Curing and thermal degradation of diglycidyl ether of bisphenol A epoxy resin crosslinked with natural hydroxy acids as environmentally friendly hardeners

2.1 Materials

Citric acid (CA) (m.p. = 152–154 °C), L-tartaric acid (TA) (m.p. = 170–172 °C) and TEBAC (m.p. = 190–193 °C) were analytical grade products and used as received. The epoxy resin of DGEBA type provided by Sintofarm, Bucharest, was a commercial product with a molecular weight of about 390 g mol⁻¹, epoxy equivalent of 190, viscosity of 15,000 mPas (measured with Brookfield Viscometer, stil 4, at 25 °C) and used without further purification.

2.2 The sample preparation for the curing test and thermal experiment

The test samples entitled DGEBA/CA and DGEBA/TA were obtained by manually mixing until to complete homogenization of the liquid DGEBA resin, CA and TA as fine powder, at molar ratio 1/0.75 epoxy ring/carboxylic proton in presence of TEBAC as catalyst (1% based on the monomers weight). For comparison, samples obtained in the absence of catalyst were also tested. The choice of the 1/0.75 epoxy ring/carboxyl proton ratio is justified by the presence of hydroxyl groups in the structure of two acids, which may exhibit etherification reactions with the epoxide groups excess, with the increase of the crosslinking temperature. Then, for the removal of the included air, the samples were placed in an oven at 40 °C and degassed under vacuum in about 30 min. After cooling at 0 °C, a small part was used for DSC measurement at three heating rates under nitrogen atmosphere. The majority of the sample was placed in aluminum moulds with the next dimensions: length = 3 · 10⁻² m, width = 2 · 10⁻² m height of about 0.5 · 10⁻² m and cured in an oven, after next thermal cycle: at 80 °C (1 h), 120 °C (1 h) and post cured at 160 °C for 5 h. The fact that DSC curves of the cured samples present only exothermic peak suggest that the curing reaction was complete. Finally, the cured samples were grounded as fine powder and used for thermal analysis.

2.3 Measurements

The epoxy equivalent was determined as in literature by titration with hydrogen bromide (ASTM D 1652-04) and has 190 value. Fourier transform infrared (FT-IR) spectra were recorded with a Vertex 70 spectrophotometer (Bruker-Germany) on KBr disks, in the range of 4000–400 cm^–1. The penetration resistance of the crosslinked products was obtained using the Shore hardness test and recorded with a Shore D-type apparatus (SAUTER HB-Germany). The depth penetration of the device head in tested sample as the average value of 5 determinations, was calculate as Shore hardness. Vickers hardness was obtained with Vickers hardness tester (Shimadzu Japan) under load of 1.961 N and dwell time of 12 s. The values of hardness were reported as an average of 6 determinations. Degree of cross-linking of epoxy resins was measured based on the literature method by extraction into solvent mixture (xylene/DMF, 1:1 v/v) [Patil et al., 2017]. The cured samples, ground as fine grain, placed in a stainless steel wire cloth, was immersed into solvent mixture, maintained at room temperature (25 °C) one day, filtered and dried under vacuum at 150 °C to constant mass. The gel content (Gc) was calculated with the next equation:

The SEM/ESEM – EDAX – QUANTA 200 apparatus was used to register the scanning electron microscopy images. The next parameters were used: high vacuum, field emission filament 15 kV accelerating voltage and magnification of 5000×. The samples obtained by fracturing after cooling with liquid nitrogen and coated with fine gold layer were used for registration.

The DSC measurements for the curing reactions were conducted on a DSC 200 F3 Maia device (Netzsch, Germany) in a dynamic mode. The DSC curves were registered at different heating rates (5, 10, 15 °C min⁻¹) under nitrogen atmosphere (50 mL min⁻¹) with the temperature ranging from 25 to 300 °C. Samples weighing up to 12 mg, were weighed carefully, encapsulated in sealed aluminum pans and heated in the presence of an empty crucible as reference, for each heating rate. It was recorded heat flow versus temperature. Previously, the device was calibrated with indium, under nitrogen atmosphere, for all the three heating rates. The measurements were made consecutive from the same sample in the same day (Vyazovkin et al., 2014).

