. 700 magnification SEM images are used. A “cleaning” process of the binary images to remove unwanted noise is performed. Several specimens from different parts of the laminate panels are observed and analysed, and the average value of the porosity degree, 4, for each system is determined. 3. Results and discussion 3.1. Characterisation of carbon fiber reinforced epoxy resin with Nylon 6,6 nanofibrous mats interlayers In order to evaluate the quality of the consolidation process of the laminates and appreciate the degree of interpenetration between nanofibers and resin, a morphological analysis on samples cross-sections is conducted. A preliminary mild etching treatment is carried out with the purpose of selectively removing the Nylon 6,6 component, thus enhancing the contrast with the resin. The same treatment is also carried out on the [0]10 system to compare systems subjected to the same surface preparation. The nonetched [0]10 laminate (here not shown for brevity) is similar to the etched one, thus confirming that the etching is not aggressive towards the epoxy resin or the carbon fibers. In Fig. 2aec representative SEM micrographs of the various systems are reported. A randomly distributed micro-porosity is evident by [0]10 and CFN1 laminates. Fewer pores are shown by CFN9. The porosity degree values, calculated by image analysis, are reported in Table 1. The void content of CFN9 is halved with respect to [0]10 and CFN1. In CFN1, a well-defined Nylon mat/resin region at the middle plain is clearly evident (see Fig. 2b). In CFN9 instead, the mat/resin interleaves are thinner and some of them are visible only at higher magnifications (see Fig. 2d,e). A close look to the Nylon mat/resin interleaf reveals the fingerprints of randomly and uniformly distributed nanofibers, mainly confined in the interlayer region (Fig. 2e). The observed improvements in laminate consolidation when the mats are placed at each interlayer is likely due to the mats acting as “resin reservoirs”. Indeed, the excess resin present on the pre-impregnated fabrics during hand lay up is absorbed by

M. Di Filippo et al. / Polymer Degradation and Stability 126 (2016) 188e195


Fig. 2. SEM images of transverse sections: (a) [0]10; (b) CFN1; (c) CFN9; (d) zoomed part from (c); (e) enlarged detail of (d). In Figure (b) and (c) the arrows on the right indicate the mat/resin interlayers. In Figure (c) the arrows on the left point to the stitching threads of UD fabric.







Matrix [0]10 [0]10


E' [Pa]

the mat when the laminate is stacked and then released when is pressed for curing. The thermal properties of the neat resin system and carbon








0 0








Temperature [°C] Fig. 3. Storage modulus (E0 ) and loss factor (tand) from DMTA of the neat resin system and the three composite laminates.

fiber/epoxy composites have been investigated by DMTA. Fig. 3 shows the curves of E0 and tand, while the relative glass transition temperatures, calculated as tand peak temperature, are reported in Table 1. With respect to the neat resin, all CF-reinforced composites show an increase of the E0 curve and a decrease of tand peak for the contribution of the high modulus carbon fibers to the storage and dissipative moduli of the resin in the composite. The thermo-mechanical behaviour of the [0]10 and CFN1 systems is similar: no effects on Tg, as a consequence of the introduction of one nanofibrous mat is evidenced (see Table 1). For the CFN9 system, an enlargement of the main peak and the presence of a second small peak, at about 180  C, can be observed, probably due to localised plasticization effects in the matrix, caused by the presence of the mats at each interlayer. The transverse flexural modulus and the interlaminar shear strength values are reported in Table 1 [36]. The addition of Nylon mats (CFN1 and CFN9 systems) does not significantly affect the transverse flexural modulus, while the interlaminar shear strength increases of about 30% for CFN9 system with respect to [0]10. Improvements in the delamination properties of the composites by recourse to electrospun polymeric mats have been already observed and explained with the above described “Velcro effect” [12e17]. In our systems, an indirect contribution from the reduced porosity to the increase of ILSS cannot be ruled out.


