Carbon Fiber-Reinforced Polyamide 6 Composites Formed by In Situ Polymerization—Experimental and Numerical Analysis of the Influence of Polymerization Temperature

Abstract

In this study, carbon fiber-reinforced polyamide 6 composites were prepared by in situ polymerization via vacuum-assisted resin transfer molding (VARTM) in order to evaluate the effect of processing parameters on the mechanical properties and mold filling ratio. At the molding temperature of 140 °C, a high degree of polymerization of the monomer and impregnation of the resin in the preform were empirically observed. The effect of the molding temperature was analyzed by monitoring the mechanical properties (tensile, bending, and impact strength). The mechanical tests showed that the tensile strength, bending strength, bending modulus, and impact strength were highest for the specimens molded at 140 °C compared to those treated at other temperatures. The experimental results were compared with numerical simulations in order to evaluate the relationship between the mold filling ratio and filling time for large-scale automotive parts with various shapes and sizes. The comparison is of great significance for predicting the mold design as well as processing parameters for large-scale automotive parts.

1. Introduction

Lightweight vehicles have recently been developed to meet the increasingly strict regulations on fuel economy and carbon dioxide emissions. Carbon fiber-reinforced plastics (CFRP) are promising materials for reducing the weight of vehicles due to the high specific strength of these materials. CFRP composites are currently used to produce panels, modules, structures, and other parts, leading to the reduction of the body weight of vehicles for a better fuel economy [1]. Traditionally, carbon fiber-reinforced composites have usually been adopted for structural applications due to their good mechanical properties and thermal stability. In recent years, high pressure resin transfer molding (HPRTM) has been used to reduce the cycle time of carbon fiber-reinforced composites, and is already being used in the series production of structural carbon fiber-reinforced plastic parts in BMW’s i3 and i8 models [2]. The difference between HPRTM and RTM is that the cycle time of HPRTM is much faster than that of standard RTM. Depending on the part size and geometry, the cycle time of standard RTM is 30–60 min with a 10–20 bar injection pressure, whereas that of HPRTM is less than 10 min with a 20–120 bar injection pressure and a pressure of up to 150 bar in the mixing head. Reducing the cost and cycle time are key issues in the automotive industry. Thus, research on fast-cure resins, such as fast-cure epoxy, is underway. According to reports from BMW, they have already produced the parts for the i3 and i8 series with a cure cycle of less than 10 min. Previously, only 1000 to 4000 parts were produced per year, whereas 12,000 to 50,000 parts are now produced annually [3,4]. However, pressures exceeding 100–110 bar may cause washout of the carbon fiber preform and prepreg; moreover, the low processability and recyclability of thermoset polymers limit their mass production.

