Effects of precure cycle on tensile and dynamical mechanical properties of carbon/benzoxazine laminates

2.1 Materials and precure degree designs

2.1.2 Preparation of specimens

Square (300×300 mm²) panels of 1.6-mm and 5-mm nominal thickness were made from Loctite BZ 9704 Aero prepregs by autoclave processing following the quasi-isotropic stacking sequence, as shown in Figure 1. Prepregs were precured in an oven (Yongguangming-DZF-1, Beijing, China) at a temperature of 80°C and pressure of 0.1 MPa for different time intentionally in order to achieve a different precure degree. It could be considered as an accelerating procedure of prepreg shelf period. But one-to-one correspondence between shelf time at storage temperature and precure time at 80°C was not included in this study because such work is time costing and the evaluation of precure degree is equivalent for both situations. The prepregs were cured in a laboratory-made autoclave at the pressure of 0.7 MPa. The cure temperature was kept 185°C for 3 h. All the heating and cooling ramps were carried out at the same rate of 2.5°C/min.

Figure 1:

Precure and cure cycles used to process carbon/ benzoxazine prepregs.

2.2 Assessment of void content

The specimens were cut from the center of each laminate, and their densities were obtained in order to estimate the void content. The composite densities were measured by Archimedes theoretical verse actual density testing technique according to ASTM D792-08. The void contents of the composites were determined referring to ASTM D2734-09.

2.3 Dynamical mechanical analysis

A dynamical mechanical analyzer (DMA) was used to evaluate the influence of precure process on the dynamical mechanical properties of the final parts. The equipment used in this study was DMA IV, TA Instruments, USA. The size of samples was 1.6 mm in thickness, 7 mm in width, and 45 mm in length. The heating rate used was 5°C/min, and the frequency was 1 Hz under amplitude control. Three-point bending mode was used within a temperature range from 25°C to 250°C.

2.4 Tensile test

Tensile tests were utilized for mechanical evaluation according to ASTM D3039. Tensile strength and tensile modulus were measured using an electrical universal testing machine (Changchun Kexin-WDW-100, Changchun, China) at a crosshead speed of 2 mm/min. The relation between void content and mechanical properties was then determined by correlating the experimental data.

2.5 Scanning electron microscope

Morphology of the carbon/benzoxazine composites was obtained on a Cambridge 3400 SEM (Cambridge, UK). The fracture surfaces of the composite samples from tensile testing were sprayed with gold and mounted onto an SEM holder with double-sided electrically conducting carbon adhesive tapes before the observation.

3.1 Precure degrees and void contents

Two aspects contributed to the precure of the Loctite BZ 9704 prepregs in this study. For one thing, the resin in the prepregs already had a precure degree of 95% before gelation in order to lower the viscosity for simple bagging procedures. For another, different heating procedures in the oven were intentionally designed to increase the precure degree of the prepregs. The precure degree of the prepregs can be measured by residual enthalpy of cross-linking reaction, which is calculated by integrating the heat flow over the entire exothermic peak on the DSC curve. Despite the inaccessibility of the total enthalpy of the cross-linking reaction of unprecured prepregs, the precure degree can be acquired through the residual enthalpy of the reaction of prepregs without precure and the residual enthalpy of the reaction of prepregs precured in the oven, as shown in Figure 2A. Thus the precure degree α can be calculated in the equation below:

Figure 2:

DSC curves for benzoxazine resin (A) showing the total enthalpy of prepregs without precure from a slow temperature scan, (B) showing the residual enthalpy of prepregs for different precure time at 80°C.

where ΔH_S (J/g) is the residual enthalpy of cross-linking reaction of prepregs that are precured in the oven, and ΔH (J/g) is the residual enthalpy of cross-linking reaction of prepregs without precure (with an accurate precure degree of 95%).

Figure 2B shows the thermal behavior of benzoxazine resin extracted from the prepregs precured at 80°C for various times by DSC measurement. The enthalpy of cross-linking reaction of the prepregs decreased with the increase of the precure time. A lower value of ΔH_S indicates a higher precure degree of the prepregs according to equation (1). Thus a positive correlation between the precure time and the precure degree of the carbon/benzoxazine prepregs at 80°C was found, and the fitting curve was plotted in Figure 3A. The decreasing curvilinear trend has been found consistent with other studies [19–21].

Figure 3:

Precure degree of carbon/benzoxazine prepregs plotted with (A) precure time at 80°C, (B) measured thickness of composite laminates.

