Introduction

Reactive polyamides of dimer fatty acids are versatile products used as epoxy-curing agents. Most of these reactive polyamides are synthesized using C36 dimer fatty acids and polyamines; their epoxy-curing products are flexible, but have poor thermal resistance and poor rigidity because of the C36 dimer fatty acid's long carboxyl chain. Therefore, the applications of conventional reactive polyamides are limited by their heat-resistance and suppleness. To solve this problem, diacid with a shorter carboxyl chain should be used to replace the C36 dimer fatty acid.

In this study, a series of reactive polyamide resins were prepared by a new methyl of C21 diacid made from tung oil and polyamines. Several studies have referred to C21 diacid; most of the C21 diacids are synthesized by the Diels–Alder reaction of tall oil fatty acids or dehydrated castor oil fatty acids and acrylic acid at a high reactive temperature.1, 2, 3 The tall oil fatty acid contains about 50% linoleic acid, which can react with acrylic acid through the Diels–Alder reaction, hence the yield of C21 diacid from tall oil fatty acid is about 50% by molar. The yield of C21 diacid from dehydrated castor oil fatty acids is about 90% by molar, but this requires many reaction steps and is thus costly. In addition, because of the unconjugated double bonds of tall oil and dehydrated castor oil fatty acids, a catalyst such as iodine or phosphate is required so that the Diels–Alder reaction can be carried out, and it is difficult to separate the catalyst from the product.

Tung oil fatty acids contain about 85% eleostearic acid, which has three conjugated double bonds in the molecular backbone. The Diels–Alder reaction tends to proceed easily for eleostearic acid molecules that have conjugated double bonds without any catalysts. As a result, the reaction temperature of tung oil fatty acids and acrylic acid is 180 °C, which is much lower than that of the reaction between tall oil fatty acids or dehydrated castor oil fatty acids and acrylic acid (about 250 °C). Energy consumption is reduced by using the methyl of tung oil fatty acids to prepare C21 diacid derivatives. In addition, as there is 85% eleostearic acid in tung oil fatty acids, the yield of C21 diacid is about 85% by molar.

Reactive polyamides synthesized by C21 diacids were reported by Vijayalakshmi et al.,4 and it was found that the coating properties of the epoxy-cured materials obtained from C21 diacid were superior to those obtained from the C36 dimer acid in tensile strength and had better hardness and adhesion. However, the C21 diacid that was reported by Vijayalakshmi et al.4 was synthesized from tall oil fatty acid, but not from tung oil fatty acid. The C21 diacid made from tung oil is a new kind of C21 diacid and has not yet been reported. Furthermore, the curing kinetics of C21 diacid reactive polyamide with epoxy resins have not been previously mentioned in the literature.

The final properties of cured epoxy are intensively dependent on the chemical structure and degree of crosslinking within the polymer network. Thus, it is necessary to be informed of the curing kinetics of the epoxy-curing agent. Curing kinetic parameters are generally determined by analyzing the experimental results obtained by different thermal analysis techniques.5, 6, 7, 8, 9 Differential scanning calorimetry (DSC), both in the scanning and non-isothermal modes, has been used extensively in this manner.10, 11, 12, 13, 14

This paper reports the mechanical properties of C21 diacid-based polyamides from tung oil curing DGEBA and curing kinetics based on non-isothermal DSC measurements, applying the simple and consistent method of Málek15, 16, 17, 18, 19, 20 to the kinetic analysis of data obtained by thermal treatment.

Experimental procedure

Materials

Tung oil (Figure 1a) that was converted into methyl fatty acids and analyzed by gas chromatography-mass spectrometry was obtained from Wuxi Jinyu Oil Corporation (Wuxi City, Jiangsu Province, China). Methyl eleostearate makes up 85.5% of the methyl of fatty acids, according to gas chromatography mass spectrometry results.

Figure 1
figure 1

Synthesis of C21-based reactive polyamide. (a) Tung oil converted into methyl fatty acids. (b) Methyl of C21 diacid. (c) C21 diacid-based reactive polyamide.

Full size image

Methyl eleostearate was obtained by transesterification of tung oil and methanol.21 The methyl of C21 diacid (Figure 1b) was prepared by the Diels–Alder reaction of methyl eleostearate with acrylic acid. DGEBA was obtained from the Wuxi Resin Factory (Wuxi City, Jiangsu Province, China) with an epoxy equivalent weight of 196.

Characterization

Fourier transform infrared spectra were recorded using a Nicolet 550 infrared spectrometer (Nicolet Corp., Madison, WI, USA) in the range of 500–4000 cm−1 on KBr pellets.

