- Email: [email protected]
The effect of the moisture content on the curing characteristics of an epoxy matrix in the presence of nanofibrous structures
Polymer Testing 40 (2014) 265e272
Contents lists available at ScienceDirect
Polymer Testing journal homepage: www.elsevier.com/locate/polytest
Material behaviour
The effect of the moisture content on the curing characteristics of an epoxy matrix in the presence of nanofibrous structures Sam van der Heijden a, Bert De Schoenmaker a, Hubert Rahier b, Guy Van Assche b, Karen De Clerck a, * a b
Ghent University, Department of Textiles, Technologiepark 907, B-9052 Gent, Belgium Vrije Universiteit Brussel, Department Materials and Chemistry, Pleinlaan 2, B-1050 Brussels, Belgium
a r t i c l e i n f o
a b s t r a c t
Article history: Received 4 July 2014 Accepted 21 August 2014 Available online 18 September 2014
This paper investigates the effect of polyamide 6 (PA 6) nanofibres and microfibres, as well as the effect of moisture present in these structures, on the curing kinetics of a diglycidyl ether of bisphenol A (DGEBA)-methylenedianiline (MDA) system. Modulated temperature differential scanning calorimetry measurements show that the initial reaction rate follows a linearly increasing trend as a function of the moisture content present in the nanofibrous structures. Compared to PA 6 microfibrous structures, incorporating PA 6 nanofibrous structures exposed to the same humidity resulted in a higher initial reaction rate, which is in agreement with the higher water absorption of the nanofibrous structure, measured with differential vapour sorption. Overall, the nanofibres themselves affect the curing characteristics, and the moisture present in the structures enhances this effect. © 2014 Elsevier Ltd. All rights reserved.
Keywords: Nano-structures Cure behaviour Thermal analysis
1. Introduction Incorporation of nanoparticles, such as carbon nanotubes, nanoclays and thermoplastic inclusions, can enhance the performance of a thermosetting epoxy matrix [1e5]. Depending on the incorporated particles, higher performance can be found in, for example, higher stiffness [3,6], increased toughness [7e9] or better electromagnetic properties [10,11]. Important disadvantages related to the use of nanoparticles are associated with safety issues [12] and dispersion problems [3,5,6]. Owing to their small dimensions, nanoparticles can easily be taken up by the human body, which is the main reason for health and safety problems. Furthermore, with the currently available techniques, it is
* Corresponding author. Tel.: þ32 9 264 57 40; fax: þ32 9 264 58 46. E-mail addresses: [email protected], sam.vanderheijden@ ugent.be (K. De Clerck). http://dx.doi.org/10.1016/j.polymertesting.2014.08.019 0142-9418/© 2014 Elsevier Ltd. All rights reserved.
still difficult to obtain homogeneous dispersion of nanoparticles in the resin. Homogeneous dispersion is required to take the full advantages of the nanoparticles in the composites. The embedding of nanofibrous nonwoven structures in general, and polyamide nanofibrous structures in particular, offers an interesting alternative to tackle these problems [13e16]. Owing to the macroscale of the nanofibrous nonwoven structures, less health hazards are involved. Moreover, since these highly porous nonwoven structures are easily wetted by the epoxy resin, they can be incorporated in epoxy composites through a straightforward impregnation step, circumventing the dispersion issue. However, the effect of these fibrous structures on the curing behaviour of the epoxy matrix needs further attention. It is known that the incorporation of fibres in a matrix may alter the curing behaviour, and thus the final performance of the matrix [17,18]. It is also known that moisture can play an important role in the curing kinetics of a thermoset matrix [19e21].
266
S. van der Heijden et al. / Polymer Testing 40 (2014) 265e272
Since nanofibrous nonwoven structures have high porosity combined with a high specific surface, they may very well interfere with the curing mechanism when incorporated in an epoxy matrix. In addition, the nanofibrous structures show a high moisture sorption capacity [22], thus the effect of moisture is also to be studied. Previous research showed that polyamide nanofibrous structures had a catalytic effect on the cure of an epoxy resin, and it was indicated that moisture may play an important role [22]. The effect of moisture present within the structures on the curing of epoxy resins has not yet been investigated. This paper examines the influence of the moisture content in polyamide 6 (PA 6) nanofibrous nonwovens on the epoxy cure kinetics and on the glass transition temperature using modulated temperature differential scanning calorimetry (MTDSC). Moreover, a comparison with conventional polyamide fibre woven structures is made to fully understand the effect of the nanofibrous structures. The moisture absorbed by the epoxy resin and the PA 6 fibres is measured using dynamic vapour sorption (DVS). 2. Materials and methods 2.1. Materials Epikote resin 828 LVEL (Hexion), which consists of the difunctional diglycidyl ether of bisphenol A (DGEBA) and the tetrafunctional hardener methylenedianiline (MDA) (Sigma-Aldrich), were used in stoichiometric quantities (molar ratio r ¼ [NH]/[Epoxy] ¼ 1). The PA 6 nanofibres were in-house produced on a multinozzle set-up [23]. Polyamide 6 pellets (16 wt%) were dissolved in a 1:1 formic acid/acetic acid solution. All chemicals were purchased from Sigma-Aldrich. The tip-tocollector distance and flow rate were set at 7 cm and 2 mL/h, respectively, while the applied voltage was adapted, between 25 and 30 kV, to obtain a stable electrospinning process. The conventional woven fabric of PA 6 microfibres was obtained from Concordia Textiles (Waregem, Belgium). All fabrics were desized and washed. The average fibre diameter of both the nanofibres and conventional microfibres was measured using CellD-software from Olympus (Table 1). The specific surface area was calculated based on a circular cross section and a PA 6 density of 1.14 kg/m3.
2.2. Methods The moisture sorption of the resin, the PA 6 microfibrous woven and nanofibrous nonwoven structures was Table 1 Characteristics of PA 6 nanofibres and microfibres.
Nanofibres Conventional micro fibres
Fibre diameter [nm]
Specific surface area [106 m2/kg]
180 ± 20 10428 ± 140
19.5 0.35
evaluated using a TA Instruments Q5000 dynamic vapour sorption (DVS) analyser. All samples were analysed in TA instruments DVS quartz pans at 23 ± 1 C. For the fibrous structures, a full sorption and desorption isotherm was measured. The method started with drying the fibres at 0 % relative humidity until the weight change of the sample was less than 0.01 % for 15 min. The sorption isotherm was measured by increasing the relative humidity in steps of 10 %, in the range of 5 to 95 %RH. Next, the desorption was measured by decreasing the relative humidity, with the same step size. For both the sorption and desorption steps, the sample was allowed to equilibrate until the weight change was less than 0.01 % during the last 15 min of the measurement. For the DGEBA-MDA resin, the maximum sorption at 95 %RH was determined as an indication of the moisture sorption of an uncured resin. A drying step was not necessary since the resin was heated to 160 C during preparation. Modulated temperature differential scanning calorimetry (MTDSC) measurements were performed using a TA Instruments Q2000 Tzero™ DSC, purged with a constant nitrogen flow of 50 mL/min. The instrument was calibrated using sapphire (Tzero calibration) and indium (heat flow rate and temperature). The modulation amplitude for the MTDSC measurements was chosen at 0.5 C with a period of 60 seconds, as used in previous work [22,24]. The samples were analysed in aluminium Tzero hermetic DSC pans (TA Instruments), which were loaded with 3.50 ± 0.05 mg fibres and 10.0 ± 1 mg resin, which implies a fibre content of 26 ± 2 wt%. The epoxy resin was heated to 160 C for 15 minutes, followed by the addition of the ground hardener and a quick stir for a few seconds. Immediately after mixing, the hot mixture was quickly poured into a syringe and quenched in liquid nitrogen (the reactive mixture was less than 10 s at 160 C). After keeping the (quenched) syringe at room temperature until the resin was sufficiently liquid, about 3 min, a drop of resin mixture was injected into the DSC pan. The resin penetrates immediately into the PA6 nanofibre web. All samples were cured quasi-isothermally at 80 C for 250 min. For the determination of the glass transition temperature and the residual reactivity, non-isothermal heating from 0 C to 195 C at 2.5 C/min was performed after the quasi-isothermal cure. The glass transition temperature of the fully cured system was obtained by a second identical heating ramp. To compare the heat flow signal profiles, which relate to variations of the reaction rate as the conversion progresses, it is necessary to have a good definition of time zero, the time at which the reaction is considered to start. In this paper, time zero is defined as the moment when 79 C was reached. The choice of 79 C was prompted by the fact that the (average) heating rate decreases significantly when approaching the final isothermal temperature of 80 C. Therefore, it may take a few minutes before 80 C is reached exactly. A statistical analysis of variance (ANOVA), was performed in SPSS to investigate the different kinetic parameters at the different relative humidity levels. Prior to the ANOVA analysis, the normality of the distribution was tested by a Shapiro-Wilk test.