The cure reactions of thermosets could be described by a series of kinetics models [20, 21]. Phenomenological reaction model is one of the most commonly used kinetics models characterizing the cure behavior semi-empirically [22]. The rate of cure reaction is defined as:

where dα/dt is the rate of reaction, K(T) is the temperature-dependent rate constant, and f(α) is a polynomial function of cure degree α corresponding to a different reaction model. More DSC data are required to acquire the kinetic parameters of the benzoxazine resin, but this is not the purpose of this study. K(T) is a constant when the temperature is fixed at 80°C. With the increase of precure degree, the concentration of function groups declined. This variation led to the increase of the viscosity of benzoxazine resin and the resistance to relative motion of chain segments. As a result, the reaction rate dropped and the cure degree increased with a decreasing growth rate.

Precure degree of the resin also has an effect on the laminate thickness. The laminate thickness measured with a vernier caliper was plotted in Figure 3B. It was positively related to the precure degree in both series of the laminates. It varied from 1.61 mm to 1.67 mm in the laminates with nominal thickness of 1.6 mm and varied from 5.28 mm to 5.40 mm in the laminates with nominal thickness of 5.3 mm. It has been explained in the literature [9, 14, 23] that the increase of laminate thickness is due to the increasing volume of voids. Therefore, the influence of precure degree on the void content in the carbon/benzoxazine laminates should also be investigated.

The void content of each type of laminate was plotted against precure degree in Figure 4. It varied from 0% to 3.21% in the laminates with nominal thickness of 1.6 mm and varied from 0% to 3.66% in the laminates with nominal thickness of 5.3 mm. The void content of laminates with the thickness of 1.6 mm was found to be lower than that of 5.3 mm at the same precure degree level. Furthermore, the void content of both laminates has a positive correlation with the precure degree, and the growth rate of the fitting curve was gradually reduced (see Figure 4).

Figure 4:

Relationship between void content and precure degree (or precure time at 80°C) of carbon/benzoxazine composite laminates.

It has been reported that gas (volatiles) evolution during curing and air bubble trapping in manufacturing process are the two main reasons of void formation [14, 23, 24]. As the carbon/benzoxazine prepregs have at least 95% precure degree and there are few volatile products in the cure reaction, the trapped air bubbles play a dominant role in void formation. The precure of the prepregs in the oven makes the resin cross-linked partially within the layer before autoclave molding, which suggests a decline of the proportion of the uncured resin that is supposed to link the plies by chemical bonds in autoclave process. Consequently, it becomes easier to embed air bubbles as voids between the layers in manufacturing process. The total volume of voids gets larger in the laminate with a larger ply amount, which means the void content of a thick laminate is larger than that of a thin one under the same precure degree. Besides, the precure pressure of the prepregs in the oven is far lower than that in autoclave. In the precure procedure, there is not enough driving force removing all trapped voids from high viscosity resin system as well as ability to limit the trapped void growth [1, 2, 24, 25]. A larger precure degree of the resin system indicates a longer time precuring under 0.1 MPa in the oven, so more voids form and grow in the laminates.

3.2 Dynamical mechanical properties

The results of the DMA tests of the laminates with different precure times are shown in Figure 5. The storage modulus E′ of the laminates reflects the storage energy by elastic deformation. Loss modulus E″ of materials is a measure of energy dissipated as heat/cycle under deformation [25]. Damping factor Tanδ is the ratio of the storage modulus to the loss modulus. A slight difference of each parameters obtained from the laminates with different precure time could be seen before the glass transition. This indicates that the precure degree has a small effect on the dynamical mechanical properties. The movability of all the chain segments was restricted, and the polymer kept a glass state when the degree was lower than 150°C. When the temperature was higher than approximately 150°C, the chain segments started releasing from the glass state and translate to the elastomeric state. The variation of precure degree in matrix (from 95% to 96.3%) was large enough to change the average length of polymer chain segment, so the dynamical mechanical temperature spectra show the slight difference. However, the values are very close to each other, so it can be confirmed that there is no significant difference in chemical compositions and structures of the final composites from the DMA results.

Figure 5:

DMA results of the carbon/benzoxazine laminate with different precure time of the prepregs.

The glass transition temperature (T_g) of the laminates acquired from E″ curve gives a further explanation. It represents the maximum service temperature of the carbon/benzoxazine composite materials. Meanwhile, it varies with the material compositions and results in a different physical property of the material. Figure 6 shows that the T_g of the laminates increased from 159.4°C to 163.5°C with the decrease of precure degree. The measured T_g values are also quite close to each other. Thus the precure degree is considered to have such little impact on the glass transition temperature of the laminates, which can be ignored. It confirms the inference once again that the chemical constituents in the laminates with different precure cycles are almost the same. To sum up, the precure degree has an ignoring effect on the dynamical mechanical properties of the final parts.

Figure 6:

T_g of the carbon/benzoxazine laminate vs. precure degree (or void content) of the prepregs.

