Study on the curing behavior of polythiol/phenolic/epoxy resin and the mechanical and thermal properties of the composites

Phenolic/epoxy resin (EP-PF) composites were prepared, in which phenolic resin and epoxy resin was used as matrix, polythiol and triethanolamine as curing accelerators. The uniform experimental design method was used to obtain the scheme, in which the resin system had minimum curing temperature. The curing kinetics and the apparent activation energy of the resin system were studied and calculated by the differential scanning calorimetry., and the mechanical and thermal properties of the composite was analyzed. The results show that when the content of polythiol was 12% and the content of triethanolamine was 11%, the peak curing temperature of the resin system was 118°C, which was lower than the phenolic resin or phenolic/epoxy resin. Tg dropped from 212 °C of PF to 157 °C of EPF and then to 147 °C of EPF-B. The thermal decomposition temperature and residual carbon rate also showed a slight downward trend, but the mechanical properties were greatly improved. The strength and flexural modulus of EPF-B have increased from 291.4 MPa and 11.2 GPa of EPF to 440 MPa and 12.3 GPa, an increase of 49% and 27%. This research provides a theoretical basis for broadening the application range of phenolic resin and epoxy resin blending system.


Introduction
In recent years, polymer matrix composites have made many breakthroughs in mechanical properties, heat resistance, and electrical properties [1,2]. The choice of resin matrix is particularly important. Phenolic resin itself has good flame retardant properties, lower smoke rate, and less harmful gases [3][4][5]. However, the curing of traditional phenolic resin requires a higher temperature and a slower rate. The use of strong acid/alkali catalysts can cause corrosion of the device, and it is easy to release small molecules such as amines during curing, which will cause the internal or surface of the product [6][7][8]. Nowadays, the requirements for various indicators of phenolic resin are getting higher and higher. In order to expand its application range and meet the needs of civilian and military industries, it is necessary to modify phenolic resin [9]. Epoxy resin is an excellent polymer matrix with excellent processing and mechanical properties [10,11]. Therefore, the blending system of phenolic resin and epoxy resin not only has the heat-resistant and flame-retardant properties of phenolic resin, but also retains the excellent mechanical properties of epoxy resin [12,13]. Moreover, the problem of easy foaming of phenolic resin is also improved in the process [14][15][16]. However, a single phenolic resin and epoxy resin blending modification method still cannot meet the requirements of practical applications [13]. The phenolicepoxy blend system still has the problems of high curing temperature and slow curing rate. Most of the reports focus on the modification of toughening and flame retardant properties of the phenolic-epoxy resin system. Deng P used a new type of phosphorus-containing phenolic resin to cure epoxy resin, and studied the curing process and the corresponding crosslinking structure and mechanical properties through differential scanning calorimeter and dynamic mechanical thermal analysis [17]. S Yue and J.-C. Munoz and others studied epoxy resin modified phenolic resin [18,19]. Reducing the curing temperature will bring great convenience in process Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. production and also save energy [20]. Therefore, how to reduce the curing temperature of the blend system is an inevitable problem in the rapid prototyping process research process [21]. The research work of S Han showed that in the same epoxy phenolic curing system, when different curing accelerators are used, the curing process presents different curing reaction mechanisms [22]. S Han studied the effect of the amount of curing accelerator on the curing of epoxy-phenolic resin system [23]. As a common epoxy resin curing agent at room temperature, Polythiol is not only low in price, but also can effectively reduce the activation energy required for epoxy resin ring opening, thereby reducing the curing temperature, and the introduction of a large number of flexible segments can also achieve the effect of toughening the resin system [24][25][26]. Modification of the phenolic-epoxy blend resin system is a necessary means to improve the process and the comprehensive performance of composite materials.
In this paper, phenolic resin and epoxy resin (EP-PF) are used as the resin matrix, Polythiol and Triethanolamine are used as curing accelerators to prepare EP-PF composite materials. The curing kinetics of the composite was studied by the differential scanning calorimetry (DSC), and the apparent activation energy was calculated by the Kissinger method and the Ozawa method. Furthermore, the thermal and mechanical properties of the composites under different proportions of curing accelerators were studied [27,28]. Studying the effect of curing accelerator on the curing process of epoxy-phenolic resin system and the effect of cured product performance has important guiding significance for optimizing the process characteristics and material properties of epoxy-phenolic resin system [29][30][31]