S. van der Heijden et al. / Polymer Testing 40 (2014) 265e272
267
3. Results and discussion 3.1. Moisture sorption behaviour of the (nano)fibrous structures and the resin To be able to relate the curing kinetics of an epoxy matrix to the moisture present in the reinforcing (nano) fibrous structures, a moisture sorption behaviour study was performed on the different components. The moisture sorption and desorption profiles of PA 6 nanofibres and microfibres were studied through DVS (Fig. 1). The small standard deviations (error bars) indicate the measurements were very reproducible. According to the IUPACclassification [25], both curves are type III-isotherms. This type describes the sorption on nanoporous and non-porous adsorbents with weak interactions [25,26]. When comparing the nanofibres to the microfibres, it can be seen that their moisture sorption is nearly the same until about 65 % relative humidity, after which the nanofibres start to show a higher moisture sorption. This higher moisture sorption can probably be explained by the larger porosity of these nanofibrous structures. It can be assumed that, from 65 % relative humidity onwards, the share of moisture sorption through filling up the pores starts to become more important. The moisture absorbs very quickly in the nanofibrous structures, as demonstrated in Fig. 2. Starting from dry nonwoven nanofibres, it took less than 2 hours to reach equilibrium at a relative humidity of 95 %. Similar results were obtained for the microfibres. This indicates that conditioning the samples at a selected relative humidity for 24 hours is more than sufficient to obtain the desired moisture content. Next to the characteristics of the moisture sorption of the fibrous structures, the sorption characteristics of the neat resin were also investigated. The moisture sorption at 95 %RH of the uncured DGEBA-MDA mixture as a function of time is presented in Fig. 3. This figure shows that the time to reach equilibrium is more than 20 hours, which is significantly longer than for the fibrous structures. Thus, the diffusion of water in the resin is significantly slower than in the fibrous structure. Furthermore, the moisture
Fig. 1. Sorption (straight line) and desorption (dashed line) of PA 6 nanofibres (black lines) and microfibres (grey lines) at 23 C. The error bars represent the standard deviation for three samples.
Fig. 2. Evolution of the moisture content in a nanofibrous structure at 95 % relative humidity and 23 C starting from dried fibres.
sorption at equilibrium is only slightly above 1 %, which is about 8 times lower than the moisture sorption of the PA 6 fibrous structures. It can, therefore, be concluded that there is a large difference in the maximal water sorption capacity between the fibres and the resin. Furthermore, above 65 %RH the nanofibrous structures absorb more water compared to the microfibrous structures. To study the effect of the moisture content (the amount of absorbed water) on the curing kinetics, the fibrous structures were saturated at different RH's. 3.2. Influence of the moisture content in the nanofibrous structures on the curing properties First, the reproducibility of the cure of neat DGEBA-MDA resin was examined with DSC. The curing characteristics of four stoichiometric DGEBA-MDA mixtures (all having a molar ratio of [NH]/[Epoxy] ¼ r ¼ 1) were investigated at a cure temperature of 80 C. Fig. 4 shows the heat flow rate of these DGEBA-MDA mixtures as a function of time, proving the high reproducibility. The heat flow rate profiles show the typical autocatalytic behaviour for amine-cured epoxy resins. Next, the cure of the resin in the presence of nanofibrous structures, conditioned at different RH's to obtain different amounts of moisture in the structures, was studied by DSC.
Fig. 3. Evolution of the moisture content of uncured DGEBA-MDA mixture at 95 % relative humidity and 23 C starting from dried resin.
268
S. van der Heijden et al. / Polymer Testing 40 (2014) 265e272
Fig. 4. Heat flow signal for the cure of four neat DGEBA-MDA samples obtained from quasi-isothermal MTDSC experiments at 80 C.
The desired amount of nanofibre nonwoven was placed in the DSC pan and a drop of resin was injected into the pan, penetrating immediately into the nanofibre web. SEM images on cross-sectioned cured DSC samples confirmed the even distribution of the nanofibres throughout the samples (as shown in Fig. 5). Fig. 6A and B present the heat flow signals per gram of resin as a function of the reaction time and reversing heat capacity change. The heat flow signals will be discussed first. Increasing moisture content in the nanofibrous structures has a pronounced accelerating influence on the curing behaviour of the DGEBA-MDA resin system. The effect of the moisture content in the fibrous structures on the curing kinetics is discussed in more detail through four parameters characterizing the cure kinetics: the initial reaction rate, the maximum reaction rate, the time to reach the maximum reaction rate and the reaction enthalpy, as presented in Fig. 7. The initial reaction rate given is the value of the heat flow signal 5 minutes after time zero, ensuring that the heat flow signal has stabilized. At this stage of the reaction, the DGEBA-MDA mixture mainly consists of unreacted resin and hardener, which means that the influence of possible side reactions or byproducts is very minor. Furthermore, the viscosity of the system at this stage of the reaction is relatively low and the reaction is chemically controlled [27]. The maximum reaction rate and the time to reach this maximum reaction
Fig. 6. Heat flow (A) and reversing heat capacity change (B) obtained from a quasi-isothermal MTDSC scan at 80 C, illustrating the effect of the nanofibrous moisture content (expressed as wt % on nanofibre weight) on the curing behaviour of the DGEBA-MDA resin.
rate (tmaxHF) are defined as the maximum value of the heat flow signal with respect to the baseline level at the end of the cure experiment and the time to reach this maximum value, respectively. By dividing the isothermal reaction enthalpy (DH) by the total enthalpy of the reaction, obtained from a non-isothermal MTDSC experiment in which the reaction reaches completion, the conversion reached at the end of the isothermal experiment is found. Although Fig. 6 gives a single curve for each moisture content, five experiments were executed in each case. Fig. 7 gives an overview of the average value as well as the standard deviation for the four kinetic parameters as a function of the moisture content of the nanofibrous
Fig. 5. SEM images of cross section of the cured samples.
S. van der Heijden et al. / Polymer Testing 40 (2014) 265e272
269
Fig. 7. Initial reaction rate (A), maximum reaction rate (B), time to maximum heat flow (tmaxHF) (C) and reaction enthalpy (DH) (D) obtained from quasiisothermal MTDSC experiments at 80 C for neat resin (D) and for samples containing moistened nanofibres (◊). The error bars represent the standard deviation for five samples.