The kinetic parameters of the curing reactions were determined using model-freeestimation methods of Kissinger (KS) and Ozawa-Flynn-Wall (OFW) (Kissinger, 1957; Ozawa, 1976; Flynn and Wall, 1966), based on the next Eqs. (2) and (3):

From the graph lnβ versus 1/T_p and ln (β · T_p⁻²) versus 1/T_p, taking into account the slope and the intercept, the pre-exponential factor and activation energy of the crosslinking reactions are obtained.

The thermal stability of the crosslinked epoxy resins were thermo gravimetrically analyzed using a STA 449 F1 Jupiter apparatus (Netzsch-Germany), coupled to a Vertex 70 spectrophotometer for FT-IR analysis and Aëolos QMS 403C mass spectrometer (Netzsch-Germany) for the mass spectroscopic analysis of the evolved gases. Samples up to 15 mg, placed in Al₂O₃ crucibles were thermal degraded at the heating rates of 5, 7.5, and 10 °C min⁻¹ under nitrogen atmosphere in the temperature range between 25 and 600 °C. Initially, the kinetic parameters with Friedman (FR) and OFW methods included in the “Thermokinetics-3” software (Netzsch “Thermokinetics-3”, version 2008.05), were calculated. Based on the activation energy variation of degradation versus conversion degree (obtained with FR method) the mode in which thermal degradation occurs can be determined. Choosing the most probable kinetic model for the thermal degradation, the obtaining of the kinetic parameters can be made using the “Multivariate regression non-linear” method from the above software. The identification of the chemical structures of the gases evolved in the degradation time was made with FT-IR analysis and mass spectroscopic analysis. The resulted gases by thermal decomposition are transferred through a line made of polytetrafluorethylene with 1.5 mm diameter and 1 m length, maintained at 190 °C to TGA-IR external modulus equipped with MCT (Mercury Cadmium Telluride) detector. FT-IR spectra were registered in 3D size with OPUS 6.5 software. For characterization of the gases by mass spectroscopy, the evolved gases are transferred at the mass spectrometer through a quartz capillary with 2 m length and 75 μm diameter, maintained at 290 °C. The working parameters were: ionizing energy with electron impact of 70 eV, vacuum 10⁻⁵ mbar, the signals m/z scale up to 200 amu, time for each cycle 100 s.