Recently, owing to their high processability and recyclability, carbon fiber-reinforced thermoplastic composites have gained prominence as one of the alternatives to mass production. In terms of processability, the cycle time of carbon fiber-reinforced thermoplastic composites is faster than that of carbon fiber-reinforced thermoset composites. Thermosetting polymers have a low recyclability because, once cured, they can neither be reheated nor reshaped, and cured thermosetting polymers will only become charred upon heating, whereas thermoplastic polymers can be reheated and reshaped a number of times, which makes carbon fiber-reinforced thermoplastic composites highly recyclable. However, compared to thermoset polymers, the high viscosity of thermoplastic polymers requires a high temperature and high pressure for the materials to be impregnated into carbon fiber reinforcements. For example, thermosetting polymers, such as epoxy, can be impregnated into carbon fiber reinforcements and cured below 200 °C. However, thermoplastic polymers must be heated to their melting temperature, which is typically well above 200 °C, and require high pressures (10–50 bar) for impregnation into carbon fiber reinforcements. In order to prevent fiber misalignment and to improve the preform permeability, thermoplastic RTM (T-RTM) has been developed as an in situ polymerization technique. The polymerization of monomers under the action of catalysts is initiated during the injection process. On the one hand, fast polymerization affords the advantage of reducing the cycle time. However, the mold processing is regarded as being difficult due to the fast gel time during polymerization. Therefore, it is a challenge to control the polymerization time of the monomers. Overall, the performance of automotive parts fabricated from CFRP composites is improving and the cycle time has been decreased to within 5 min [5,6]. For the resin transfer molding (RTM) process, the polymerization kinetics have a key effect on the properties of the products. The in situ polymerization of ε-caprolactam has been studied, with Ahmadi et al. suggesting that the correct ratio of the monomer, catalyst, and activator is key for successful anionic ε-caprolactam polymerization in order to achieve the least monomer residue and the best properties of PA6 samples [7,8,9]. Various types of catalysts and activators have been used for the anionic polymerization of ε-caprolactam, where the polymerization rate, monomer conversion, mechanical properties, electrical conductivity, etc. have been investigated [10,11,12,13,14,15]. Generally, mechanical properties of composites can be affected by factors such as the weight fraction of nanofillers, the degree of dispersion, processing temperatures, pressure, time, etc. [16,17,18]. Using these polymerization techniques, Gong et al. fabricated polyamide single polymer composites via RTM; however, the reaction time is too long for the automotive industry [19,20]. On the other hand, the polymerization of ε-caprolactam has been numerically performed, especially for the catalysts’ and activators’ dependencies. However, only a few studies have focused on reducing the cycle time of the molding processes via the formation of ε-caprolactam-based CFRP composites within several minutes using RTM. Recently, studies of glass fiber-reinforced thermoplastic composites and carbon fiber-reinforced thermoplastic composites with A (caprolactam & activator) mixtures and B (caprolactam & catalyst) mixtures were examined while observing their mechanical properties [21,22]. However, for the studies on polymerization kinetics, few studies were examined. Considering large-scale products, polymerization kinetics and modeling based on experimental data must be examined in order to control the manufacturing processes and cycle time.

In this study with a focus on the conditions for manufacturing carbon fabric-reinforced PA6 composites by vacuum-assisted resin transfer molding (VARTM) method, the effect of the molding temperature on the mechanical properties (tensile, bending, and impact) is investigated. The fracture morphology of the CFRTP specimens manufactured at different temperatures are observed using scanning electron microscopy (SEM). The experimental results are compared with a numerical simulation to evaluate the influence of the mold filling ratio and filling time on the shapes and sizes of large-scale automotive parts.

2. Materials and Methods

2.1. Materials

To form polyamide 6, ε-Caprolactam (DSM, Heerlen, The Netherlands) was used as a monomer. Sodium metal (Daejung Chemical & Metals Co., Ltd., Siheung-si, Korea) and hexamethylene diisocynate (Wako Chemicals, Wako-Shi, Japan) were used as the catalyst and the activator, respectively. For the reinforcement, plane carbon fabric (MUHAN COMPOSITES, Gwangju-si, Korea) with a tow of 3 K (3 thousand filaments per tow) was used (Figure 1), and the properties of the fabric are shown in

Table 1. Specification of carbon fabrics.

Product	Weaving Methods	Width (mm)	Wrap	Fill	Weight (g/m2)
C-120	Plain	1000/1500	Carbon 3 K	Carbon 3 K	200

.

2.2. Specimen Preparation

The carbon fiber-reinforced composite material was manufactured via the VARTM process. The experimental apparatus is schematically illustrated in Figure 2. The monomer (ε-caprolactam) was dried at 40 °C for 24 h in a vacuum oven. For the anionic polymerization of ε-caprolactam, the catalyst (sodium metal) was melted at 110 °C for 1 h in an oil bath under vacuum conditions. To maintain a fiber volume fraction of ca. 45%, thirteen layers of dried carbon fabric were used before injection into the well-sealed mold. The activator was then added to the monomer and catalyst mixture and then injected into the mold at a predetermined temperature (140, 160, 180, or 200 °C) for 5 min for polymerization. Subsequently, the carbon fiber-reinforced polyamide 6 composites (CFRPs) were released from the mold without cooling (Figure 3a,b).

2.3. Characterization and Mechanical Test

The Raman spectra of the samples were measured with a confocal Micro-Raman Spectrometer (NRS-3100, Tokyo, Japan) equipped with a 532 nm wavelength laser in order to determine the polymerization and crystallization temperatures. The samples were heated from 80 °C to 110 °C at a heating rate of 5 °C/min.