structures. The triangles in the different graphs in Fig. 7 represent the values of the neat dry resin for that specific parameter. The initial reaction rate follows a linearly increasing trend as a function of the moisture content. This implies a first order kinetics in moisture content. Moreover, even when no moisture is present in the fibres, the initial reaction rate is higher: the initial reaction rate increased from 0.017 ± 0.003 W/g for a neat resin to 0.024 ± 0.002 W/g when dry nanofibres are incorporated. This indicates that the nanofibres themselves also have a catalysing effect on the resin cure, which is attributed to a catalysing effect of the functional groups on the fibre surface. A further increase of the initial reaction rate with moisture content is observed up to 0.044 ± 0.006 W/g. Fig. 7B and C again show a linear correlation for both the maximum reaction rate and the time to maximum reaction rate as a function of the moisture content in the nanofibrous structures. The catalytic effect of the dry nanofibrous structure itself is only significant in the early stages of the curing process, as the maximum reaction rate of the neat resin is at the same level as the sample containing nanofibres at 0 wt% moisture content (Fig. 7B). The observed correlation of the maximum reaction rate with the moisture content shows a somewhat lower R2-value, probably due to competition of the influence of absorbed water with the catalysis by hydroxyl groups formed during the reaction. Thus, the initial reaction rate is the most suitable parameter to evaluate the effect of the moisture content on the cure kinetics, since at this stage of the reaction the system mainly contains a mixture of uncured resin, hardener and moisture. The total enthalpy of the reaction, obtained from a nonisothermal curing cycle, and the enthalpy of the quasiisothermal cure experiment were 435 ± 40 J/g and 326 ± 15 J/g for the neat resin (triangle in Fig. 7D), respectively, giving a final isothermal conversion of 75 % for the neat DGEBA-MDA mixture. Statistical ANOVA-tests
indicated that the moisture content of the nanofibrous structures had no significant influence on the reaction enthalpy (see also Fig. 7D). Thus, although the moisture content affects the reaction rate markedly, it has no significant effect on the degree of conversion of the DGEBAMDA mixture attained at the end of the isothermal cure experiments. In addition to the heat flow, the change of the reversing heat capacity is measured in the quasi-isothermal MTDSC experiments (Fig. 6B). The stepwise decrease in heat capacity observed can be attributed to the gradual vitrification of the resin due to the increase of the glass transition temperature of the reacting system to the cure temperature [24,27]. This decrease is due to the loss of segmental or cooperative mobility as the system gradually evolves from a liquid/rubbery material to a glass due to the progressing reaction. The initial rise in reversing heat capacity can be attributed to the higher heat capacity of the reaction products compared to the initial resin [28e30]. The heat capacity evolution also depends on the moisture content of the fibres. Fig. 6B shows that the reversing heat capacity change on vitrification of the neat resin sample is significantly higher than for the nanofibre containing samples: the average value of the reversing heat capacity change expressed per grams of resin was around 0.32 ± 0.03 J/(g C) for all moisture contents, compared to 0.45 ± 0.03 J/(g C) for the neat resin. The difference between the two cases seems to indicate vitrification has a less pronounced effect on the cooperative mobility lost in the case when nanofibres are present. On the other hand, no pronounced effect of the moisture content on the step height of the reversing heat capacity change was observed, indicating that once fibres are present, a similar degree of cooperative mobility is lost on vitrification. Another important parameter that can be obtained from the reversing heat capacity signals is the time to vitrification, characterized here by the time at the half of the decrease in reversing heat capacity. In contrast to the step
270
S. van der Heijden et al. / Polymer Testing 40 (2014) 265e272
height, the time to vitrification is affected by the moisture content. Vitrification occurs sooner with increasing moisture content of the nanofibrous structure (see Fig. 8). This indicates that Tg is rising more quickly at higher moisture content, which is in agreement with catalysis of the DGEBA-MDA reaction by (the moisture present in) the nanofibrous structures. 3.3. Comparison between nanofibrous and conventional structures Since it is now proven that the nanofibres and the moisture in these nanofibrous structures have a catalysing effect, it is also interesting to compare the findings to microfibrous structures. Fig. 9 shows clear differences in initial reaction rate between nanofibrous and microfibrous samples at each moisture content (the mass fraction of fibres is 26 ± 2 wt% in each case). While for the nanofibrous structures the initial reaction rate linearly increases with increasing moisture content, it seems to stagnate for the microfibres around a moisture content of 2.7 wt%. Due to the larger average diameter of the microfibres (Table 1), the moisture present within a microfibre has to diffuse over a larger distance to reach the fibre/resin interface. Moreover, the specific surface area is about 56 times smaller for the microfibrous structures compared to the nanofibrous structures (Table 1). Thus, less surface is available for moisture to diffuse into the resin matrix. In addition, the moisture has to diffuse over larger distances in the resin, due to the larger size of the pores in between the thicker fibres. The diffusion of moisture from the fibres into the bulk of the resin is most probably the rate determining factor, as observed from the DVS measurements. Thus, although the total amount of moisture in the conventional microfibres is approximately the same as in the nanofibres, there is not enough time to complete the diffusion of moisture into the bulk of the resin due to the larger distances over which diffusion is required to take place. 3.4. Influence of the moisture content on the final glass transition temperature of the resin Besides the effect of the moisture content on the curing characteristics, the effect on the final glass transition
Fig. 8. The time to vitrification of the neat DGEBA-MDA system (D) and for samples containing moistened nanofibres (◊). The error bars represent the standard deviation for five samples.
Fig. 9. Comparison between the initial reaction rate of the DGEBA-MDA curing reaction at different fibre moisture contents for neat resin (D), PA 6 nanofibres (◊) and conventional PA 6 micro fibres (▫). The error bars represent the standard deviation for five samples.
temperature was also investigated. Fig. 10 shows the final glass transition temperature for the resin containing nanofibres, compared to the neat resin. An ANOVA-test proved there is no statistically significant difference between the final glass transition temperatures reached with the addition of nanofibres with different moisture contents. There is, however, a statistically significant difference between the final glass transition temperature of the nanofibre-containing resin samples and the neat resin samples. The final glass transition temperature is 169 ± 4 C for the nanofibre-containing resin samples, whereas for the neat resin samples this is 172 ± 3 C. This indicates that the Tg-reducing effect of the addition of nanofibres is to be found in the addition of the nanofibres themselves, and not in the moisture which is brought into the DGEBA-MDA resin by the nanofibrous structures. It is suggested that the cause of the Tg-reducing effect can be found in a preferential migration (or adsorption) of one of the resin components to (or on) the polyamide nanofibres. Similar observations have been made in the past [31e33]. Furthermore, it is also well known that the epoxy-amine ratio has a strong influence on the final glass transition temperature of epoxy resins, the highest glass transition temperature being achieved with stoichiometric ratios [27]. Through preferential migration of a fraction of one of
Fig. 10. Final glass transition temperature Tg of a fully cured DGEBA-MDA resin containing PA 6 nanofibres with different moisture contents (◊), and dry neat resin without fibres (D).
S. van der Heijden et al. / Polymer Testing 40 (2014) 265e272
the components to the nanofibre surface, the stoichiometry of the DGEBA-MDA resin mixture is disturbed throughout the matrix, resulting in the lower final glass transition temperature [28]. 4. Conclusions Nanofibrous structures and the moisture present in them clearly affect the curing characteristics and final glass transition temperature of a DGEBA-MDA mixture. The addition of dry nanofibrous structures has a small catalytic effect on the cure reaction of the DGEBA-MDA resin, observed as an increase in the initial reaction rate. Moisture present in the nanofibrous structures additionally catalyzes the reaction, with the initial reaction rate following a linearly increasing trend as a function of the moisture content. Adversely, conventional PA 6 microfibres result in a smaller catalytic effect, which stagnates for 2.7 wt% moisture and above. This difference was attributed to the larger specific surface area of the nanofibrous structures as well as the smaller distances that need to be covered for diffusion of moisture out of the fibre and into the matrix. Statistical tests proved that the moisture content in the nanofibrous structures had no influence on the reaction enthalpy, and thus also not on the degree of conversion of the DGEBA-MDA mixture reached during isothermal cure. Nevertheless, the final glass transition reached is clearly lower for the nanofibrous materials compared to neat resin. This lowering of the final glass transition temperature in the case of nanofibrous structures was attributed to preferential migration to and/or adsorption of one of the resin components on the fibre surface. Both the catalytic and plasticizing effects are important for the application of these nanofibrous structures in epoxy resin-PA 6 nanocomposites. Acknowledgements Financial support from The Agency for Innovation by Science and Technology, Flanders (IWT) is gratefully acknowledged. Results in this paper were obtained within the framework of the IWT Strategic Basic Research Grant 121156. References [1] B. Nuhiji, D. Attard, G. Thorogood, T. Hanley, K. Magniez, B. Fox, The effect of alternate heating rates during cure on the structureproperty relationships of epoxy/MMT clay nanocomposites, Compos. Sci. Technol. 71 (2011) 1761e1768. [2] B. Ashrafi, J.W. Guan, V. Mirjalili, Y.F. Zhang, L. Chun, P. Hubert, B. Simard, C.T. Kingston, O. Bourne, A. Johnston, Enhancement of mechanical performance of epoxy/carbon fiber laminate composites using single-walled carbon nanotubes, Compos. Sci. Technol. 71 (2011) 1569e1578. [3] J.N. Coleman, U. Khan, W.J. Blau, Y.K. Gun'ko, Small but strong: a review of the mechanical properties of carbon nanotube-polymer composites, Carbon 44 (2006) 1624e1652. [4] L.C. Tang, Y.J. Wan, K. Peng, Y.B. Pei, L.B. Wu, L.M. Chen, L.J. Shu, J.X. Jiang, G.Q. Lai, Fracture toughness and electrical conductivity of epoxy composites filled with carbon nanotubes and spherical particles, Compos. Part A-Appl. S 45 (2013) 95e101. [5] T.W. Chou, L.M. Gao, E.T. Thostenson, Z.G. Zhang, J.H. Byun, An assessment of the science and technology of carbon nanotube-based fibers and composites, Compos. Sci. Technol. 70 (2010) 1e19.