3.1 Curing of epoxy resin with HA as hardeners

The possible crosslinking reaction between DGEBA epoxy resin and hydroxy acids (TEBAC as catalyst) is presented in Fig. 1. In the presence of TEBAC, the acid is deprotonated and occurs an anioncarboxylate as a nucleophilic agent. Next, the esterification reaction takes place by the attack of the anioncarboxylate to the epoxide ring, preventing thus the homopolymerization of epoxy resin. (Matejka and Dusek, 1986). In our case, TEBAC was used as catalyst, but is it embedded in the final product. This type of reaction is confirmed by FT-IR analysis. As it can be seen in Fig. 2 (FT-IR spectrum for DGEBA/TA and DGEBA/CA) the increasing of the value of the absorption band located at 3411 cm⁻¹ confirms the formation of the new hydroxyl groups due to the reaction between epoxy and carbonyl groups. Also, this is confirmed by the disappearance of the signal from 915 cm⁻¹ specific to epoxy ring and the appearance of the new signal located at 1741 cm⁻¹ specific to C⚌O group from the new ester groups, as well as the signal located at 1181 cm⁻¹ (C—O—C group) specific to ester group which overlaps by the C—O—C group of ether type from epoxy resin. Because of the very similar chemical structure of the two acids, the IR spectra of the crosslinked products are almost identical. The only observable difference is located in the 850–960 cm⁻¹ range, the peaks for DGEBA/TA crosslinked product in this region being larger than that of DGEBA/CA crosslinked product. These peaks can be attributed to ether linkages, which may be more because TA possesses two hydroxyl groups. In Fig. 3 are presented the DSC curves recorded at three heating rates for DGEBA/HA samples (in presence of TEBAC) which show a single exothermic peak for each heating rate. This suggests that reactions between carboxylic protons and epoxy ring take place simultaneously at all two or three carboxyl groups. With the temperature increase the crosslinked products are obtained. The peak temperature values for both systems are presented in Table 1. They grow with increasing of heating rates and are different for each system. With the increase of the heating rate the peak temperature is shifted to higher values, this fact can be attributed both to the kinetic considerations and to the delay of the instrument response (Rosu et al., 2001; Roudsari et al., 2014; Tripathi et al., 2015) observable degradation using Eqs. (2) and (3), the kinetic parameters of curing reaction can be calculated. The different values of the activation energy obtained for the same sample is due to the different approximation methods used to solve the two equations. For DGEBA/CA sample, these values ranges from 61 and 67 kJ mol⁻¹ and for DGEBA/TA sample varies between 65 and 71 kJ mol⁻¹. This fact is in concordance with the literature for DGEBA crosslinked with acids as hardeners (Mustata and Tudorachi, 2016). For non-catalyzed samples, the shift of the exothermic peak (Fig. 4) to the higher temperatures justifies the use of the catalyst in the crosslinking processes. The variation of activation energies versus conversion degree calculated with FR method (Netzsch “Thermokinetics-3”, version 2008.05) (Fig. 5) indicates how the crosslinking process occurs. The fact that the shape of activation energies curves presents maxima and minima suggests that the crosslinking processes are complex processes (Worzakowska, 2007; Edelmann et al., 2007). The most probable kinetic model and the kinetic parameters can be obtained using the “Multivariate regression non-linear” method from Netzsch “Thermokinetics-3” software. Based on the initial values of Ea and pre-exponential factor obtained with Friedman method, the multivariate linear regression software resolve of the numerical differential equations, using 16 different kinetic models included in program. It may determine the most probable global kinetic model based on differences between the calculated and experimental data. The software uses a modified Marquardt procedure using Runge–Kutta analysis. The calculations conducted with multivariate linear regression software were based on the experimental data collected at three heating rates and were performed by software on the conversion range of 0.15 and 0.85. The next curing reaction mechanisms with successive reactions were chosen (see Table 2):

Close

Fig. 1

The possible mechanism of DGEBA the crosslinking with CA.

Export to PPT

Close

Fig. 2

FT-IR spectra of: (a) DGEBA, (b) (DGEBA/CA), and (c) (DGEBA/TA).

Export to PPT

Close

Fig. 3

Simulated and experimental DSC curves of (a) DGEBA/CA and (b) DGEBA/TA at: (■) 5 °C min⁻¹, (●) 10 °C min⁻¹ and (▴) 15 °C min⁻¹ (symbols represent the simulated curves and lines represent experimental curves).

Export to PPT

Close

Fig. 4

Experimental DSC curves of (a) DGEBA/CA and (b) DGEBA/TA obtained at 10 °C min⁻¹ in absence of catalyst.

Export to PPT

Close

Fig. 5

Dependence of crosslinking activation energy and pre-exponential factor versus the conversion degree for: (a) DGEBA/CA and (b) DGEBA/TA (calculated with Friedman method).

Export to PPT

For DGEBA/TA sample: $$\text{A-}1\to \text{B-}2\to \text{C-}3\to \text{D}\text{with}\text{t}:\text{f,f;}\text{Bna,}\text{An,}\text{Fn}\text{reaction}\text{code}$$ where: A are the initial reactants, B and C are the intermediate products and D is the final crosslinked product, t:f,f; represent the three-steps successive reaction schemes and 1, 2, 3 denote the reaction steps.