M. Di Filippo et al. / Polymer Degradation and Stability 126 (2016) 188e195

3.2. Study of water adsorption and desorption processes in the resin and laminates Preliminarily to the assessment of the hydrothermal aging behaviour of the composite systems, a study of the aging behaviour of the neat epoxy matrix has been performed. In Fig. 4 the water absorption/desorption profile of the epoxy matrix, measured on DMTA-type samples, is reported. In the initial absorption phase, the amount of absorbed water, Mt, increases linearly with the square root of time, indicating a Fickian-type transport mechanism of water into the matrix. Prolonging the immersion time, the rate of absorption gradually decreases but it does not reach a full saturation condition in the explored time window. The first stage of the water desorption process, carried out at room temperature, is slower than the water uptake and it reaches a plateau after about three months, the material retaining almost 50% of the absorbed water. A more drastic thermal treatment, at T ¼ 150  C for about 50 h, eliminates the residual water. The water absorption and desorption profiles of DMTA samples with transverse reinforcement (30  8 mm, DMTA type) for all composite systems, and for the resin as reference, are presented in Fig. 5a,b, respectively. Samples with longitudinal reinforcement (20  6 mm, SBS type) were also subjected to the same hydrothermal aging system, showing a very similar behaviour with respect to the ones with transverse reinforcement (water absorption data here not presented for brevity). This evidence supports the assumption that border effects can be neglected and an isotropic diffusion model can be applied to describe the behaviour of the composite systems. In panel (a), the absorption data for the laminates ([0]10, CFN1 and CFN9 systems) are normalized with respect to the weight fraction of the matrix, so that the values can be directly compared with those of the neat resin, and the lines are the fitting curves determined using Eq. (2). In absorption (Fig. 5a), both resin and composites show a same behaviour: initially, Mt increases linearly with t1/2, then it continues to rise with a constant but lower slope. An equilibrium or saturation state is not reached in the time window observed. In Table 2, D, k and M∞0 values, resulting from the fitting of the curves reported in Fig. 5a using the two-stage model and the coefficient of determination, R2, are reported. The quality of fitting is satisfactory for all systems. The apparent diffusivity value, D, calculated for the neat resin system is close to the values reported in the literature for similar systems [23]. D values for the composite systems are significantly higher than for the neat resin; this


A 5

Desorption @RT

Mt %

4 3


Absorption @90 °C


Desorption @150 °C

1 0 0







Time1/2 [h1/2] Fig. 4. Water absorption/desorption curve of the neat resin system. The solid line is only a guide for the eye.

suggests that capillary transport of water into pores and microcracks collaborates to the diffusional transport of water through the matrix [28,32]. Porosity is highly irregular in shape and not necessarily interconnected, that makes very arduous any attempt to model it. On the other hand, pores and microcracks can act as reservoirs and, can be more or less rapidly filled up with water depending on their accessibility. Water will then diffuse through the matrix under a concentration gradient. For this reason, we attempted to fit the water absorption curves with the same equation used to fit the behaviour of the neat resin, with D being now an “apparent” diffusion coefficient. The data shown in Table 2 indicate that the presence of carbon fibers and Nylon mats influences significantly the apparent diffusion coefficient but not so much the parameter k, which is related to matrix plasticization effects. This evidence also supports the hypothesis that an imperfect fibermatrix interface and the presence of voids in the composites are the main causes of the observed differences in water-uptake. Moreover, M∞0 value is higher for the system with the highest degree of porosity (CFN1), as expected. The fitting parameters determined for water-uptake curves of samples with different geometry (SBS-type), confirm the same trends. Desorption curves (see Fig. 5b) show a faster water release for the systems with higher porosity. A plateau is reached after about 200 h irrespective of system composition. As for the “pure” matrix, a thermal treatment at higher temperature is required to eliminate the residual water. 3.3. Influence of hydrothermal aging on the thermal and mechanical properties of the laminates Firstly, the influence of hydrothermal aging on the neat resin is examined. In particular, Fig. 6 shows storage modulus and loss factor curves from DMTA analysis for samples tested in conditions corresponding to points A, B and C of Fig. 4. E0 and tand curves for the non-aged system are also reported for comparison. The tand curve of the non-aged neat resin shows only one relaxation peak at high temperature (Tg ¼ 264  C), while after immersion at 90  C for 4 weeks (Point A) it shows two broad peaks, at about 190  C and 250  C. These two broad relaxations suggest the co-existence of material portions or domains with different crosslinking densities. The observed behaviour can be explained with an uneven distribution of the absorbed water into the epoxy network, inducing a non uniform swelling and a different extent of plasticization in the different domains, and/or to the occurrence of degradation phenomena [27]. The first effect should be reversible upon drying, while the latter is irreversible. After the first step of water desorption (Point B of the curve in Fig. 4), a main peak with a shoulder is evident in the tand curve, indicating only a partial recovery of the initial conditions. After the thermal treatment at 150  C, a full recovery of the thermomechanical behaviour of the matrix is achieved (with a single peak at the 268  C), ruling out irreversible degradation phenomena in the investigated conditions. Indeed, the fact that the samples immersed in water did not reach an equilibrium state after four weeks and/or the presence of molecular heterogeneities in the epoxy network (formation of loops, dangling ends, presence of residual unreacted monomer, etc) [23,25] can be at the basis of the non-uniform distribution of water. In Fig. 7a,b DMTA results of the [0]10 e CFN9 composite systems before and after aging are reported. For ease of comparison, all values of peak temperature in tand curve, for the [0]10 and CFN9 systems, at the different tested aging conditions are reported in Table 3. For the [0]10 system that is tested after the absorption step (Fig. 7a), a shift of the main peak of the tand curve to lower