271
[6] H. Qian, E.S. Greenhalgh, M.S.P. Shaffer, A. Bismarck, Carbon nanotube-based hierarchical composites: a review, J. Mater. Chem. 20 (2010) 4751e4762. [7] L. Boogh, B. Pettersson, J.A.E. Manson, Dendritic hyperbranched polymers as tougheners for epoxy resins, Polymer 40 (1999) 2249e2261. [8] K.Y. Kim, L. Ye, Interlaminar fracture toughness of CF/PEI composites at elevated temperatures: roles of matrix toughness and fibre/matrix adhesion, Compos. Part A-Appl. S 35 (2004) 477e487. [9] S.P. Wilkinson, T.C. Ward, J.E. Mcgrath, Effect of thermoplastic modifier variables on toughening a bismaleimide matrix resin for high-performance composite-materials, Polymer 34 (1993) 870e884. [10] I.W. Nam, H.K. Lee, J.H. Jang, Electromagnetic interference shielding/ absorbing characteristics of CNT-embedded epoxy composites, Composites Part A-applied Science and Manufacturing 42 (2011) 1110e1118. [11] B.P. Singh, V. Choudhary, P. Saini, R.B. Mathur, Designing of epoxy composites reinforced with carbon nanotubes grown carbon fiber fabric for improved electromagnetic interference shielding, AIP Adv. 2 (2012). [12] K. Savolainen, L. Pylkkanen, H. Norppa, G. Falck, H. Lindberg, T. Tuomi, M. Vippola, H. Alenius, K. Hameri, J. Koivisto, D. Brouwer, D. Mark, D. Bard, M. Berges, E. Jankowska, M. Posniak, P. Farmer, R. Singh, F. Krombach, P. Bihari, G. Kasper, M. Seipenbusch, Nanotechnologies, engineered nanomaterials and occupational health and safety e a review, Saf. Sci. 48 (2010) 957e963. [13] 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. [14] K. Bilge, S. Venkataraman, Y.Z. Menceloglu, M. Papila, Global and local nanofibrous interlayer toughened composites for higher inplane strength, Compos. Part A: Appl. Sci. Manuf. 58 (2014) 73e76. [15] B. De Schoenmaker, S. Van der Heijden, I. De Baere, W. Van Paepegem, K. De Clerck, Effect of electrospun polyamide 6 nanofibres on the mechanical properties of a glass fibre/epoxy composite, Polym. Test. 32 (2013) 1495e1501. [16] P. Akangah, S. Lingaiah, K. Shivakumar, Effect of Nylon-66 nanofiber interleaving on impact damage resistance of epoxy/carbon fiber composite laminates, Compos. Struct. 92 (2010) 1432e1439. [17] D. Aussawasathien, E. Sancaktar, Effect of non-woven carbon nanofiber mat presence on cure kinetics of epoxy nanocomposites, Macromol. Symp. 264 (2008) 26e33. [18] G.R. Palmese, O.A. Andersen, V.M. Karbhari, Effect of glass fiber sizing on the cure kinetics of vinyl-ester resins, Compos. Part AAppl. S 30 (1999) 11e18. [19] J.C. Cabanelas, S.G. Prolongo, B. Serrano, J. Bravo, J. Baselga, Water absorption in polyaminosiloxane-epoxy thermosetting polymers, J. Mater. Process. Tech. 143 (2003) 311e315. [20] N.D. Danieley, Effects of curing on the glass-transition temperature and moisture absorption of a neat epoxy-resin, J. Polym. Sci. Polym. Chem. 19 (1981) 2443e2449. [21] P. Musto, E. Martuscelli, G. Ragosta, P. Russo, The curing process and moisture transport in a tetrafunctional epoxy resin as investigated by FT-NIR spectroscopy, High Perform. Polym. 12 (2000) 155e168. [22] B. De Schoenmaker, S. van der Heijden, S. Moorkens, H. Rahier, G. van Assche, K. De Clerck, Effect of nanofibres on the curing characteristics of an epoxy matrix, Compos Sci. Technol. 79 (2013) 35e41. [23] P. Westbroek, C.T. Van, V.S. De, C.K. De, Production and use of laminated nanofibrous structures, in, Google Patents, 2008. [24] M. Reading, D. Hourston, B. Mele, H. Rahier, G. Assche, S. Swier, The application of modulated temperature differential scanning calorimetry for the characterisation of curing systems, in: Modulated Temperature Differential Scanning Calorimetry, Springer, Netherlands, 2006, pp. 83e160. [25] C. Sangwichien, G.L. Aranovich, M.D. Donohue, Density functional theory predictions of adsorption isotherms with hysteresis loops, Colloid. Surf. A 206 (2002) 313e320. [26] O. Ceylan, L. Van Landuyt, F. Meulewaeter, K. De Clerck, Moisture sorption in developing cotton fibers, Cellulose 19 (2012) 1517e1526. [27] G. Van Assche, A. Van Hemelrijck, H. Rahier, B. Van Mele, Modulated differential scanning calorimetry: isothermal cure and vitrification of thermosetting systems, Thermochim. Acta 268 (1995) 121e142. [28] S. Swier, G. Van Assche, B. Van Mele, Reaction kinetics modeling and thermal properties of epoxy-amines as measured by modulated-
272
S. van der Heijden et al. / Polymer Testing 40 (2014) 265e272
temperature DSC. II. Network-forming DGEBA plus MDA, J. Appl. Polym. Sci. 91 (2004) 2814e2833. [29] S. Swier, G. Van Assche, B. Van Mele, Reaction kinetics modeling and thermal properties of epoxy-amines as measured by modulatedtemperature DSC. I. Linear step-growth polymerization of DGEBA plus aniline, J. Appl. Polym. Sci. 91 (2004) 2798e2813. [30] S. Swier, B. Van Mele, Reaction thermodynamics of amine-cured epoxy systems: validation of the enthalpy and heat capacity of reaction as determined by modulated temperature differential scanning calorimetry, J. Polym. Sci. Polym. Phys. 41 (2003) 594e608.
[31] G.R. Palmese, R.L. Mccullough, Kinetic and thermodynamic considerations regarding interphase formation in thermosetting composite systems, J. Adhes. 44 (1994) 29e49. [32] M. Munz, Evidence for a three-zone interphase with complex elastic-plastic behaviour: nanoindentation study of an epoxy/thermoplastic composite, J. Phys. D Appl. Phys. 39 (2006) 4044e4058. [33] M. Munz, On the nature of the multi-zone interphase of a thermoset/thermoplastic composite e an analysis employing dynamicmechanical thermal analysis and nanoindentation, J. Adhes. 84 (2008) 445e482.