And for DGEBA/CA sample: $$\text{A-}1\to \text{B-}2\to \text{C}\text{with}\text{d}:\text{f;}\text{CnB,}\text{An,}\text{reaction code}$$ where the meanings are the same as above.

The conversion functions are:

• –

  expanded Prout–Tompkins equation

• –

  reaction order n_th with autocatalysis model CnB

• –

  n_th reaction order, model,

• –

  Avrami–Erofeev reaction model

Using the data presented in Table 3, the recalculated curves (Fig. 3) for the reaction models taking place into three-steps with consecutive reactions of t:f,f; Bna, An, Fn (sample DGEBA/TA) and d:f; CnB, An, (Sample DGEBA/CA) types, are in good concordance with the experimental data, suggesting that the kinetic models fairly describe them. Thus, for the DGEBA/TA sample, the activation energies varies between 59 and 101 kJ mol⁻¹ with correlation coefficient of 0.9580 and for DGEBA/CA sample between 56 and 76 kJ mol⁻¹ with correlation coefficient of 0.9955.

3.2 TG/FT-IR/MS characterization of the cured products

The thermal behavior of the cured products was estimated using simultaneous TG/FT-IR/MS analysis. In Fig. 1 is presented the most probable chemical structure of the crosslinked DGEBA/CA system. In Fig. 6, are shown TG and DTG curves for these systems registered at three heating rates (5, 7.5, 10 °C min⁻¹) under nitrogen atmosphere. From these curves, the main parameters of the degradation process: T_onset - temperature for initial decomposition, T_peak - temperature at maximum decomposition, T₁₀ - temperature at 10 wt% weight loss, T₂₀ - temperature at 20 wt% weight loss, T₅₀ - temperature at 50 wt% weight loss, T_GS - the maximum temperature of Gram-Schmidt curve from IR spectrum and W - weight loss at 600 °C) are summarized in Table 3. As it can be seen from Fig. 6, with the increase of the heating rate TG curves are shifted to higher temperatures. This phenomenon is often found in this kind of determinations, because the sample temperature is proportionally exceeded by the furnace temperature due to their thermal inertia at high heating rate (Rosu et al., 2011; Mustata and Tudorachi, 2010). As it can be seen in Fig. 6 and in Table 3, the thermal decomposition begin near 200 °C for DGEBA/TA system and the weight loss does not exceed 5 wt%. For DGEBA/CA system the thermal decomposition begins near 300 °C. Also, it can be seen that DGEBA/CA sample shows only observable degradation peak and the degradation onset temperature has much higher values than DGEBA/TA sample. For the same heating rate the peak temperature has up to 30 °C higher values. As can be seen, DGEBA/CA presents only one thermal degradation process with significant mass losses (about 90%), while DGEBA/TA presents three thermal degradation processes with mass loss for each degradation step: (I = 2.78–4.10%, II = 9.75–10.20%, III = 75.45–76.12%). This confirms that DGEBA resin was almost completely crosslinked with CA, whereas in the case of TA the crosslinking degree is much lower. This may be a consequence of the higher number of carboxyl groups from CA. For both samples the residual mass at 600 °C was about 10–11%. The difference between measured temperatures for 10 and 50 wt% mass losses for both samples, is preserved. The experimental data were processed using “Thermokinetics 3” software. Initially, this software allows the calculation of the kinetic parameters using FR and OFW model free methods. These methods permit to calculate the variation of the activation energy of the thermal degradation versus the conversion degree. The data from Table 3, the shape of DTG curves (Fig. 6) and variation of activation energies (Fig. 7) indicate that the thermal degradation processes are complex processes (Worzakowska, 2007; Edelmann et al., 2007). The shape of DTG and Ea curves suggests the entire process of degradation is governed by at least two different reaction mechanisms. These processes can take place in more steps and can run as consecutive, parallel or competitive processes. The identification of the thermal degradation mechanisms and the calculation of the corresponding kinetic parameters, were made with isoconversional differential method of Friedman and “Multivariate regression non-linear” method from “Thermokinetics-3” software based on the data recorded at different heating rates (Netzsch “Thermokinetics-3”, version 2008.0519; Opfermann, 2000). The software has done the calculations based on the experimental data collected at three heating rates on the conversion range of 0.15 and 0.85.