M. Di Filippo et al. / Polymer Degradation and Stability 126 (2016) 188e195 12








Matrix CF


CFN1 2



Mt %



Mt %





1.5 1 0.5 0











Time1/2 [h1/2]

Fig. 5. Water (a) absorption and (b) desorption of the neat resin (matrix) and composite laminates. In (a), the lines are the fitting curves obtained using Eq. (2).

Table 2 Coefficient of Determination (R2), apparent diffusion coefficient (D), polymer relaxation (k) and quasi-equilibrium water uptake value (M∞0), determined according to Eq. (2) for the neat resin and composite systems. The two different geometries are denoted as DMTA and SBS. R2


0.99788 0.99002 0.98284 0.98874 0.98714 0.99361 0.99562

4.17 1.49 9.95 3.85 1.43 1.08 3.78


K [s0.5]


1.06 5.78 3.57 2.48 6.55 4.09 3.39

10 104 105 105 104 104 105


M∞0% 104 104 104 104 104 104 104

4.58 5.45 6.32 5.04 4.55 5.26 4.51




0,4 Non aged



Absorp. [email protected] °C


E' [Pa]


D [mm2/s]

Desorp. [email protected] Desorp. [email protected] °C






0 0








Temperature [°C] Fig. 6. Storage modulus (E0 ) and loss factor (tand) from DMTA of the neat epoxy resin, before and after aging. Fig. 7. Storage modulus (E0 ) and loss factor (tand) from DMTA before and after aging of: (a) [0]10 system; (b) CFN9 system.

temperatures (260  C) and the appearance of a secondary peak at about 170  C, corresponding to a flex point in E0 curve, are shown. These are the effects of plasticization operated by water on different network portions, as discussed for the neat resin. For this system, once the water desorption process is completed, a full recovery of the thermal properties is observed. Actually, a small increase in Tg with respect to the non aged system is observed and can be attributed to a slow “post-curing” treatment, occurring at relatively low temperature on the plasticised system during the prolonged immersion in water at 90  C. DMTA analysis on the CFN9 system (see Fig. 7b) after water absorption shows a single tand peak with a broad shoulder. The temperature at peak shifts towards higher values (266  C). The shoulder is no longer present after

Table 3 Peak temperature in tand curve for [0]10 and CFN9, before and after aging. System

[0]10 CFN9

T of tand curve at peak, [ C] Non aged

Absorp. [email protected]  C

Desorp. [email protected]

Desorp. [email protected]  C

265 263e180

260e170 266-shoulder

270 266

270 270

partial or complete water desorption and the higher peak temperature is retained.


M. Di Filippo et al. / Polymer Degradation and Stability 126 (2016) 188e195

The mechanical behaviour of the aged laminates is investigated by both transverse flexural tests and short beam shear tests. Flexural modulus and interlaminar shear strength values are shown in Fig. 8a and b. Wet samples show a decrease of about 15% in flexural modulus that is recovered after complete water desorption (the differences observed are within the experimental error). Such a decrease in stiffness is likely induced by the increased mobility of the resin due to the above discussed plasticization effects. No significant differences are shown by the different systems. A similar trend is observed for the ILSS data. The reduction in strength caused by water uptake is almost entirely recovered upon drying. The improvement of 30% of ILSS achieved by the introduction of mats at each interlayer with respect to [0]10 is maintained at each stage of the hydrothermal aging process. 4. Conclusions Carbon fiber reinforced epoxy laminates with Nylon 6,6 electrospun mats as interleaves have been prepared and characterized. An increase of interlaminar shear strength up to 30% with respect to the composite with no mats is observed when these elements are placed at each interlayer. No detrimental effects on Tg and transversal flexural modulus are observed. Interestingly, a significant reduction in porosity is obtained, the mats probably acting as resin reservoirs and collaborating to a more uniform distribution of the resin across the laminate upon curing. The reduction in porosity is at the basis of the reduced water uptake from the composite with the nanofibrous mat reinforced interlayers. All the systems aged in the conditions here investigated (immersed in water at constant temperature of 90  C) are