Contents lists available at ScienceDirect
Polymer Testing journal homepage: www.elsevier.com/locate/polytest
Material behaviour
The effect of the moisture content on the curing characteristics of an epoxy matrix in the presence of nanofibrous structures Sam van der Heijden a, Bert De Schoenmaker a, Hubert Rahier b, Guy Van Assche b, Karen De Clerck a, * a b
Ghent University, Department of Textiles, Technologiepark 907, B-9052 Gent, Belgium Vrije Universiteit Brussel, Department Materials and Chemistry, Pleinlaan 2, B-1050 Brussels, Belgium
a r t i c l e i n f o
a b s t r a c t
Article history: Received 4 July 2014 Accepted 21 August 2014 Available online 18 September 2014
This paper investigates the effect of polyamide 6 (PA 6) nanofibres and microfibres, as well as the effect of moisture present in these structures, on the curing kinetics of a diglycidyl ether of bisphenol A (DGEBA)-methylenedianiline (MDA) system. Modulated temperature differential scanning calorimetry measurements show that the initial reaction rate follows a linearly increasing trend as a function of the moisture content present in the nanofibrous structures. Compared to PA 6 microfibrous structures, incorporating PA 6 nanofibrous structures exposed to the same humidity resulted in a higher initial reaction rate, which is in agreement with the higher water absorption of the nanofibrous structure, measured with differential vapour sorption. Overall, the nanofibres themselves affect the curing characteristics, and the moisture present in the structures enhances this effect. © 2014 Elsevier Ltd. All rights reserved.
Keywords: Nano-structures Cure behaviour Thermal analysis
1. Introduction Incorporation of nanoparticles, such as carbon nanotubes, nanoclays and thermoplastic inclusions, can enhance the performance of a thermosetting epoxy matrix [1e5]. Depending on the incorporated particles, higher performance can be found in, for example, higher stiffness [3,6], increased toughness [7e9] or better electromagnetic properties [10,11]. Important disadvantages related to the use of nanoparticles are associated with safety issues [12] and dispersion problems [3,5,6]. Owing to their small dimensions, nanoparticles can easily be taken up by the human body, which is the main reason for health and safety problems. Furthermore, with the currently available techniques, it is
* Corresponding author. Tel.: þ32 9 264 57 40; fax: þ32 9 264 58 46. E-mail addresses: [email protected], sam.vanderheijden@ ugent.be (K. De Clerck). http://dx.doi.org/10.1016/j.polymertesting.2014.08.019 0142-9418/© 2014 Elsevier Ltd. All rights reserved.
still difficult to obtain homogeneous dispersion of nanoparticles in the resin. Homogeneous dispersion is required to take the full advantages of the nanoparticles in the composites. The embedding of nanofibrous nonwoven structures in general, and polyamide nanofibrous structures in particular, offers an interesting alternative to tackle these problems [13e16]. Owing to the macroscale of the nanofibrous nonwoven structures, less health hazards are involved. Moreover, since these highly porous nonwoven structures are easily wetted by the epoxy resin, they can be incorporated in epoxy composites through a straightforward impregnation step, circumventing the dispersion issue. However, the effect of these fibrous structures on the curing behaviour of the epoxy matrix needs further attention. It is known that the incorporation of fibres in a matrix may alter the curing behaviour, and thus the final performance of the matrix [17,18]. It is also known that moisture can play an important role in the curing kinetics of a thermoset matrix [19e21].
266
S. van der Heijden et al. / Polymer Testing 40 (2014) 265e272
Since nanofibrous nonwoven structures have high porosity combined with a high specific surface, they may very well interfere with the curing mechanism when incorporated in an epoxy matrix. In addition, the nanofibrous structures show a high moisture sorption capacity [22], thus the effect of moisture is also to be studied. Previous research showed that polyamide nanofibrous structures had a catalytic effect on the cure of an epoxy resin, and it was indicated that moisture may play an important role [22]. The effect of moisture present within the structures on the curing of epoxy resins has not yet been investigated. This paper examines the influence of the moisture content in polyamide 6 (PA 6) nanofibrous nonwovens on the epoxy cure kinetics and on the glass transition temperature using modulated temperature differential scanning calorimetry (MTDSC). Moreover, a comparison with conventional polyamide fibre woven structures is made to fully understand the effect of the nanofibrous structures. The moisture absorbed by the epoxy resin and the PA 6 fibres is measured using dynamic vapour sorption (DVS). 2. Materials and methods 2.1. Materials Epikote resin 828 LVEL (Hexion), which consists of the difunctional diglycidyl ether of bisphenol A (DGEBA) and the tetrafunctional hardener methylenedianiline (MDA) (Sigma-Aldrich), were used in stoichiometric quantities (molar ratio r ¼ [NH]/[Epoxy] ¼ 1). The PA 6 nanofibres were in-house produced on a multinozzle set-up [23]. Polyamide 6 pellets (16 wt%) were dissolved in a 1:1 formic acid/acetic acid solution. All chemicals were purchased from Sigma-Aldrich. The tip-tocollector distance and flow rate were set at 7 cm and 2 mL/h, respectively, while the applied voltage was adapted, between 25 and 30 kV, to obtain a stable electrospinning process. The conventional woven fabric of PA 6 microfibres was obtained from Concordia Textiles (Waregem, Belgium). All fabrics were desized and washed. The average fibre diameter of both the nanofibres and conventional microfibres was measured using CellD-software from Olympus (Table 1). The specific surface area was calculated based on a circular cross section and a PA 6 density of 1.14 kg/m3.
2.2. Methods The moisture sorption of the resin, the PA 6 microfibrous woven and nanofibrous nonwoven structures was Table 1 Characteristics of PA 6 nanofibres and microfibres.
Nanofibres Conventional micro fibres
Fibre diameter [nm]
Specific surface area [106 m2/kg]
180 ± 20 10428 ± 140
19.5 0.35
evaluated using a TA Instruments Q5000 dynamic vapour sorption (DVS) analyser. All samples were analysed in TA instruments DVS quartz pans at 23 ± 1 C. For the fibrous structures, a full sorption and desorption isotherm was measured. The method started with drying the fibres at 0 % relative humidity until the weight change of the sample was less than 0.01 % for 15 min. The sorption isotherm was measured by increasing the relative humidity in steps of 10 %, in the range of 5 to 95 %RH. Next, the desorption was measured by decreasing the relative humidity, with the same step size. For both the sorption and desorption steps, the sample was allowed to equilibrate until the weight change was less than 0.01 % during the last 15 min of the measurement. For the DGEBA-MDA resin, the maximum sorption at 95 %RH was determined as an indication of the moisture sorption of an uncured resin. A drying step was not necessary since the resin was heated to 160 C during preparation. Modulated temperature differential scanning calorimetry (MTDSC) measurements were performed using a TA Instruments Q2000 Tzero™ DSC, purged with a constant nitrogen flow of 50 mL/min. The instrument was calibrated using sapphire (Tzero calibration) and indium (heat flow rate and temperature). The modulation amplitude for the MTDSC measurements was chosen at 0.5 C with a period of 60 seconds, as used in previous work [22,24]. The samples were analysed in aluminium Tzero hermetic DSC pans (TA Instruments), which were loaded with 3.50 ± 0.05 mg fibres and 10.0 ± 1 mg resin, which implies a fibre content of 26 ± 2 wt%. The epoxy resin was heated to 160 C for 15 minutes, followed by the addition of the ground hardener and a quick stir for a few seconds. Immediately after mixing, the hot mixture was quickly poured into a syringe and quenched in liquid nitrogen (the reactive mixture was less than 10 s at 160 C). After keeping the (quenched) syringe at room temperature until the resin was sufficiently liquid, about 3 min, a drop of resin mixture was injected into the DSC pan. The resin penetrates immediately into the PA6 nanofibre web. All samples were cured quasi-isothermally at 80 C for 250 min. For the determination of the glass transition temperature and the residual reactivity, non-isothermal heating from 0 C to 195 C at 2.5 C/min was performed after the quasi-isothermal cure. The glass transition temperature of the fully cured system was obtained by a second identical heating ramp. To compare the heat flow signal profiles, which relate to variations of the reaction rate as the conversion progresses, it is necessary to have a good definition of time zero, the time at which the reaction is considered to start. In this paper, time zero is defined as the moment when 79 C was reached. The choice of 79 C was prompted by the fact that the (average) heating rate decreases significantly when approaching the final isothermal temperature of 80 C. Therefore, it may take a few minutes before 80 C is reached exactly. A statistical analysis of variance (ANOVA), was performed in SPSS to investigate the different kinetic parameters at the different relative humidity levels. Prior to the ANOVA analysis, the normality of the distribution was tested by a Shapiro-Wilk test.