Close

Fig. 6

TG and DTG curves recorded at 5, 7.5 and 10 °C min⁻¹ for: (a) DGEBA/CA and (b) DGEBA/TA, (symbols represent the experimental values and lines represent the calculated values).

Export to PPT

Close

Fig. 7

Dependence of thermal decomposition activation energy and pre-exponential factor versus conversion degree for: (a) DGEBA/TA and (b) DGEBA/CA (calculated with Friedman method).

Export to PPT

The next conversion Eqs. (3)–(6) for one single step was used:

• –

  three-dimensions diffusion (Jander’s type) D3:

• –

  three-dimensions diffusion (Ginstling-Brounstein type) D4:

• –

  Avrami–Erofeev reaction model An:

• –

  reaction order n_th model, with Fn:

The calculations were effectuated for a greater number of mechanisms from which were selected the following successive processes which had place in three or four steps: $$\text{A-}1\to \text{B-}2\to \text{C-}3\to \text{D}$$ and $$\text{A-}1\to \text{B-}2\to \text{C-}3\to \text{D-}4\to \text{E}$$ with the following types of the processes schemes: t:f,f; D3,Fn,An (the intermediate solid products are A, B, C and D is the solid residue) and q:f,f,f; D4,An,An,Fn, (the intermediate solid products are A, B, C, D and E is solid residue), the numbers 1, 2, 3, 4 representing the reaction steps.

Their choice was made using the greatest statistic coefficients [the experimental F-value (Fexp), Fcrit (0.95) and correlation coefficients]. In Table 4 the obtained kinetic and statistic parameters for the above models, are listed. Based on the above mechanisms and kinetic parameters listed in Table 4, the recalculated TG curves on the interval 200–600 °C show a good agreement with experimental data, suggesting that the theoretical models fit very well with the real phenomenon (Fig. 6). As it can be seen in Table 4 for some degradation stages, the reaction order greater than 1 was recorded, suggesting that during the degradation process may appear some degradation products with high molecular weight (Rosu et al., 2011). The value of reaction orders less than 1 suggests that at this thermal degradation stage, the majority of volatile products are low molecular weight products. The obtained activation energies and pre-exponential factors (Table 4) are different in function of the chemical structure of curing agents. DGEBA/TA activation energy values between 61 and 234 kJ mol⁻¹ while for DGEBA/CA between 55 and 314 kJ mol⁻¹. This fact is in concordance with the chemical structure of CA which has three carboxyl groups and can induce higher degree of crosslinking with increase of thermal stability, consequently. Also, the conclusion that CA vs. TA can induce a higher crosslinking degree was confirmed by the Shore and Vickers hardness tests, as well as the gel percentage. Also, the conclusion that CA vs. TA can induce a higher crosslinking degree was confirmed by Shore and Vickers hardness tests as well as the percentage of gel. If considering T_peak,W₆₀₀ parameters and activation energies of the thermal decomposition as thermal stability criteria, it can be seen that only DGEBA/CA sample has the same degree of thermal stability in comparison with DGEBA crosslinked with p-aminobenzoic acid or with dicarboxylic aromatic acids [Rosu et al., 2011; Mustata et al., 2015; Mustata and Tudorachi, 2016]. By comparison (Fig. 8) it can be seen that for the samples obtained in the absence of the catalyst, several degradation peaks occur with the maxima at lower temperatures than those obtained in the presence of the catalyst. This can be a consequence of a lower crosslinking degree in absence of the catalyst.

Close

Fig. 8

TG and DTG experimental curves recorded at 10 °C min⁻¹ for: (▴) DGEBA/CA and (○) DGEBA/TA, crosslinked without catalyst.