With CDCl3 solvent, 1H nuclear magnetic resonance (1H NMR) spectra were recorded on a Bruker DPX-300 NMR (Bruker Corp., Beijing, China), and tetramethylsilane was used as the internal standard. The operating parameters for 1H NMR were as follows: sweep width, 300 MHz; pulse width, 5.80 ms; number of scans, 16.

Synthesis

Here, the typical procedure for the preparation of C21 diacid-based reactive polyamide (Figure 1c) is described. Methyl of C21 diacid (100 g) was placed in a 500 ml, four-necked round-bottom flask equipped with a stirrer, a nitrogen inlet tube, a thermometer and a Dean–Stark apparatus. After heating to 140 °C in 0.5 h, the required quantity of polyamine was added dropwise over a period of 1.5 h. The temperature was then gradually increased to 200 °C over a period of 1.5 h, and the reaction was continued at this temperature for a further 3 h. The water and methanol liberated during the reaction were collected in the Dean–Stark apparatus. The reaction was monitored by determining the acid values. When the acid value of the product decreased below 1 mg g−1, nitrogen supply was discontinued and volatiles were removed under vacuum (0.67–1.33 kPa) for 1 h; the product was then cooled and stored. The steps involved in the preparation are shown in Figure 1.

Sample preparation

Epoxy resin DGEBA and the C21-based reactive polyamide were carefully and homogenously mixed at the optimum ratio. Samples of 3–4 mg were separated from the freshly prepared compositions for calorimetric measurements. Samples for mechanical property tests were prepared by molding the epoxy mixtures in iron molds, and samples were cured in a vacuum oven using a defined cure program.

Methods

The mechanical properties of cured materials were examined using a CMT4303 Sans testing machine (Sans Corp., Shenzhen City, Guangdong Province, China) according to standards as follows:

Tensile strength, tensile strain and tensile modulus were tested following GB/T 2568-1995. The bending strength and modulus were tested following GB/T 2570-1995. Compression strength was examined following GB/T 2569-1995 and lap shear strength was tested following GB 7124-86.

Differential scanning calorimetry experiments were run on a Perkin–Elmer Diamond DSC calorimeter (Perkin-Elmer Corp., Waltham, MA, USA) to analyze data. Samples were placed in an aluminum pan and then covered by an aluminum lid before undergoing a specific cure program, depending on the mode of measurement. Scanning DSC measurements were carried out at four heating rates of 5, 10, 20 and 40 °C min−1 from 30 to 250 °C in a nitrogen atmosphere. Pure indium was used as the standard for calorimetric calibration. Runs were carried out using an empty cell as reference. Heat flow data, as a function of temperature and time, were obtained using the area under the peak of the exothermal. These data were processed further to obtain the fractional conversion and rate of reaction. Raw data were converted into Excel files and then kinetic analysis was applied.

The basic assumption for the application of the DSC technique to the curing of thermosetting polymers is that the rate of the kinetic process (dα/dt) is proportional to the measured heat flow φ.18, 20

where ΔH is the enthalpy of the curing reaction.

The rate of the kinetic process in kinetic analysis can be described by Equation (2)

where k(T) is a temperature-dependent reaction rate constant and f(α) is a mechanism-dependent kinetic model function.20, 21

The rate constant is dependent on temperature as shown in Equation (3)

where A is the preexponential factor, and Ea is the apparent activation energy.

The kinetic parameters of the curing reaction, with special reference to Ea, can be calculated using various computational methods.22, 23 Activation energy was determined by the iso-conversion method,16 using the logarithmic form of the kinetic rate (Equation (2))

The slope of ln (dα/dt) versus 1/T for the same value of α gives the value of Ea. The value of Ea can then be used to find the appropriate kinetic model that best describes the conversion function of the studied process. For this, it is necessary to appeal to the special functions, y(α) and z(α).15, 16

where x is the reduced activation energy (Ea/RT), β is the heating rate (K min−1), T is the absolute temperature (K) and π(x) is the expression of the temperature integral. As pointed out previously,15 the π(x) function can be well approximated using the fourth rational expression of Senum and Yang24 as in Equation (7)

Function y(α) is proportional to f(α) function, being characteristic for a given kinetic model. The shape and maximum of both y(α) and f(α) functions for several models, normalized within the (0, 1) interval, provide valuable information for determining the most suitable kinetic model to characterize the process studied.

Results and discussions

Apparent properties of reactive polyamides

Reactive polyamides having amine values ranging from 374 to 496 were prepared by reacting the C21 diacid with different polyamines, such as diethylene triamine, triethylene tetramine and tetraethylene pentamine, at two different molar ratios (1:1.2 and 1:1.5). Polyamides were prepared from the C36 dimer acid and diethylene triamine (in the mole ratio 1:2) for comparison. The apparent properties of these polyamides are given in Table 1.