Specimens for the mechanical tests (tensile, three-point bending, and Izod impact tests) were prepared according to the ASTM D3039, ASTM D790-10, and ASTM D256 specifications, respectively (Figure 3c). An Izod impact tester (SALT, Incheon, Korea) was used to measure the impact strength. The three-point bending properties and tensile properties were measured using a universal materials testing machine (Instron-5584, Buckinghamshire, UK).

2.4. Morphology Observation

The fractured morphology of the CFRTP specimen was examined with a scanning electron microscope (FE-SEM, 7800F Prime, Tokyo, Japan). To examine the morphology of the interface between the carbon fibers and the resin, the fractured specimen was obtained from the test piece. To avoid a surface charge before scanning, each sample was coated with gold. The accelerating voltage was applied with 5 kV to prevent sample degradation, and the working distance was 9.8 ± 0.4 mm.

3. Numerical Modeling and Implementation

Because ε-caprolactam rapidly undergoes polymerization, the entire process must be carried out within a critical processing time, where the time at which the molding filling process is completed is defined as the critical processing time. Additionally, the critical processing time is an important parameter in molding parts of various sizes and shapes. The viscosity of ε-caprolactam also changes rapidly depending on the polymerization time. Therefore, a critical processing time was defined, and a numerical simulation was performed to evaluate the relationship between the mold filling ratio and filling time.

3.1. Modeling

The resin flow in the carbon fabric in the mold can be modeled as a viscous flow in a porous medium using Darcy’s law:

$${\mathrm{u}}_{\mathrm{i}}=-\frac{{\mathrm{K}}_{\mathrm{ij}}}{\mathsf{\mu }}\frac{\mathrm{dp}}{{\mathrm{dx}}_{\mathrm{j}}}$$

(1)

where ${\mathrm{u}}_{\mathrm{i}}$, K, μ, and p represent the volume-averaged resin velocity, permeability tensor of the reinforcement, resin viscosity, and fluid (resin) pressure, respectively. The resin flow during the impregnation process can be expressed by the incompressible Navier–Stokes equation presented below:

$$\mathsf{\rho }\frac{\partial \mathrm{u}}{\partial \mathrm{t}}+\mathsf{\rho }\mathrm{u}·\nabla \mathrm{u}=-\nabla \mathrm{p}+\mathsf{\mu }∆\mathrm{u}-\mathsf{\mu }\frac{\mathrm{u}}{\mathrm{K}}$$

(2)

where $\mathsf{\rho }$ is the viscosity of the fluid. For the mold cavity, it is assumed that the cavity is uniformly compressed through the thickness; therefore, the continuity equation can be expressed with an incompressible flow as shown below:

$$\nabla ·\mathrm{u}=-\frac{1}{\mathrm{h}}·\frac{\mathrm{dh}}{\mathrm{dt}}$$

(3)

where $\mathrm{h}$ is the thickness of the mold cavity.

To deal with the moving free surface for the current numerical simulation, a fixed grid method was used. Here, the mesh was treated as a fixed reference frame through which the fluid moves [23]. The volume of fluid (VOF) method was employed to treat the change in the calculation domain. By using the fractional fluid volume in a cell on fixed grids, the free surface is shown, with each rectangle in Figure 4 denoting a unit cell [21]. The fractional volume of the fluid, f, is defined for an element variable. The fractional volume in the total domain is divided into two regions, which are either occupied or empty according to the volume fraction of fluid [24]. Every cell is represented with the value of f in the numerical calculation. When the value of f lies between 0 and 1 (0 < f < 1), a cell is considered to be on the free surface [25]. Discontinuity in f is presented according to the following transport equation [26]:

$$\frac{\partial \mathrm{f}}{\partial \mathrm{t}}+\mathrm{u}\cdot \nabla \mathrm{f}=0$$

(4)

where u denotes the velocity vector found by solving the flow field.