3.3 Tensile properties

The tensile test was performed on all the composite specimens with the thickness of 1.6 mm according to ASTM D3039 to evaluate the influence of precure degree on static mechanical properties. Figure 7 presents the results of tensile strength σ_max and modulus E of the laminates. The values of the two properties were negatively related to the precure degree as well as void content, which is consistent with literatures [1, 2]. The tensile strength decreased by 9.88% when the void content increased from 0% to 3.2%. Besides, there appeared to be 10.02% decrease when the void content increased by 3.2% with respect to tensile modulus. However, the Poisson’s ratio ν is less sensitive to the void content, as it was nearly the same under different precure time (see Table 1).

Figure 7:

Relationship of carbon/benzoxazine composite laminates between precure degree (or void content) and (A) tensile strength, (B) tensile modulus.

It is very likely that the voids introduced by precure process are mainly accounted for the decrease of tensile strength and modulus of the laminates [5, 14]. The micro-sized voids were hardly found in the laminate without precuring in the oven while easily found distributed all over the fracture surface when the precure time was prolonged to 30 min (see Figure 8). The increasing presence of micro-sized voids can result in initiating a micro-sized crack in the matrix when applying load. The local failure density and crack size grow with the load increasing; subsequently, fiber breakage occurs and clusters with micro-sized cracks in the adjacent matrix as a macro-sized crack. Stress concentration around the crack tips leads to the continuous crack propagation. However, the laminates without precuring in the oven had less damage under the same load. Thus the tensile strength and modulus decrease with the increase of the void content. The high fiber volume fractions after curing in the autoclave contributes to the insensitivity of the Poisson’s ratio to the void content [2]. It can be explained using the rule of mixtures. The presence of voids can lead to the decrease of matrix strength and modulus but not that of the Poisson’s ratios of fibers and matrix. Hence, the Poisson’s ratio of the carbon/benzoxazine composites remained unchanged with the increase of void content.

Figure 8:

Microstructure of the 0° layer in carbon/benzoxazine composite specimen precured for different times at 80°C: (A) 0 min and (B) 30 min.

The microstructures of the specimens after fracture were observed by SEM. Most of the specimens failed in the middle of the gauge length. In all these cases, fiber breakage, fiber/matrix interface debonding and matrix hackles were the dominant failure modes in the 0° plies, 90° plies and ±45° plies, respectively. The failure modes of the specimens under different precure degree were nearly the same. The laminate with the precure time of 30 min at 80°C was chosen as a typical example to specifically analyze the failure mechanisms from the fracture microstructures.

Figure 9A shows the general morphology of the fracture surface of the specimen. Delaminations among all the layers exhibited a completely uncoupled fracture. The failure feature of 90° ply in Figure 9B shows that fiber debonding was interrupted by deep matrix cracks every several diameters long in the marked zone, and these matrix hackles inclined towards the direction of crack propagation. Besides, a few fiber breaks were observed in the 90° ply as a result of fiber bridging. Actually, fiber/matrix interface debonding with a slick debonded surface also existed in the failure process [26]. The interfacial strength decides whether matrix hackle or interface debonding was the dominant failure mode. A relatively high strength of carbon/benzoxazine interface enables the cracks to propagate in the matrix rather than along the fiber/matrix interface, which forms matrix hackles. Oppositely, a low strength leads to interface debonding. The matrix hackles can still be observed in the failure of ±45° plies (see Figure 9C). This suggests the shear force inclined towards the direction of crack propagation [5, 26]. The shear mode of the matrix reveals the stress redistribution after fiber breaking. In addition, the fiber/matrix interface debonding still played an important role in the failure of ±45° plies. The progressive breakage of the fibers was the major cause of failure in the 0° plies, according to Figure 9D, owing to the Weibull distribution of fiber strength [27, 28]. The mechanism of fiber pull-out was revealed distinctly: some fractured fibers were exposed without the package of matrix, and some matrix was left with a deep fiber-shaped hole.

Figure 9:

Microstructure of different layers in carbon/benzoxazine composite specimen precured for 30 min at 80°C after fracture by SEM: (A) [90°/45°/0°/-45°]₃, (B) 0°, (C) 45° and (D) 90°.

It should be noticed that there was a slight difference between the microstructures of 90° plies of the specimens with different precure time, as shown in Figure 8. The interface debonding was the dominant failure mode of the specimen precured for 0 min, while matrix hackles were the dominant failure mode of the one precured for 30 min. As analyzed above, precure procedure can improve the carbon/benzoxazine interfacial strength. The precure procedure allows some chain segments of the matrix to relax to a viscous state, causing less high residual thermal stress concentration to the interfaces of the laminates. Therefore, the premature failure is more likely to happen in matrix instead of interface, which indicates an improvement on the interfacial strength. Moreover, plenty of micro-sized voids generated in the resin system precured for 30 min. This is an important reason for strength reduction of the laminates, which has also been explained in detail before.