Uniform test plan design
The three factors set in this experiment are the mass ratio of phenolic resin to epoxy resin, the mass ratio of polythiol to epoxy resin, and the mass ratio of triethanolamine to epoxy resin. The peak temperature T p measured by DSC is used as the evaluation. Select the appropriate level range to design a three-factor seven-level uniform test plan, and choose U 7 (7 4 ) uniform design table as shown in table 1 [32][33][34].

Preparation of fast-curing phenolic epoxy resin composite
Blend phenolic resin, epoxy resin, polythiol and triethanolamine according to the proportions in table 2 and mix them to form a low-viscosity resin mixture. Brush them evenly on the cut glass fiber cloth, so that it will be saturated. After the glass fiber cloths were placed at room temperature for 3-4 days, they were stacked and placed

Characterization
Differential Scanning Calorimetry (DSC): The curing reaction peak temperature T p and curing reaction kinetics of the blend system are carried out by a DSC4000 differential scanning calorimeter. Test conditions: the temperature range is 30°C-250°C; the heating rates are 5°C min −1 , 10°C min −1 , 15°C min −1 , 20°C min −1 ; nitrogen atmosphere. Thermogravimetric analysis (TGA): The test instrument is the STA449F3 synchronous thermal analyzer from NETZSCH from Germany. Test conditions: temperature range is 30°C-800°C; heating rate is 10°C min −1 ; nitrogen atmosphere.
Dynamic thermomechanical analysis (DMA): The glass transition temperature T g of the sample is analyzed by the Pyris Diamond DMA dynamic thermomechanical analyzer. Test conditions: the temperature range is 30°C-300°C; the heating rate is 5°C min −1 ; and the fixed frequency is 1 Hz. The sample size is: 60 mm×10 mm×3 mm.
Mechanical performance test: The flexural modulus and strength of samples were tested by CSS-44300 electronic testing machine according to the standard of GB/T 2567-2008. The tensile strength and elastic     Regression analysis was performed based on the results of the above-mentioned test scheme [35], and the results obtained are shown in tables 4-6.
Four regression equation models can be obtained through stepwise regression analysis. The closer the model R 2 is to 1, the higher the credibility of the model. From table 4, it can be seen that the R 2 of Model 4 is the closest to 1. Therefore, model 4 is selected as the fitting reference model, and the regression equation of model 4 is as follows: It can be seen from the regression equation that within the test range, when x 2 takes the maximum value and x 1 takes the minimum value, y has the minimum value. That is x 1 =10 and x 2 =1.2. The partial derivative of the regression equation can be obtained when x 3 =1.09, and the optimal conditions are x 1 =10, x 2 =1.2, x 3 =1.1. The optimized formula is: Phenolic resin 10 g, polythiol 1.2 g, and triethanolamine 1.1 g. According to the residual statistics in table 6, the minimum predicted value of the regression equation is 118.3942, which is less than the T p value of either the uniform test or the single factor test.

Optimal scheme test verification
According to the preferred solution (10 g Phenolic resin, 10 g Epoxy resin, 1.2 g Polythiol, 1.1 g Triethanolamine), after preparing the resin, perform non-isothermal DSC test with heating rate of 5°C min −1 , 10°C min −1 , 15°C min −1 , 20°C min −1 respectively. The test result is shown in figure 1.             figure 3. We can see that the curing exothermic peak shifts to the left, which means that with the addition of epoxy resin and polythiol, the curing reaction temperature shows a downward trend. The curing reaction temperature of EPF-B is much lower than that of EPF and PF.

Curing kinetics of optimal scheme
It is a common analysis method to use DSC data to derive the curing kinetic parameters. Assuming that the curing process conforms to one dynamic model by using the model fitting method, then the kinetic parameters can be obtained. Different kinetic parameters are mainly based on the Kissinger equation to calculate the reaction activation energy and pre-exponential factor of the resin system, the Flynn-Wall-Ozawa equation to calculate the apparent activation energy of different reaction degrees, and the Crane equation to calculate the curing reaction order n [36,37]. The main kinetic models and equations of calculation parameters are shown in table 8.