Table 1 Properties and sample numbers of curing agents derived from different monomers
Full size table

The data in Table 1 show that the polyamides with lower amine values have higher viscosities than those with higher amine values because of their lower molecular weights. Polyamides derived from C36 dimer acids are lighter in color than those derived from the methyl of C21 diacid.

Figure 2 shows the 13C-NMR spectra of the methyl of C21 diacid (Figure 1b). The peaks at δ 14.1, 51.5, 174.5 and 181.7 can be assigned to –CH3, –OOCH3, –COO– and –COOH, respectively. Peaks from δ 22.9–43.5 can be attributed to methylene (–CH2–). Peaks at δ 127.4 and 128.3 can be assigned to cyclohexene ( ).

Figure 2
figure 2

13C-NMR spectra of the methyl of C21 diacid.

Full size image

Figures 3a–c show the Fourier transform infrared spectra of methyl eleostearate and (a) acrylic acid, (b) C21 diacid and (c) C21-based reactive polyamide, respectively. For (a), the reactant acrylic acid was confirmed by the appearance of C=C stretching and vibration absorption peaks at 1635, 966 and 993 cm−1, and in product (b) these peaks disappeared. The methyl of C21 diacid (b) was already confirmed by 13C-NMR, as shown in Figure 2. Figure 3c shows the C21-based polyamide's structure. The condensation reaction between the methyl of C21 diacid (b) and polyamines was confirmed by the appearance of a C=O stretching absorption peak at 1650 cm−1, as well as by the appearance of a strong broad characteristic absorption peak of –N–H stretching vibration at 3280 cm−1 and that of an absorption peak of C–N stretching vibration at 1120 cm−1.

Figure 3
figure 3

Fourier transform infrared spectra of reactants and products, methyl eleostearate and acrylic acid (a); C21 diacid (b) and C21-based reactive polyamide (c).

Full size image

Glass transition temperature and mechanical properties

The glass transition temperatures of cured DGEBA for samples 1–4 were measured by DSC. The results are shown in Figure 4. The glass transition temperatures of the four samples were found to be 53, 108, 109 and 116 °C, respectively. The glass transition temperature of the cured resin increased as the amine value increased; this might be because of the increasing crosslink density and polarity of polyamides. The glass transition temperature of C21-based polyamides is about twice that of the C36-based polyamide because of shorter molecular chains in C21-based polyamides.

Figure 4
figure 4

Differential scanning calorimetry curves for the glass transition temperatures of cured diglycidyl ether of bisphenol A.

Full size image

The mechanical strength and modulus were tested by a Sans testing machine. The tensile strength and strain are shown in Figure 5. As the amine values of samples increased from samples 1 to 3, tensile strength increased. However, sample 4 was found to be more flexible than the C21-based polyamide-cured resins, because it had a low tensile strain, similar to that of sample 3. C21-based polyamides tend to be hard and rigid, whereas C36-based polyamides are more flexible. The flexural strengths of the four samples are shown in Figure 6. By increasing amine values, the flexural strength and modulus increased significantly. This is because of the increased crosslink density of the polymers.

Figure 5
figure 5

Tensile strength and strain of the polyamide/diglycidyl ether of bisphenol A system.

Full size image
Figure 6
figure 6

Flexural strength and modulus of the polyamide/diglycidyl ether of bisphenol A system.

Full size image

The shear strength, thermal deformation temperature and shore D hardness values are shown in Table 2. Shear strength is strongly dependent on the flexibility of the polyamide backbone. Owing to the shorter and more rigid molecular chain in C21-based polyamides, their shear strengths are lower than those of C36-based polyamides. As amine values increased, crosslink density increased, producing a more rigid polyamide backbone. Thus, the shear strength was increased by raising the amine values of polyamides. The thermal deformation temperature and shore D hardness values have the same trend as the glass transition temperature.

Table 2 Properties of four samples cured diglycidyl ether of bisphenol A
Full size table

Curing kinetics

Dynamic DSC scans for mixtures of C21-based polyamides and DGEBA were collected at different heating rates (5, 10, 20 and 40 °C min−1). Typical DSC curves for sample 2 are presented in Figure 7. Similar results were found for the other C21-based polyamides with different amine values. As can be seen in Figure 7 and the resulting exotherm, the rate of curing reaction increases with increasing temperature.

Figure 7
figure 7

Typical differential scanning calorimetry curves recorded for the sample 2/diglycidyl ether of bisphenol A system at different heating rates.

Full size image

Figure 8 plots the conversion against the dynamic cure temperature at various heating rates. It shows that, at the same conversion, the iso-conversion temperature increases when the heating rate is increased. Plots of cure rate versus cure temperature are shown in Figure 9. The cure rate passes through a maximum point and decreases as a function of the curing time. Furthermore, the peak cure rate increases with increasing cure temperature.