S. van der Heijden et al. / Polymer Testing 40 (2014) 265e272
267
3. Results and discussion 3.1. Moisture sorption behaviour of the (nano)fibrous structures and the resin To be able to relate the curing kinetics of an epoxy matrix to the moisture present in the reinforcing (nano) fibrous structures, a moisture sorption behaviour study was performed on the different components. The moisture sorption and desorption profiles of PA 6 nanofibres and microfibres were studied through DVS (Fig. 1). The small standard deviations (error bars) indicate the measurements were very reproducible. According to the IUPACclassification [25], both curves are type III-isotherms. This type describes the sorption on nanoporous and non-porous adsorbents with weak interactions [25,26]. When comparing the nanofibres to the microfibres, it can be seen that their moisture sorption is nearly the same until about 65 % relative humidity, after which the nanofibres start to show a higher moisture sorption. This higher moisture sorption can probably be explained by the larger porosity of these nanofibrous structures. It can be assumed that, from 65 % relative humidity onwards, the share of moisture sorption through filling up the pores starts to become more important. The moisture absorbs very quickly in the nanofibrous structures, as demonstrated in Fig. 2. Starting from dry nonwoven nanofibres, it took less than 2 hours to reach equilibrium at a relative humidity of 95 %. Similar results were obtained for the microfibres. This indicates that conditioning the samples at a selected relative humidity for 24 hours is more than sufficient to obtain the desired moisture content. Next to the characteristics of the moisture sorption of the fibrous structures, the sorption characteristics of the neat resin were also investigated. The moisture sorption at 95 %RH of the uncured DGEBA-MDA mixture as a function of time is presented in Fig. 3. This figure shows that the time to reach equilibrium is more than 20 hours, which is significantly longer than for the fibrous structures. Thus, the diffusion of water in the resin is significantly slower than in the fibrous structure. Furthermore, the moisture
Fig. 1. Sorption (straight line) and desorption (dashed line) of PA 6 nanofibres (black lines) and microfibres (grey lines) at 23 C. The error bars represent the standard deviation for three samples.
Fig. 2. Evolution of the moisture content in a nanofibrous structure at 95 % relative humidity and 23 C starting from dried fibres.
sorption at equilibrium is only slightly above 1 %, which is about 8 times lower than the moisture sorption of the PA 6 fibrous structures. It can, therefore, be concluded that there is a large difference in the maximal water sorption capacity between the fibres and the resin. Furthermore, above 65 %RH the nanofibrous structures absorb more water compared to the microfibrous structures. To study the effect of the moisture content (the amount of absorbed water) on the curing kinetics, the fibrous structures were saturated at different RH's. 3.2. Influence of the moisture content in the nanofibrous structures on the curing properties First, the reproducibility of the cure of neat DGEBA-MDA resin was examined with DSC. The curing characteristics of four stoichiometric DGEBA-MDA mixtures (all having a molar ratio of [NH]/[Epoxy] ¼ r ¼ 1) were investigated at a cure temperature of 80 C. Fig. 4 shows the heat flow rate of these DGEBA-MDA mixtures as a function of time, proving the high reproducibility. The heat flow rate profiles show the typical autocatalytic behaviour for amine-cured epoxy resins. Next, the cure of the resin in the presence of nanofibrous structures, conditioned at different RH's to obtain different amounts of moisture in the structures, was studied by DSC.
Fig. 3. Evolution of the moisture content of uncured DGEBA-MDA mixture at 95 % relative humidity and 23 C starting from dried resin.
268
S. van der Heijden et al. / Polymer Testing 40 (2014) 265e272
Fig. 4. Heat flow signal for the cure of four neat DGEBA-MDA samples obtained from quasi-isothermal MTDSC experiments at 80 C.
The desired amount of nanofibre nonwoven was placed in the DSC pan and a drop of resin was injected into the pan, penetrating immediately into the nanofibre web. SEM images on cross-sectioned cured DSC samples confirmed the even distribution of the nanofibres throughout the samples (as shown in Fig. 5). Fig. 6A and B present the heat flow signals per gram of resin as a function of the reaction time and reversing heat capacity change. The heat flow signals will be discussed first. Increasing moisture content in the nanofibrous structures has a pronounced accelerating influence on the curing behaviour of the DGEBA-MDA resin system. The effect of the moisture content in the fibrous structures on the curing kinetics is discussed in more detail through four parameters characterizing the cure kinetics: the initial reaction rate, the maximum reaction rate, the time to reach the maximum reaction rate and the reaction enthalpy, as presented in Fig. 7. The initial reaction rate given is the value of the heat flow signal 5 minutes after time zero, ensuring that the heat flow signal has stabilized. At this stage of the reaction, the DGEBA-MDA mixture mainly consists of unreacted resin and hardener, which means that the influence of possible side reactions or byproducts is very minor. Furthermore, the viscosity of the system at this stage of the reaction is relatively low and the reaction is chemically controlled [27]. The maximum reaction rate and the time to reach this maximum reaction
Fig. 6. Heat flow (A) and reversing heat capacity change (B) obtained from a quasi-isothermal MTDSC scan at 80 C, illustrating the effect of the nanofibrous moisture content (expressed as wt % on nanofibre weight) on the curing behaviour of the DGEBA-MDA resin.
rate (tmaxHF) are defined as the maximum value of the heat flow signal with respect to the baseline level at the end of the cure experiment and the time to reach this maximum value, respectively. By dividing the isothermal reaction enthalpy (DH) by the total enthalpy of the reaction, obtained from a non-isothermal MTDSC experiment in which the reaction reaches completion, the conversion reached at the end of the isothermal experiment is found. Although Fig. 6 gives a single curve for each moisture content, five experiments were executed in each case. Fig. 7 gives an overview of the average value as well as the standard deviation for the four kinetic parameters as a function of the moisture content of the nanofibrous
Fig. 5. SEM images of cross section of the cured samples.
S. van der Heijden et al. / Polymer Testing 40 (2014) 265e272
269
Fig. 7. Initial reaction rate (A), maximum reaction rate (B), time to maximum heat flow (tmaxHF) (C) and reaction enthalpy (DH) (D) obtained from quasiisothermal MTDSC experiments at 80 C for neat resin (D) and for samples containing moistened nanofibres (◊). The error bars represent the standard deviation for five samples.
