Interply friction of carbon fiber/epoxy prepreg stacks under different processing conditions

Abstract

The interply friction of prepreg is commonly found from the pre-forming to the final curing process during the fabrication of composite materials. In this study, a testing device was established to evaluate the slipping behavior of two kinds of carbon fiber/epoxy prepreg systems and the dominated friction mechanisms under different processing parameters, including temperature, pressure, and pulling rate, are determined. The interply frictional resistances of prepreg stacks were studied with the surface morphology observation and the surface roughness measurements. The results indicate that the friction mechanism of the two prepreg systems is the mixed friction, which could change from the Coulomb-dominated friction to the hydrodynamic-dominated friction under different processing conditions. The surface morphology and surface roughness of prepreg have significant effects on the slipping behavior controlled by the mixed friction. The Coulomb-dominated friction is not conducive to the slippage of prepreg and may result in fiber wrinkles in the laminates. The hydrodynamic-dominated friction should be achieved by adjusting the processing conditions to avoid manufacturing defects.

Keywords

Carbon fiber reinforced-polymer composite prepreg interply friction forming surface analysis

Introduction

As composite structures used in aerospace industry are increasingly large and complex, prepreg lay-up process as the first step of the composite manufacturing need to be with higher efficiency and lower labor. The traditional manual lay-up process is very time-consuming and the forming quality is not repeatable, especially for large components. To enhance product quality and increase productivity, automated tape laying (ATL) technique is developed, which is suitable for producing parts with flat or large curvature radius surfaces.^1,2 However, ATL is unable to directly lay-up complex composite parts with angle-shaped structure, such as L-shaped stringers and C-beams. For this kind of components, a flat prepreg laminate can be first formed by ATL and then deformed into the designed shape with abrupt contours using the hot diaphragm forming process. This technique realizes laying up complex structures by automated technology.^3,4

A schematic of the deformation for the prepreg laminate fabricated by the hot diaphragm forming process is shown in Figure 1. When the pressure P is applied, interfacial shear stresses τ₁ and τ₂ arise between the prepreg layers. The individual plies of prepreg stacks slip against each other to form parts with the designed angle-shape. Previous study⁵ has shown that interply and intraply deformation contributes to the forming of prepreg laminate and our previous work on the hot double-diaphragm forming process for thermosetting prepreg laminates^6,7 indicates that there is a close relationship between the slipping ability of prepreg layers and the forming quality of composite parts. This article focuses on the interply friction of carbon fiber prepreg systems by measuring the frictional resistance of prepreg under different processing conditions.

Figure 1.

Schematic of hot diaphragm forming process and sliding condition occurring at the prepreg–prepreg interfaces.

Over the past three decades, the slipping behavior of prepreg during the composite manufacturing has attracted considerable research efforts. Martin et al.⁸ developed a testing device to measure the frictional resistance of woven prepreg and found out that core crush of honeycomb composites was strongly influenced by the frictional resistance. Ersoy et al.⁹ used a similar apparatus to characterize the frictional processes of prepreg–prepreg and prepreg–tool during autoclave manufacturing and indicated that the friction of prepreg existed even at low degrees of cure. Several researchers worked on the influences of processing parameters on the frictional behavior of prepreg systems,^5,10–16 such as the temperature, the pressure, and the sliding rate, and found out that all these processing conditions had different effects on the slippage between prepreg. Others did experimental and numerical efforts on the influences of the frictional behavior of prepreg on the composite manufacturing,^17–25 such as residual stresses and shape distortions. The slipping ability of prepreg also influences the consolidation process of the L-shaped laminate manufactured by the autoclave process.^26,27 For the studies on the diaphragm forming technique,^28–34 Mallon and O'Bradaigh^28–30 observed that the relative movement of adjacent plies occurred during the process and the major defect in the forming was the fiber buckling in the laminate.³⁴