Fig. 8. (a) Transverse flexural modulus and (b) interlaminar shear strength of composites for different aging conditions.

plasticised by water but they recover their original thermal and mechanical properties after drying. No evidence of irreversible degradation phenomena caused the hygroscopic Nylon 6,6 nanofibers is found. These results encourage in proceeding with testing the performance of these nanofibrous mats toughened composites in more severe environmental aging conditions. References [1] A.J. Kinloch, S.J. Shaw, D.A. Tos, D.L. Hunston, Deformation and fracture behavior of a rubber-toughened epoxy: 1. Microstructure and fracture studies, Polymer 24 (1983) 1341e1354. [2] K. Yamanaka, Y. Takagi, T. Inoue, Reaction-induced phase separation in rubber-modified epoxy resins, Polymer 60 (1989) 1839e1844. [3] C. Dispenza, J.T. Carter, P.T. Mcgrail, G. Spadaro, Reactive blending of functionalized acrylic rubbers and epoxy resins, Polym. Eng. Sci. 41 (9) (2001) 1486e1496. [4] C. Yan, K. Xiao, L. Ye, Y.W. Mai, Numerical and experimental studies on the fracture behavior of rubber-toughened epoxy in bulk specimen and laminated composites, J. Mater. Sci. 37 (2002) 921e927. [5] S. Sprenger, M.H. Kothmann, V. Altstaedt, Carbon fiber-reinforced composites using an epoxy resin matrix modified with reactive liquid rubber and silica nanoparticles, Compos. Sci. Technol. 105 (2014) 86e95. [6] B.S. Kim, T. Chiba, T. Inoue, Morphology development via reaction-induced phase separation in epoxy/poly(ether sulfone) blends: morphology control using poly(ether sulfone) with functional end-groups, Polymer 36 (1) (1995) 43e47. [7] D.J.P. Turmel, I.K. Paetridge, Heterogeneous phase separation around fibres in epoxy/PEI blends and its effect on composite delamination resistance, Compos. Sci. Technol. 57 (1997) 1001e1007. [8] F. Mujika, A. De Benito, B. Fernandez, A. Vazquez, R. Llano-Ponte, I. Mondragon, Mechanical properties of carbon woven reinforced epoxy matrix composites. A study on the influence of matrix modification with polysulfone, Polym. Compos. 23 (3) (2002) 372e382. [9] Y. Zeng, H.Y. Liu, Y.W. Mai, X.S. Du, Improving interlaminar fracture toughness of carbon fibre/epoxy laminates by incorporation of nano-particles, Compos. B Eng. 43 (2012) 90e94. [10] Y. Ye, H. Chen, J. Wua, C.M. Chan, Interlaminar properties of carbon fiber composites with halloysite nanotube-toughened epoxy matrix, Compos. Sci. Technol. 71 (2011) 717e723. [11] R.S. Dubrow, United States patent 7,651,769 B2: Structures, systems and methods for joining articles and materials and uses therefore, (2010). [12] R. Palazzetti, A. Zucchelli, C. Gualandi, M.L. Focarete, L. Donati, G. Minak, S. Ramakrishna, Influence of electrospun Nylon 6,6 nanofibrous mats on the interlaminar properties of Gr-epoxy composite laminates, Compos. Struct. 94 (2012) 571e579. [13] G. Li, P. Li, C. Zhang, Y. Yu, H. Liu, S. Zhang, X. Jia, X. Yang, Z. Xue, S. Ryu, Inhomogeneous toughening of carbon fiber/epoxy composite using electrospun polysulfone nanofibrous membranes by in situ phase separation, Compos. Sci. Technol. 68 (2008) 987e994. [14] J. Zhang, T. Yang, T. Lin, C.H. Wang, Phase morphology of nanofibre interlayers: critical factor for toughening carbon/epoxy composites, Compos. Sci. Technol. 72 (2012) 256e262. [15] R. Palazzetti, X. Yan, A. Zucchelli, Influence of geometrical features of electrospun nylon 6,6 interleave on the CFRP laminates mechanical properties, Polym. Compos. 35 (2014) 137e150. [16] S. Hamer, H. Leibovich, A. Green, R. Intrater, R. Avrahami, E. Zussman, A. Siegmann, D. Sherman, Mode I interlaminar fracture toughness of nylon 66 nanofibrilmat interleaved carbon/epoxy laminates, Polym. Compos. 32 (2011) 1781e1789. [17] S. Alessi, M. Di Filippo, C. Dispenza, M.L. Focarete, C. Gualandi, R. Palazzetti, G. Pitarresi, A. Zucchelli, Effects of nylon 6,6 nanofibrous mats on thermal properties and delamination behavior of high performance CFRP laminates, Polym. Compos. 36 (2015) 1303e1313. [18] P. Akangah, S. Lingaiah, K. Shivakumar, Effect of nylon-66 nano.fiber interleaving on impact damage resistance of epoxy/carbon fiber composite laminates, Compos. Struct. 92 (2010) 1432e1439. [19] M. Leali Costa, S.F. Muller de Almeida, M.C. Rezende, Hygrothermal effects on dynamic mechanical analysis and fracture behavior of polymeric composites, Mater. Res. 8 (3) (2005) 335e340. [20] S. Alessi, G. Pitarresi, G. Spadaro, Effect of hydrothermal ageing on the thermal and delamination fracture behaviour of CFRP composites, Compos. B Eng. 67 (2014) 145e153. [21] C.J. Tsenoglou, S. Pavlidou, C.D. Papaspyrides, Evaluation of interfacial relaxation due to water absorption in fiber-polymer composites, Compos. Sci. Technol. 66 (2006) 2855e2864. [22] B.C. Ray, Temperature effect during humid ageing on interfaces of glass and carbon fibers reinforced epoxy composites, J. Colloid Interface Sci. 298 (2006) 111e117. [23] J. Zhou, J.P. Lucas, Hygrothermal effects of epoxy resin. Part I: the nature of water in epoxy, Polymer 40 (1999) 5505e5512. [24] S. Alessi, D. Conduruta, G. Pitarresi, C. Dispenza, G. Spadaro, Accelerated