structures. The triangles in the different graphs in Fig. 7 represent the values of the neat dry resin for that specific parameter. The initial reaction rate follows a linearly increasing trend as a function of the moisture content. This implies a first order kinetics in moisture content. Moreover, even when no moisture is present in the fibres, the initial reaction rate is higher: the initial reaction rate increased from 0.017 ± 0.003 W/g for a neat resin to 0.024 ± 0.002 W/g when dry nanofibres are incorporated. This indicates that the nanofibres themselves also have a catalysing effect on the resin cure, which is attributed to a catalysing effect of the functional groups on the fibre surface. A further increase of the initial reaction rate with moisture content is observed up to 0.044 ± 0.006 W/g. Fig. 7B and C again show a linear correlation for both the maximum reaction rate and the time to maximum reaction rate as a function of the moisture content in the nanofibrous structures. The catalytic effect of the dry nanofibrous structure itself is only significant in the early stages of the curing process, as the maximum reaction rate of the neat resin is at the same level as the sample containing nanofibres at 0 wt% moisture content (Fig. 7B). The observed correlation of the maximum reaction rate with the moisture content shows a somewhat lower R2-value, probably due to competition of the influence of absorbed water with the catalysis by hydroxyl groups formed during the reaction. Thus, the initial reaction rate is the most suitable parameter to evaluate the effect of the moisture content on the cure kinetics, since at this stage of the reaction the system mainly contains a mixture of uncured resin, hardener and moisture. The total enthalpy of the reaction, obtained from a nonisothermal curing cycle, and the enthalpy of the quasiisothermal cure experiment were 435 ± 40 J/g and 326 ± 15 J/g for the neat resin (triangle in Fig. 7D), respectively, giving a final isothermal conversion of 75 % for the neat DGEBA-MDA mixture. Statistical ANOVA-tests
indicated that the moisture content of the nanofibrous structures had no significant influence on the reaction enthalpy (see also Fig. 7D). Thus, although the moisture content affects the reaction rate markedly, it has no significant effect on the degree of conversion of the DGEBAMDA mixture attained at the end of the isothermal cure experiments. In addition to the heat flow, the change of the reversing heat capacity is measured in the quasi-isothermal MTDSC experiments (Fig. 6B). The stepwise decrease in heat capacity observed can be attributed to the gradual vitrification of the resin due to the increase of the glass transition temperature of the reacting system to the cure temperature [24,27]. This decrease is due to the loss of segmental or cooperative mobility as the system gradually evolves from a liquid/rubbery material to a glass due to the progressing reaction. The initial rise in reversing heat capacity can be attributed to the higher heat capacity of the reaction products compared to the initial resin [28e30]. The heat capacity evolution also depends on the moisture content of the fibres. Fig. 6B shows that the reversing heat capacity change on vitrification of the neat resin sample is significantly higher than for the nanofibre containing samples: the average value of the reversing heat capacity change expressed per grams of resin was around 0.32 ± 0.03 J/(g C) for all moisture contents, compared to 0.45 ± 0.03 J/(g C) for the neat resin. The difference between the two cases seems to indicate vitrification has a less pronounced effect on the cooperative mobility lost in the case when nanofibres are present. On the other hand, no pronounced effect of the moisture content on the step height of the reversing heat capacity change was observed, indicating that once fibres are present, a similar degree of cooperative mobility is lost on vitrification. Another important parameter that can be obtained from the reversing heat capacity signals is the time to vitrification, characterized here by the time at the half of the decrease in reversing heat capacity. In contrast to the step
270
S. van der Heijden et al. / Polymer Testing 40 (2014) 265e272
height, the time to vitrification is affected by the moisture content. Vitrification occurs sooner with increasing moisture content of the nanofibrous structure (see Fig. 8). This indicates that Tg is rising more quickly at higher moisture content, which is in agreement with catalysis of the DGEBA-MDA reaction by (the moisture present in) the nanofibrous structures. 3.3. Comparison between nanofibrous and conventional structures Since it is now proven that the nanofibres and the moisture in these nanofibrous structures have a catalysing effect, it is also interesting to compare the findings to microfibrous structures. Fig. 9 shows clear differences in initial reaction rate between nanofibrous and microfibrous samples at each moisture content (the mass fraction of fibres is 26 ± 2 wt% in each case). While for the nanofibrous structures the initial reaction rate linearly increases with increasing moisture content, it seems to stagnate for the microfibres around a moisture content of 2.7 wt%. Due to the larger average diameter of the microfibres (Table 1), the moisture present within a microfibre has to diffuse over a larger distance to reach the fibre/resin interface. Moreover, the specific surface area is about 56 times smaller for the microfibrous structures compared to the nanofibrous structures (Table 1). Thus, less surface is available for moisture to diffuse into the resin matrix. In addition, the moisture has to diffuse over larger distances in the resin, due to the larger size of the pores in between the thicker fibres. The diffusion of moisture from the fibres into the bulk of the resin is most probably the rate determining factor, as observed from the DVS measurements. Thus, although the total amount of moisture in the conventional microfibres is approximately the same as in the nanofibres, there is not enough time to complete the diffusion of moisture into the bulk of the resin due to the larger distances over which diffusion is required to take place. 3.4. Influence of the moisture content on the final glass transition temperature of the resin Besides the effect of the moisture content on the curing characteristics, the effect on the final glass transition
Fig. 8. The time to vitrification of the neat DGEBA-MDA system (D) and for samples containing moistened nanofibres (◊). The error bars represent the standard deviation for five samples.
Fig. 9. Comparison between the initial reaction rate of the DGEBA-MDA curing reaction at different fibre moisture contents for neat resin (D), PA 6 nanofibres (◊) and conventional PA 6 micro fibres (▫). The error bars represent the standard deviation for five samples.
temperature was also investigated. Fig. 10 shows the final glass transition temperature for the resin containing nanofibres, compared to the neat resin. An ANOVA-test proved there is no statistically significant difference between the final glass transition temperatures reached with the addition of nanofibres with different moisture contents. There is, however, a statistically significant difference between the final glass transition temperature of the nanofibre-containing resin samples and the neat resin samples. The final glass transition temperature is 169 ± 4 C for the nanofibre-containing resin samples, whereas for the neat resin samples this is 172 ± 3 C. This indicates that the Tg-reducing effect of the addition of nanofibres is to be found in the addition of the nanofibres themselves, and not in the moisture which is brought into the DGEBA-MDA resin by the nanofibrous structures. It is suggested that the cause of the Tg-reducing effect can be found in a preferential migration (or adsorption) of one of the resin components to (or on) the polyamide nanofibres. Similar observations have been made in the past [31e33]. Furthermore, it is also well known that the epoxy-amine ratio has a strong influence on the final glass transition temperature of epoxy resins, the highest glass transition temperature being achieved with stoichiometric ratios [27]. Through preferential migration of a fraction of one of
Fig. 10. Final glass transition temperature Tg of a fully cured DGEBA-MDA resin containing PA 6 nanofibres with different moisture contents (◊), and dry neat resin without fibres (D).
S. van der Heijden et al. / Polymer Testing 40 (2014) 265e272
the components to the nanofibre surface, the stoichiometry of the DGEBA-MDA resin mixture is disturbed throughout the matrix, resulting in the lower final glass transition temperature [28]. 4. Conclusions Nanofibrous structures and the moisture present in them clearly affect the curing characteristics and final glass transition temperature of a DGEBA-MDA mixture. The addition of dry nanofibrous structures has a small catalytic effect on the cure reaction of the DGEBA-MDA resin, observed as an increase in the initial reaction rate. Moisture present in the nanofibrous structures additionally catalyzes the reaction, with the initial reaction rate following a linearly increasing trend as a function of the moisture content. Adversely, conventional PA 6 microfibres result in a smaller catalytic effect, which stagnates for 2.7 wt% moisture and above. This difference was attributed to the larger specific surface area of the nanofibrous structures as well as the smaller distances that need to be covered for diffusion of moisture out of the fibre and into the matrix. Statistical tests proved that the moisture content in the nanofibrous structures had no influence on the reaction enthalpy, and thus also not on the degree of conversion of the DGEBA-MDA mixture reached during isothermal cure. Nevertheless, the final glass transition reached is clearly lower for the nanofibrous materials compared to neat resin. This lowering of the final glass transition temperature in the case of nanofibrous structures was attributed to preferential migration to and/or adsorption of one of the resin components on the fibre surface. Both the catalytic and plasticizing effects are important for the application of these nanofibrous structures in epoxy resin-PA 6 nanocomposites. Acknowledgements Financial support from The Agency for Innovation by Science and Technology, Flanders (IWT) is gratefully acknowledged. Results in this paper were obtained within the framework of the IWT Strategic Basic Research Grant 121156. References [1] B. Nuhiji, D. Attard, G. Thorogood, T. Hanley, K. Magniez, B. Fox, The effect of alternate heating rates during cure on the structureproperty relationships of epoxy/MMT clay nanocomposites, Compos. Sci. Technol. 71 (2011) 1761e1768. [2] B. Ashrafi, J.W. Guan, V. Mirjalili, Y.F. Zhang, L. Chun, P. Hubert, B. Simard, C.T. Kingston, O. Bourne, A. Johnston, Enhancement of mechanical performance of epoxy/carbon fiber laminate composites using single-walled carbon nanotubes, Compos. Sci. Technol. 71 (2011) 1569e1578. [3] J.N. Coleman, U. Khan, W.J. Blau, Y.K. Gun'ko, Small but strong: a review of the mechanical properties of carbon nanotube-polymer composites, Carbon 44 (2006) 1624e1652. [4] L.C. Tang, Y.J. Wan, K. Peng, Y.B. Pei, L.B. Wu, L.M. Chen, L.J. Shu, J.X. Jiang, G.Q. Lai, Fracture toughness and electrical conductivity of epoxy composites filled with carbon nanotubes and spherical particles, Compos. Part A-Appl. S 45 (2013) 95e101. [5] T.W. Chou, L.M. Gao, E.T. Thostenson, Z.G. Zhang, J.H. Byun, An assessment of the science and technology of carbon nanotube-based fibers and composites, Compos. Sci. Technol. 70 (2010) 1e19.