The above studies demonstrate that the slipping behavior of prepreg is commonly found from the pre-forming to the final curing process during the fabrication of composite materials. Moreover, the friction shows different behaviors under different processing conditions. As is well known, friction behaviors are controlled by different friction mechanisms and the mechanisms can be divided into three groups according to the tribology. When there is no fluid between the contact surfaces, the friction should be generally described as the Coulomb friction. When two surfaces in relative motion are completely separated by a fluid lubricating film, the friction between the two surfaces in contact may be purely hydrodynamic friction. When the friction consists both of the situations above, it is referred to the mixed friction.³⁵

Prepreg is a mixture of fibers and resin, which has different processable states under different processing conditions. Therefore, the friction mechanism may change with the parameters and further influence the composite manufacturing. Take the hot diaphragm forming process for example. The interply deformation of prepreg is one of the main mechanisms and a proper slipping ability of prepreg under certain processing condition is needed for the deformation to fabricate composite parts without fiber winkles. However, until now few studies have been conducted on the changes of friction mechanisms under different processing conditions, which is important to control the slipping ability of prepreg and determine the manufacturing parameters.

This study aims to determine the governing mechanism controlling the interply friction of carbon fiber/epoxy prepreg materials and the changes of the mechanism under different processing conditions. A testing system was set up to evaluate the frictional resistances of prepreg plies, and the slipping behaviors of the prepreg systems at different temperatures, pressures, and pulling rates were measured. The surface morphology and the surface roughness of two kinds of prepreg systems were studied for explaining the differences of their slipping behaviors. Based on these results of experiments, the interply friction mechanisms of prepreg systems at different processing parameters were discussed.

Experimental

Materials

The prepreg materials used in the study were unidirectional carbon fiber/epoxy resin 12500 prepreg (Guangwei Co.) and unidirectional carbon fiber/epoxy resin X850 prepreg (Cytec Co.). The curing temperature of the 12500 prepreg is 130℃, while the curing temperature of the X850 prepreg is 180℃. The resin weight contents of the 12500 prepreg and the X850 prepreg are 31.7% and 35.5%, respectively.

Rheological measurements of the two kinds of epoxy resin used in the prepreg stacks were conducted at a heating rate of 10℃/min using a rheometer (Gemini, Bohlin Instruments). The stress was 10 Pa and oscillation frequency was 1 Hz. The dissolution method was used to obtain the resin of prepreg. First, the prepreg sample with the size of 100 × 100 mm² was placed in a beaker containing 150 mL of acetone. The beaker was covered by a glass-surface vessel and was shaken gently during the soaking. The fibers were then removed from the beaker and the beaker was placed at 25℃ until the acetone evaporated. Finally, the beaker was put into the vacuum oven at 30℃ for 12 h and the resin of prepreg can be obtained. However, there are insoluble thermoplastic ingredients in the resin of the X850 prepreg, which makes it difficult to obtain the pure resin of this prepreg system. The measured resin viscosity of the X850 prepreg cannot represent the actual viscosity of the resin and can only demonstrate the changing trend of resin viscosity with the temperature.

Friction measurements

To characterize the slipping ability between individual prepreg plies, a testing device was designed to be mounted in a standard tensile testing machine (SANS Ltd. Co., 5 kN) and measure the friction behavior of the prepreg. The testing principle of this device is similar to the one of the apparatus used by Martin et al.⁸ and Ersoy et al.⁹ to measure the frictional resistance of thermosetting prepreg.

The testing device used in this study includes a heating system and a pressurizing system to simulate the forming conditions of the composite manufacturing process, as shown in Figure 2. The pressure was applied to the surfaces of the sample by means of four spring-screw sets, which were calibrated according to the load–displacement curves of the compression springs. Before the friction test, the actual force applied to the testing surface by these four spring-screw sets was measured using the tensile testing machine. However, the normal load may change as the plies were pulled out during the test, which was not considered in this study. The sample was heated by a couple of silicone rubber heating films, and a programmable profiling temperature controller was used to control the temperature cycle. Aluminum plates and rubber pads were used between the heating films and the specimen to insure that the temperature and pressure were uniformly distributed over the testing areas. A locking fixture was applied to grip the steel backplates and fix the whole apparatus when the specimen was pulled during the test.