Export to PPT

3.3 Chemical composition of the evolved gases

In Fig. 9 is shown the most likely mechanism of thermal destruction of the cured sample. The chemical structure of gases that result from thermal degradation was identified using FT-IR and MS analysis. In Fig. 10 are shown the two and three-dimensional FT-IR spectra (2D and 3D spectra) for the gases resulted at the thermal degradation of DGEBA/TA sample registered at 10 °C min⁻¹ heating rate the temperature range of 40–600 °C. 2D and 3D spectra of the evolved gases at the degradation of DGEBA/CA sample, recorded in the same conditions as above are similar and are not presented from the reason of brevity. In FT-IR 3D spectrum is shown the absorbance versus the wavelength and temperature, while in 2D spectrum is plotted the absorbance versus the wavelength at the maximum temperature from Gram Schmidt plot (424 °C). As it can be seen in 3D spectrum at the temperature of 424 °C, the intensity of the absorption bands shows a maximum. In Fig. 10b can be observed a relative large number of signals that are located at 3655, 3254, 2963, 2439, 2354, 1729, 1605, 1510, 1250, 1180, 1039 and 827 cm⁻¹. The significance of these signals was assigned based on visual observation and on the data from literature (Silverstein et al., 2005). Thus, the signals located between 3200 and 3700 cm⁻¹ can be assigned to water vapors and alcohols which can appear at the thermal degradation of the secondary hydroxyl, ester or ether groups. The signals present at 3040, 2800 cm⁻¹ and 1400, 1300 cm⁻¹ can be assigned to the vibration of CH, CH₂ and CH₃ groups located in the chemical structure of saturated and unsaturated aliphatic hydrocarbons or aromatic compounds, which originate from the degradation of acids or DGEBA moieties. The strongest signal present at 2354 cm⁻¹ and can be certainly attributed to carbon dioxide, which can occur as result of ester group degradation (Worzakowska and Scigalski, 2014). Also, at 2180 cm⁻¹ is present one small signal which suggests that an amount of carbon monoxide can appear. It would be derived from the degradation of the hydroxyl groups located on the hydroxy acid molecule. The signal located at 1729 cm⁻¹ can be assigned to the C⚌O groups from acids, esters, anhydrides, aldehydes products occurring at the degradation of ester groups formed after cross-linking reaction. The signals located at 1605, 1510 and 829 cm⁻¹ are specific for the aromatic derivatives such as phenol, benzene, toluene, α-methyl styrene etc. resulted at the cracking of DGEBA moieties. Also, for these substances the signal located at 3040 cm⁻¹ which is specific to the C⚌C aromatic bonds, could be assigned. The signals situated at 1250 and 1180 cm⁻¹ can be assigned to C—O group, located in ester type gases.

Close

Fig. 9

The possible mechanism of cured DGEBA/TA thermal degradation.

Export to PPT

Close

Fig. 10

Stacked plot diagram (a), FT-IR spectrum (b) and MS spectrum (c) of evolved gases obtained at 424 °C for DGEBA/TA.