Figure 8
figure 8

Conversion as a function of temperature for the sample 2/diglycidyl ether of bisphenol A system.

Full size image
Figure 9
figure 9

The kinetic rate as a function of temperature for the sample 2/diglycidyl ether of bisphenol A system.

Full size image

Plots of ln(dα/dt) versus 1/T at different conversions are shown in Figure 10. The fractional activation energy for the conversion from 0.1 to 0.9 can be calculated from the slope, on the basis of Equation (4). These results are summarized in Table 3. Arrhenius plots from DSC scans recorded for cured samples using Equation (4) and the same α-value are straight lines, the slopes of which give the value of Ea. Figure 9 shows the plot obtained for sample 2, in the interval of 0.1<α<0.9, whereas Figure 11 shows the variation of Ea for all studied samples as a function of the conversion. Figure 11 plots the values of the activation energy for different values of conversion, showing that Ea gradually increases at the early stage of the reaction. This may be because of consumption of the accelerator and a decrease in the mobility of the reactive groups of the partially cured epoxy. The value of Ea is roughly constant in the interval of 0.2<α<0.8, with a slight tendency to increase or decrease for α-values beyond the above-mentioned interval. These changes are probably due to errors in the baseline approximation for peak tails, which lead to high error levels in the calculated activation energy.25 It is well known that primary amine is more reactive than secondary amine. It can be observed that sample 1, with the highest primary amine value (Table 1), has the lowest activation energy (Table 3). This further demonstrates that sample 1 has a higher reactivity than the other two samples.

Figure 10
figure 10

Arrhenius plots for sample 2 in the interval of 0.1<α<0.9.

Full size image
Table 3 The kinetics parameters of three samples
Full size table
Figure 11
figure 11

Variation of Ea versus conversion for sample 2 in the interval of 0.1<α<0.9.

Full size image

The value of Ea was used to calculate both y(α) and z(α) functions using Equations (2) and (6), respectively. Figures 12 and 13 show the variation of y(α) and z(α) versus conversion. Table 4 lists the maxima values αM and αp∞ corresponding to functions y(α) and z(α) for all studied samples.

Figure 12
figure 12

Variation of the y(α) function versus conversion for sample 2.

Full size image
Figure 13
figure 13

Variation of the z(α) function versus conversion for sample 2.

Full size image
Table 4 The values of αM and αp∞ for each sample
Full size table

As noted, the data in Figures 12 and 13 and those in Table 4 show that the values of αM are lower than those of αp∞, whereas the values of αp∞ are lower than 0.632. This indicates that the studied curing process can be described using the two-parameter autocatalytic kinetic Šesták–Berggren model (Equation (9)).26

where m and n are the kinetic exponents.16

The kinetic exponent n can be obtained by the slope of ln[(dα/dt)ex] versus ln[αp(1−α)] (from Equation (2)), and m=pn, where p=αM/1−αM.

Table 3 lists some kinetic parameters evaluated for the proposed Šesták–Berggren kinetic model. As shown in Table 3, the variation of kinetic parameter values with the heating rate is within the experimental error limit (within 10% of the average value).

The accuracy of the kinetic model proposed using the Šesták–Berggren equation was verified by plotting dα/dt versus temperature (experimental curves), using the data listed in Table 3, and compared with the curves obtained by computational processing of thermograms (calculated curves). An example is given in Figure 14 for the sample 2/DGEBA system. It was found that the two-parameter Šesták–Berggren model describes the studied curing process well.

Figure 14
figure 14

Comparison of experimental (symbols) and calculated (full lines) differential scanning calorimetry curves for the sample 2/diglycidyl ether of bisphenol A system.

Full size image

Conclusions

The epoxy-cured material obtained by using C21-based polyamides had better tensile strength, flexural strength and flexural modulus values than those cured by C36-based polyamides. In addition, the thermal properties of the C21-based polyamide-cured material are superior to those obtained from C36-based polyamides with regard to glass transition temperature and thermal deformation temperature.

The curing kinetics of DGEBA in the presence of C21-based polyamides were examined by DSC at various heating rates. The Ea values of samples 1–3 determined by the iso-conversional method were 52.4, 55.4 and 54.4 kJ mol−1. The values of Ea are dependent on the amine values of polyamides. The Ea value of the curing reaction decreased as the primary amine values of curing agents increased. It was established that the two-parameter autocatalytic model (the Šesták–Berggren equation) is suitable for describing the studied curing process. The kinetic parameters that were experimentally determined were used to calculate theoretical DSC curves. These showed a good agreement with experimental data.