Figure 2.

Frictional resistance testing device: (a) test configuration and (b) photograph.

The specimen used for measuring interply frictional resistance consists of three parts, as shown in Figure 3(a). The upper plies were two pieces of 150 × 35 mm² prepreg strips where the fibers were aligned along the direction of y. The plies in the middle were two pieces of 50 × 35 mm² prepreg strips where the fibers were aligned along the direction of z. The lower plies were prepared by folding two pieces of 350 × 35 mm² prepreg strips where the fibers were aligned along the direction of y and then looped around the bottom spar. The spar was fixed on the testing device to prevent the movement of the lower plies, when the upper plies were pulled by the tensile testing machine. The role of middle plies was to fill the empty area between the upper and lower plies and insure that the pressure was uniformly distributed over the surfaces of the sample.

Figure 3.

Schematic of the specimen used in the testing device.

During the test, the specimen was mounted on the testing device and was heated to the testing temperature. When the temperature was increased, the compaction of prepreg resulted in a thickness reduction of prepreg and the actual force exerted on the testing surface decreased. To maintain the force, the springs were adjusted according to the changes of prepreg thickness. When the temperature and the thickness of the sample were stable, the test began. The equilibrium time was no longer than 1 min and the change of the curing degree for the prepreg can be ignored according to the rheological measurements. The upper plies were pulled by the tensile testing machine at a certain pulling rate, while the middle and lower plies were fixed by the device. A slippage occurred at the [0/0] interfaces a–b and c–d between the upper and lower plies during the test (Figure 3), so the testing region was the overlapping area with 100 × 35 mm². The tensile load and the displacement were recorded during the testing process. Figure 3(b) demonstrates the force condition at the testing region, where F is the tensile load, N the normal force exerted on the surfaces, and f the friction force between the upper and the lower plies. Based on the force equilibrium equation, the relationship between the tensile load and the friction can be expressed as follows

F = 2 f

(1)

The tensile force F equals the total frictional resistance of prepreg, so the maximum tensile load was chosen to represent the slipping ability of prepreg layers.

According to the tribology, when there is no fluid between the surfaces, the friction should be generally described as the Coulomb friction. It is governed by the force normal to the interface and the friction coefficient is generally described as follows

μ = f / N = F / 2 N

(2)

When the two surfaces in relative motion are completely separated by a fluid lubricating film, the friction between the two surfaces in contact is believed to be purely hydrodynamic friction. In this case, the friction may be predicted in terms of the traction forces acting on the film.³⁶ Neglecting the surface roughness of the surfaces, it can be written as a function of matrix viscosity, film thickness and pulling rate

f = \frac{η}{d} ν

(3)

where η is the matrix viscosity, d the film thickness, and v the relative speed of slipping. However, for actual surfaces, the friction may consist of both the situations. It is between the two friction models and can be referred to the mixed friction.

In this article, the frictional resistances of prepreg systems at different processing conditions were measured. The temperature, pressure, and pulling rate were chosen according to the actual processing parameters during the composite manufacturing. During the test, the force applied on the prepreg surface was constant. As the prepreg was pulled, the testing area changed and the effective pressure was most likely changing accordingly. To simplify the comparisons of the friction behavior of the prepreg under different processing conditions, the initial pressure was used in the discussions and the effective pressure increased about 10% during the whole test. Table 1 gives the test configurations in this study.

Table 1.

Test configurations for interply friction of prepreg stacks.