M. Di Filippo et al. / Polymer Degradation and Stability 126 (2016) 188e195






ageing due to moisture absorption of thermally cured epoxy resin/polyethersulphone blends. Thermal, mechanical and morphological behavior, Polym. Degrad. Stab. 96 (2011) 642e648. M. Jackson, M. Kaushik, S. Nazarenko, S. Wars, R. Maskell, J. Wiggins, Effect of free volume hole-size on fluid ingress of glassy epoxy networks, Polymer 52 (2011) 4528e4535. S.G. Prolongo, M.R. Gude, A. Urena, Water uptake of epoxy composites reinforced with carbon nanofillers, Compos. Part A Appl. Sci. Manuf. 43 (2169) (2012) 2175. G. Pitarresi, M. Scafidi, S. Alessi, M. Di Filippo, C. Billaude, G. Spadaro, Absorption kinetics and swelling stresses in hydrothermally aged epoxies investigated by photoelastic image analysis, Polym. Degrad. Stab. 111 (2015) 55e63. A. Zafar, F. Bertocco, J. Schjodt-Thomsen, J.C. Rauhe, Investigation of the long term effects of moisture on carbon fibre and epoxy matrix composites, Compos. Sci. Technol. 72 (2012) 656e666. L.R. Bao, A.F. Yee, C.Y.C. Lee, Moisture absorption and hygrothermal aging in a bismaleimide resin, Polymer 42 (2001) 7327e7333.


[30] L.R. Bao, A.F. Yee, Moisture diffusion and hygrothermal aging in bismaleimide matrix carbon fiber composites-partI: uni-weave composites, Compos. Sci. Technol. 62 (2002) 2099e2110. [31] V.M. Karbhari, G. Xian, Hygrothermal effects on high VF pultruded unidirectional carbon/epoxy composites: moisture uptake, Compos. B Eng. 40 (2009) 41e49. [32] Z. Lu, G. Xian, H. Li, Effects of thermal aging on the water uptake behavior of pultruded BFRP plates, Polym. Degrad. Stab. 110 (2014) 216e224. [33] UNI EN ISO 14125, Fibre-reinforced Plastic Composites e Determination of Flexural Properties, 2000. [34] ASTM D 2344/D2344M e 13, Standard Test Method for Short-beam Strength of Polymer Matrix Composite Materials and Their Laminates, 2013. [35] ImageJ image processing and analysis in Java. Available from: http://rsbweb. nih.gov/ij/index.html. [36] P.C. Varelidis, R.L. McCulloughb, C.D. Papaspyrides, The effect on the mechanical properties of carbon/epoxy composites of polyamide coatings on the fibers, Compos. Sci. Technol. 59 (1999) 1813e1823.