Kinetic analysis of non-isothermal curing
As is shown in figure 4, ln(β/T p 2 )∼1/T p is used to plot a linear fit. The result of the linear fitting is: y=8.680-7.594x. E a =63.14 KJ mol −1 , A 0 =4.468×107/s can be calculated by Kissinger equation. The addition of Polythiol and Triethanolamine reduces the activation energy of the epoxy resin. The simultaneous presence of thiol groups and tertiary amines makes it easier to open the epoxy ring, thereby reducing the overall activation energy of the resin system.
In the Flynn-Wall-Ozawa formula in table 9, since G(α) is a function of the curing reaction degree α, even at different heating rates, when the curing reaction degree α is the same, the corresponding G(α) value is also equal [38]. Plot lnβ versus 1/T, as is shown in figure 5. According to the Flynn-Wall-Ozawa equation, the slope of the fitted curve is the apparent activation energy Comparing the activation energy obtained by the two methods, the apparent activation energy obtained by the Kissinger method is slightly higher than that of the Flynn-Wall-Ozawa method, but there is not much difference between the two values. It can be considered that the value is reliable. Calculate the average of the two as the final apparent activation energy, E a =60.62 KJ mol −1 .
The reaction order can be calculated by the Crane equation, in which T p is much smaller than E/nR, so 2T p is ignored in the analysis and fitting [39], and the curve shown in figure 6 is obtained by linear fitting lnβ to 1/T p . According to figure 6, the fitting equation is y=22.74-8.43x, and the reaction order n=0.904 can be calculated by Crane equation.  Integrating all parameters into the n-order catalytic reaction model, the dynamic equation of the optimal scheme is obtained as follows:

Analysis of isothermal curing kinetic
In the isothermal DSC test, whether the kinetic model of the reaction is an n-order reaction or an autocatalytic reaction can be judged by the time when the maximum exothermic reaction rate occurs. When the maximum reaction rate occurs at t=0, it means that the reaction is an n-order reaction. And the maximum reaction rate of the autocatalytic reaction is generally at 30∼40% of the reaction degree [40]. The relationship between the curing reaction rate and time obtained after the curve of the above figure is differentiated. It can be seen from figure 7 that the curing reaction rate reaches the maximum in a short time and then gradually decreases, so we can determine that the resin system conforms to the Kamal autocatalytic model. After differential treatment, plot dα/dt against α, as is shown in figure 8, it can be found that the curing reaction rate first increases and then decreases with the increase of curing degree. At the same time, the higher the constant curing temperature, the higher the peak value of the curve, which means that the curing reaction rate is faster.
According to figure 8, the Kamal autocatalysis model and the n-level reaction model were used to perform nonlinear fitting to it, and the results are as follows.
From the non-linear fitting results of the two models in figure 9, the fitting results of the Kamal autocatalytic model are highly consistent with the experimental calculation results, and the multiple regression coefficient R 2 is close to 1, while the fitting results of the n-order reaction model are consistent. The results of the Kamal autocatalysis model fitted under four constant temperature conditions are shown in table 10. The kinetic equation of the Kamal autocatalysis model can be obtained by averaging its parameters.
The kinetic equation of the Kamal autocatalysis model can be written as follows:

Thermomechanical properties
Dynamic thermomechanical analysis is used to study the thermomechanical properties of EPF resin cured products. The following figure 10 shows that the relationship between storage modulus (E′) and loss tangent (Tanδ) with temperature. In the curve of Tanδ and temperature, the temperature corresponding to its peak is defined as the glass transition temperature (T g ). It can be seen from the table 11 that the glass transition temperature and storage modulus of the three composite material samples have gradually decreased. The T g of the phenolic resin composite material is the highest, reaching 212°C. After adding epoxy resin, the Tg decreases to 157°C. After adding polythiol and triethanolamine, T g decreases slightly, but the storage modulus decreases significantly. This is because epoxy groups can react with phenolic hydroxyl groups, and polythiol can be regarded as both a curing agent and a toughening agent. Their addition makes a large number of flexible segments connected to increase the chain length, which increases the relative molecular weight between the cross-linking points to reduce the cross-link density. This lowers the glass transition temperature. The introduction of a large number of flexible segments also played a role in toughening, making the storage modulus gradually decrease. By comparison, it can be found that the addition of a small amount of polythiol will slightly reduce the glass transition temperature, but it can greatly reduce the modulus to achieve a toughening effect.