3.2. Numerical Implementation

For the current numerical simulation study, a structured and non-adaptive grid (hexahedral grid) with 50,000 elements was used. The convergence for each time step was based on the residual below 10⁻⁴ and 10⁻⁶ for the velocity and for the mass, energy balance, respectively. With the finite volume method, the governing equations and boundary conditions were solved in the commercial CFD package ANSYS-Fluent (ANSYS, version 16.0, Canonsburg, PA, USA).

The time-dependent viscosity of caprolactam was implemented with user-defined function (UDF) subroutines in order to adopt experimentally determined viscosity models. The resin is injected into the mold at a constant pressure. Therefore, the pressure boundary condition was applied for the inlet boundary conditions. The permeability of woven carbon fabric used in this study is K₁ = K₂ = 8.0 × 10⁻¹¹ m², K₃ = 4.0 × 10⁻¹² m². To evaluate the processing time for the overall mold filling process using the preform, a rectangular mold including upper (Figure 5a) and lower (Figure 5b) mold as well as the mesh (Figure 5c), and a mold for an automotive part (Figure 5d) as well as its mesh (Figure 5e) were considered.

4. Results and Discussion

4.1. Analysis of Raman Spectra of ε-Caprolactam

The mixture solution containing ε-caprolactam, the catalyst, and the activator was prepared and frozen under a liquid nitrogen condition to prevent the polymerization reaction. The frozen mixture samples were heated to 80 °C, and then from 80 °C to 180 °C at a heating rate of 5 °C/min; the corresponding Raman spectra are shown in Figure 6. Spectral differences were evident in the CC stretch region (1000–1150 cm⁻¹) based on the temperature at which polymerization begins. The peak corresponding to the reactant (~1025 cm⁻¹) persisted from 80 °C to 100 °C, but started to lose intensity at 110 °C, indicating that the polymerization of ε-caprolactam starts at this temperature.

According to a previous report [27], different types of crystal structures, including the α-phase, β-phase, and γ-phase, can be formed in the region between 1100 and 1150 cm⁻¹. In the present study, a weak band was observed at 1124 cm⁻¹ for the sample treated at 160 °C, indicating that the γ-phase crystal structure starts to form at 160 °C and continues to be formed until 180 °C. Hence, temperature is a critical point for the in situ polymerization and crystallization of ε-caprolactam, which may affect its processing time and mechanical properties.

4.2. Mechanical Properties

The tensile and bending properties of CFRTP composites were determined through a tensile and three-point bending test, respectively. First, the specimen for the tensile test was cut to dimensions of 100 × 13 × 3.75 mm³ according to ASTM D790-10. Figure 7 shows the results of the tensile test for the CFRP composites treated at different molding temperatures. As shown in Figure 7, the tensile strength and the strain of the CFRPs molded at 140 °C were slightly higher than for the samples molded at other temperatures.

To determine the bending strength and bending modulus, each specimen was cut to dimensions of 100 × 13 × 3.75 mm³ according to ASTM D790-10. Figure 8a,b show the bending strength and bending modulus, respectively. The bending strength and bending modulus of the CFRPs molded at 140 °C were higher than those of the samples molded at other temperatures.

The higher tensile and bending strengths of the sample molded at 140 °C indicate a relatively large degree of polymerization and crystallization at this temperature, as deduced from our previous study on the effect of temperature on the polymerization and crystallization of PA 6 [28]. Overall, the sample molded at 140 °C had superior properties.

The Izod impact test was performed based on ASTM D256 in order to evaluate the impact resistance of the specimen in each case. The dimensions of each specimen were 61.5 mm (height) × 11.5 mm (depth) × 3.75 mm (width), with a notch depth of 2.54 mm. Figure 9 shows the impact strength of the CFRP composites that were molded at 200 °C, where the impact strength was much higher than that of the samples molded at other temperatures.

The impact strength of the sample impregnated with carbon fibers and molded at 200 °C was higher than that of the other samples. When the sample was molded at 200 °C, a relatively higher amount of unreacted monomer persisted. In the fracture test, the specimens molded at 200 °C fractured via the partial break mode, whereas the samples molded at other temperatures fractured via the complete break mode. Therefore, the impact strength of the CFRPs molded at 200 °C was higher than that of the samples molded at other temperatures.