Export to PPT

The assignments in the IR spectra are supported by MS representations from Fig. 10c. From Fig. 10c (for brevity only MS spectrum for DGEBA/TA is presented) the majority of signals are located up to m/z = 70 and are specific to more species. The identification of the chemical structure of gases resulted after thermal degradation was made using data from NIST MS library on the base of molecular weight of ionic fragments (http://webbook.nist.gov/chemistry/name-ser.html). The ionic fragments located up to 50 amu can be assigned to water (m/z = 18, 17), carbon dioxide (m/z = 44, 28, 16, 12), saturated and unsaturated aliphatic hydrocarbons [methane (m/z = 16, 15, 14, 13, 12), propane (m/z = 44, 43, 39, 29), ethylene (m/z = 28, 27, 26, 24, 14) propene (m/z = 42, 41, 40, 39, 38, 37, 27, 26, 15), acetylene (m/z = 26, 25, 24, 13)]. These ionic fragments may be derived by degradation of aliphatic moieties of the two-components. The fragments which appear in the range of 50–60 amu can be allocated to aldehyde, ketone, ester or acid type products, and may originate especially from degradation of the hardener moieties [2-propenal (m/z = 56, 55, 37, 29, 28, 27, 26, 25), acetone (58, 43, 42, 27,15), formic acid (m/z = 46, 45, 42, 29, 28, 17), acetylaldehyde (44, 43, 42, 29, 15), acetic acid (m/z = 60, 45, 43, 42, 29, 15), methyl formate (m/z = 60, 32, 31, 29, 15)]. The above substances can result mainly by degradation of the ester linkage obtained from the reaction between epoxide groups and carboxyl groups of the hydroxy acids. The signals with values over 70 amu can be assigned to the components of aromatic type [benzene (m/z = 78, 77, 50, 39), phenol (m/z = 94, 66, 65, 39), toluene (92, 91, 65, 51, 39) isopropylbenzene (m/z = 120, 105, 103, 91, 79, 77, 51, 39)] obtained from the thermal degradation of DGEBA moieties. Such gaseous products have been identified at the degradation of DGEBA thermosetting resins (Mustata et al., 2011). In Fig. 11, are shown the evolved gases profiles on the temperature range 100–600 °C for some MS signals (for clarity were chosen for presentation only few signals which appear at the thermal degradation of DGEBA/TA sample) specific to water, carbon dioxide, methane, propane, acetylene, acetic acid, benzene, phenol, isopropyl benzene. For the DGEBA/CA sample, the evolved gases profile is relatively similar. According to the degradation model from Fig. 9 correlated with Figs. 10 and 11, it can be considered that at lower temperatures (about 200 °C) takes place the breakdown of the secondary hydroxyl and ester groups, with the obtaining of water, carbon dioxide and carboxyl products (acids, esters, ketones, aldehydes), while at higher temperatures (about 400 °C) aromatic products (benzene, phenol, toluene) are resulted.

Close

Fig. 11

The variation of ion current intensity with temperature at heating rate of 10 °C min⁻¹ for the main signals of the evolved gases at DGEBA/TA sample.thermal degradation.

Export to PPT

3.4 Morphological study of the cured samples

In Table 5 is shown the Shore and Vickers hardness for the samples crosslinked in the presence and absence of TEBAC. As can be seen, the DGEBA/CA samples have higher hardness compared to DGEBA/TA samples crosslinked in presence of the catalyst. The same trend is maintained in the case of crosslinked samples in absence of the catalyst. The toughness of DGEBA/HA systems can be correlated with the chemical composition of the hardener. The aspect of the fracture surfaces (SEM micrographs-magnification of 5000) obtained after cooling at liquid nitrogen temperature for DGEBA/HA crosslinked samples are shown in Fig. 12. The appearance of the fractured surface allows making a correlation between the morphology and toughness of samples. In Fig.12a for DGEBA/CA sample, the cracks of the surface appear as smooth glass in which the fracture lines are parallel lines without any proof of deformation, indicating that the thermoset product has a poor impact strength (Mustata et al., 2013b). This is due to the chemical structure of CA which induces a higher degree of crosslinking of DGEBA/CA sample, so a more rigid structure resulted with, a brittle behavior. In Fig.12b is shown the surface fracture for DGEBA/TA sample where it can be observed that the surfaces present the fracture lines with more ridges and tortuous cracks. These forms of the fracture lines show that, due to chemical structure of DGEBA/TA sample (lower degree of crosslinking), the breaking has ductile characteristics. The gel content also confirms the different crosslinking degrees between DGEBA/CA and DGEBA/TA samples. Similar characteristics were observed on the fracture surface of DGEBA thermoset resin crosslinked with maleopimaric acid (Mustata et al., 2014).

Close

Fig. 12

SEM micrographs of: (a) DGEBA/CA; (b) DGEBA/TA at magnification of 5000.

Export to PPT