Thermal stability
The following figure 11 shows the TG and DTG curves of the cured resin. When the weight loss rate is 5%, the temperature point is the initial thermal decomposition temperature (T 5% ). The peak temperature of the DTG curve is the temperature (T max ) at the maximum weight loss rate, and the maximum weight loss rate (R max ) is at 800°C. The mass retention rate is regarded as the residual carbon rate (CY). All data are listed in table 12.
The DTG curve of all resin cured products has only one peak, that is, there is only one weight loss stage. From the above table, we can see that with the addition of epoxy resin, polythiol and triethanolamine, the T 5% , T max , R max and CY of the cured resin decreased slightly. But they basically maintained high ablation resistance and residual carbon rate. This phenomenon is due to the fact that the addition of epoxy resin and polythiol introduces a large number of flexible segments. Compared with the more benzene ring structure in phenolic resin, the thermal stability is poor, which makes T 5% show a significant downward trend. The addition of amine curing accelerator will make the phenolic resin and epoxy resin crosslink more completely, so that the T max and R max of the cured product of EPF-B component are not much different from EPF. Phenolic resin will gradually be carbonized into residues at high temperatures. The higher the content of phenolic resin, the higher the residual carbon rate of the cured product. Therefore, with the addition of other components, the residual carbon rate will gradually decrease.

Mechanical properties of the composite
The flexural strength, flexural modulus, tensile strength and elastic modulus of composite materials are shown in table 13. Compared with PF, composite materials of EPF component has increased flexural strength, flexural modulus, tensile strength, and elastic modulus by 61%, 99%, 28% and 19% respectively. The addition of epoxy resin can appropriately increase the cross-linking density of the cross-linking system to make the internal connection of the cured product closer, thereby improving the mechanical properties of the composite materials. Compared with EPF, composite material of EPF-B component has increased flexural strength and flexural modulus by 49% and 27% respectively. And the tensile strength and elastic modulus are slightly reduced but still higher than composite material of PF component. Since the addition of polythiol will introduce a large amount of flexible chain structure, it is equivalent to adding a toughening agent. This will greatly increase the flexural strength and flexural modulus of the composite material. The addition of tougheners tends to decrease the tensile strength and elastic modulus.

Conclusions
Design a suitable uniform test plan to test, and perform stepwise regression analysis on the results to get the regression equation as: = --+ + y 147.630 47.065x 14.126x 0.306x x 20.179x x 3 2 13 33 The optimal solution EPF-B can be obtained through the regression equation as x 1 (phenolic resin)=10 g, x 2 (Polythiol)=1.2 g, x 3 (Triethanolamine)=1.1 g. The minimum value of statistical prediction is 118.3924, and the test value of verification test 118.24 is close to the statistical forecast value. The apparent activation energy obtained by the Kissinger method and the Flynn-Wall-Ozawa method were 63.137 KJ mol −1 and 58.10 KJ mol −1 , respectively. Take their average value as E a . The pre-referential factor A 0 is 4.468×10 7 /s, and the reaction order n is 0.904. The curing kinetic equation is obtained as follows: DMA research shows that T g dropped from 212°C of PF to 157°C of EPF and then to 147°C of EPF-B. The thermal decomposition temperature and residual carbon rate also showed a slight downward trend. The mechanical properties were greatly improved. The strength and flexural modulus of EPF-B have increased from 291.4 MPa and 11.2 GPa of EPF to 440 MPa and 12.3 GPa, an increase of 49% and 27%. The tensile strength and elastic modulus are also greatly improved compared to PF. The curing system obtained by uniform experimental design solves the problem of high curing temperature and slow curing speed. At the same time, through this curing method, a cured product with excellent thermal and mechanical properties is obtained.