4.3. Morphology

To observe the morphology of the interface between the fibers and resin, the fracture surfaces of the CFRTP specimens manufactured at different temperatures were examined via a SEM observation, as shown in Figure 10. The image of the fracture surface shows that the carbon fibers were uniformly impregnated into the carbon fabrics for the polyamide 6 resins molded at 140, 160, and 180 °C, whereas the fibers pulled out of the resins that were molded at 200 °C. This is due to unreacted monomer in the latter, and the results are consistent with the higher impact strength of the CFRPs molded at 200 °C compared with that of the samples molded at other temperatures.

4.4. Numerical Simulation

The mold filling process, where fluid resin is placed in the mold, was investigated through a numerical study using a rectangular mold and a mold for an automotive part. A 180 × 180 mm² sized rectangular mold and 1200 × 1700 mm² sized automotive part (hood) mold were considered for the mold filling process. From our previous research [28], the experimentally determined time-dependent viscosity of caprolactam was employed, and the critical processing time for ε-caprolactam was defined as 20 s. A numerical analysis was performed with four different temperature conditions: 140, 160, 180, and 200 °C, where a difference in the filling behavior with the processing time was expected based on the difference in the viscosity of the resin. The inlet pressure was set to 0.85 bar, and the outlet pressure was atmospheric pressure for all cases. To simulate the mold filling process, the flow in the mold was simplified in the porous zone, with four inlets at each edge and one outlet in the middle part for the rectangular mold (Figure 11a) and with three inlets at each edge and one outlet for the automotive part mold (Figure 11b), because, in the actual case, the resin flows along the runner first and then flows into the porous zone.

Figure 11 shows the phase contour of the resin in the mold filling process for the rectangular mold with carbon fabrics and the automotive part mold with carbon fabrics. The resin phase contour is used to observe the filling behavior of the resin during injection into the mold at a constant predetermined pressure. The mold filling process using the resin was completed within 5 s for the rectangular mold, which coincides with the experimental data in Figure 12a. For the automotive part mold with carbon fabrics, different predetermined pressures of 10, 20, 30, 40, and 50 bar were applied to determine the critical processing time, and the time-dependent viscosity of polymerizing ε-caprolactam at 140 °C was set by using a user-defined function for this numerical analysis. The simulation shows that the mold filling process using the resin can be completed within 20 s at a pressure of 40 bar, which is the lowest pressure at which to fill the mold completely (Figure 12b). The results show that the entire manufacturing process should be conducted within this critical processing time (20 s).

5. Conclusions

Carbon fabric-reinforced PA6 composites were manufactured by vacuum-assisted resin transfer molding (VARTM) at various molding temperatures. The effect of the molding temperature on the mechanical properties (tensile, bending, and impact strengths) were examined. The observation of the fracture surfaces of the CFRTP specimens manufactured at different temperatures were conducted. The analysis of the Raman spectra of caprolactam showed evident spectral differences in the CC stretch region (1000–1150 cm⁻¹) based on the temperature at which polymerization began. The peak of the reactant (~1025 cm⁻¹) persisted from 80 °C to 100 °C, followed by a loss of intensity from 110 °C, indicating that the polymerization of ε-caprolactam started at 110 °C. The tensile strength and strain of the CFRTP composites molded at 140 °C were slightly higher than those of the samples molded at other temperatures. The bending strength and bending modulus of the samples molded at 140 °C were also higher than those of the other samples. These superior properties may result from the relatively large degree of polymerization and crystallization at this temperature. The impact strength of the CFRP composites molded at 200 °C was much higher than that of the samples molded at other temperatures due to poor impregnation at this temperature, with fractures occurring via the partial break mode. The experimental data were compared with a numerical simulation. The filling time obtained from the simulation was similar to the experimental results. A numerical study was also performed to evaluate the processing conditions for automotive parts with various shapes and sizes. From the experimental and simulation results, the optimum temperature for polymerization is considered as being in the range of 140 °C–160 °C. Additionally, the molding process must be conducted within 20 s at a pressure of 40 bar. The present study on the molding conditions for carbon fabric-reinforced PA6 composites can be used in future work to determine a varying range of design factors for the manufacturing process.