271
[6] H. Qian, E.S. Greenhalgh, M.S.P. Shaffer, A. Bismarck, Carbon nanotube-based hierarchical composites: a review, J. Mater. Chem. 20 (2010) 4751e4762. [7] L. Boogh, B. Pettersson, J.A.E. Manson, Dendritic hyperbranched polymers as tougheners for epoxy resins, Polymer 40 (1999) 2249e2261. [8] K.Y. Kim, L. Ye, Interlaminar fracture toughness of CF/PEI composites at elevated temperatures: roles of matrix toughness and fibre/matrix adhesion, Compos. Part A-Appl. S 35 (2004) 477e487. [9] S.P. Wilkinson, T.C. Ward, J.E. Mcgrath, Effect of thermoplastic modifier variables on toughening a bismaleimide matrix resin for high-performance composite-materials, Polymer 34 (1993) 870e884. [10] I.W. Nam, H.K. Lee, J.H. Jang, Electromagnetic interference shielding/ absorbing characteristics of CNT-embedded epoxy composites, Composites Part A-applied Science and Manufacturing 42 (2011) 1110e1118. [11] B.P. Singh, V. Choudhary, P. Saini, R.B. Mathur, Designing of epoxy composites reinforced with carbon nanotubes grown carbon fiber fabric for improved electromagnetic interference shielding, AIP Adv. 2 (2012). [12] K. Savolainen, L. Pylkkanen, H. Norppa, G. Falck, H. Lindberg, T. Tuomi, M. Vippola, H. Alenius, K. Hameri, J. Koivisto, D. Brouwer, D. Mark, D. Bard, M. Berges, E. Jankowska, M. Posniak, P. Farmer, R. Singh, F. Krombach, P. Bihari, G. Kasper, M. Seipenbusch, Nanotechnologies, engineered nanomaterials and occupational health and safety e a review, Saf. Sci. 48 (2010) 957e963. [13] 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. [14] K. Bilge, S. Venkataraman, Y.Z. Menceloglu, M. Papila, Global and local nanofibrous interlayer toughened composites for higher inplane strength, Compos. Part A: Appl. Sci. Manuf. 58 (2014) 73e76. [15] B. De Schoenmaker, S. Van der Heijden, I. De Baere, W. Van Paepegem, K. De Clerck, Effect of electrospun polyamide 6 nanofibres on the mechanical properties of a glass fibre/epoxy composite, Polym. Test. 32 (2013) 1495e1501. [16] P. Akangah, S. Lingaiah, K. Shivakumar, Effect of Nylon-66 nanofiber interleaving on impact damage resistance of epoxy/carbon fiber composite laminates, Compos. Struct. 92 (2010) 1432e1439. [17] D. Aussawasathien, E. Sancaktar, Effect of non-woven carbon nanofiber mat presence on cure kinetics of epoxy nanocomposites, Macromol. Symp. 264 (2008) 26e33. [18] G.R. Palmese, O.A. Andersen, V.M. Karbhari, Effect of glass fiber sizing on the cure kinetics of vinyl-ester resins, Compos. Part AAppl. S 30 (1999) 11e18. [19] J.C. Cabanelas, S.G. Prolongo, B. Serrano, J. Bravo, J. Baselga, Water absorption in polyaminosiloxane-epoxy thermosetting polymers, J. Mater. Process. Tech. 143 (2003) 311e315. [20] N.D. Danieley, Effects of curing on the glass-transition temperature and moisture absorption of a neat epoxy-resin, J. Polym. Sci. Polym. Chem. 19 (1981) 2443e2449. [21] P. Musto, E. Martuscelli, G. Ragosta, P. Russo, The curing process and moisture transport in a tetrafunctional epoxy resin as investigated by FT-NIR spectroscopy, High Perform. Polym. 12 (2000) 155e168. [22] B. De Schoenmaker, S. van der Heijden, S. Moorkens, H. Rahier, G. van Assche, K. De Clerck, Effect of nanofibres on the curing characteristics of an epoxy matrix, Compos Sci. Technol. 79 (2013) 35e41. [23] P. Westbroek, C.T. Van, V.S. De, C.K. De, Production and use of laminated nanofibrous structures, in, Google Patents, 2008. [24] M. Reading, D. Hourston, B. Mele, H. Rahier, G. Assche, S. Swier, The application of modulated temperature differential scanning calorimetry for the characterisation of curing systems, in: Modulated Temperature Differential Scanning Calorimetry, Springer, Netherlands, 2006, pp. 83e160. [25] C. Sangwichien, G.L. Aranovich, M.D. Donohue, Density functional theory predictions of adsorption isotherms with hysteresis loops, Colloid. Surf. A 206 (2002) 313e320. [26] O. Ceylan, L. Van Landuyt, F. Meulewaeter, K. De Clerck, Moisture sorption in developing cotton fibers, Cellulose 19 (2012) 1517e1526. [27] G. Van Assche, A. Van Hemelrijck, H. Rahier, B. Van Mele, Modulated differential scanning calorimetry: isothermal cure and vitrification of thermosetting systems, Thermochim. Acta 268 (1995) 121e142. [28] S. Swier, G. Van Assche, B. Van Mele, Reaction kinetics modeling and thermal properties of epoxy-amines as measured by modulated-
272
S. van der Heijden et al. / Polymer Testing 40 (2014) 265e272
temperature DSC. II. Network-forming DGEBA plus MDA, J. Appl. Polym. Sci. 91 (2004) 2814e2833. [29] S. Swier, G. Van Assche, B. Van Mele, Reaction kinetics modeling and thermal properties of epoxy-amines as measured by modulatedtemperature DSC. I. Linear step-growth polymerization of DGEBA plus aniline, J. Appl. Polym. Sci. 91 (2004) 2798e2813. [30] S. Swier, B. Van Mele, Reaction thermodynamics of amine-cured epoxy systems: validation of the enthalpy and heat capacity of reaction as determined by modulated temperature differential scanning calorimetry, J. Polym. Sci. Polym. Phys. 41 (2003) 594e608.
[31] G.R. Palmese, R.L. Mccullough, Kinetic and thermodynamic considerations regarding interphase formation in thermosetting composite systems, J. Adhes. 44 (1994) 29e49. [32] M. Munz, Evidence for a three-zone interphase with complex elastic-plastic behaviour: nanoindentation study of an epoxy/thermoplastic composite, J. Phys. D Appl. Phys. 39 (2006) 4044e4058. [33] M. Munz, On the nature of the multi-zone interphase of a thermoset/thermoplastic composite e an analysis employing dynamicmechanical thermal analysis and nanoindentation, J. Adhes. 84 (2008) 445e482.