Material	Temperature (℃)	Pressure (MPa)	Pulling rate (mm/min)
12500 Prepreg	18	0.1	1
	30	0.1	1
	45	0.1	1
	60	0.1	1
	80	0.1	1
X850 Prepreg	18	0.1	1
	30	0.1/0.6	1/0.2/0.1
	45	0.1/0.37/0.45/0.6	1
	60	0.1	1
	80	0.1	1

Surface characterization of prepreg

Surface roughness of the prepreg systems was studied and the samples were 100 × 100 mm² in size. Each of the samples was put in a vacuum bag and compacted under the vacuum pressure for 10 min, then they were placed in an oven at 18℃, 30℃, 45℃, 60℃ and 80℃ for 10 min, respectively. The samples were removed from the oven and cooled quickly to ambient temperature for a best effort at maintaining the profile of prepreg layers. The pressure applied to the samples was 0.1 MPa from the beginning to the cooling stage. Therefore, the samples processed as above could characterize the surface roughness of prepreg at the temperature of 18℃, 30℃, 45℃, 60℃, and 80℃, when the pressure was 0.1 MPa.

The surface profile of the uncured prepreg was measured using a surface roughness detector (TR110®, Peak Times of Beijing Co.) and the four testing positions for one sample are shown in Figure 4. The size of the surface roughness detector is 24 × 70 mm². The scanning route is perpendicular to the fiber direction (the direction of x). The working principle is as follows. When the sensor is moved by the drive mechanism and does a uniform linear motion along the testing surface, the internal probe perpendicular to the working surface develops a vertical displacement with the outline of the uneven surface. Through the sensor, the displacements are transferred into power changes. The electric signals are then amplified, filtered, and converted to digital signals.

Figure 4.

Testing positions of surface roughness of prepreg.

The measurement of surface roughness of a prepreg is shown in Figure 5. The value of R_a, which is the arithmetic mean height of each point on the contour within the sampling length, is chosen to represent the surface roughness of prepreg layers, as given by equation (4).

R_{a} = \frac{1}{l} \int_{0}^{l} | y (x) | d x = \frac{1}{n} \sum_{i = 1}^{n} | y_{i} |

(4)

where y(x) is the height of each point on the profile, l the sampling length, n the measurement points, and y_i the height of each measurement point on the profile. The larger the value of R_a is, the rougher the surface is. The sampling length l, the evaluation length l_n, and the datum line are three basic parameters for testing the surface roughness. The datum line is the arithmetic average midline of the profile. The sampling length l is a section of the datum line which can represent the characteristics of surface roughness. The evaluation length l_n is a required length of the measurement of surface profile. For uneven surfaces, l_n ≥ 5l. During the test, the sampling and evaluation lengths were chosen to be 0.8 and 4 mm according to the Chinese standard,³⁷ respectively.

Figure 5.

Measurement of surface roughness of prepreg.

Surface morphology of the uncured prepreg systems at 18 ℃ was observed using the optical microscope (Olympus BX51M). The samples were 100 × 100 mm² in size.

Results and discussion

Data repeatability for friction measurements

To demonstrate the effectiveness of the friction measurements, the frictional resistances of the 12500 prepreg at 25℃ were measured by the designed testing device at a pressure of 0.1 MPa. The pulling rate was 1 mm/min. Figure 6 shows the reproducibility of data for the frictional resistance testing. The results are quite consistent and the coefficient of variance for the maximum frictional resistances is 0.5%.

Figure 6.

Reproducibility on frictional resistance testing of prepreg.

Influence of temperature

The processing properties of a prepreg system, especially the resin viscosity of prepreg, are dependent on the forming temperature, and the temperature might have significant impact on the slipping behavior of prepreg layers. Figure 7 shows the frictional resistances of the 12500 prepreg and the X850 prepreg at different temperatures, respectively. The pressure was 0.1 MPa and the pulling rate was 1 mm/min. The maximum frictional resistances are listed in Table 2.

Figure 7.

Frictional resistances of two prepreg systems at different temperatures: (a) 12500 prepreg and (b) X850 prepreg (pressure 0.1 MPa, pulling rate 1 mm/min, testing area 3500 mm²).

Table 2.

Frictional resistances of two prepreg systems at different temperatures.

Temperature (℃)	12500 prepreg		X850 prepreg
Temperature (℃)	Maximum frictional resistance (N)	Resin viscosity (Pa.s)	Maximum frictional resistance (N)	Resin viscosity (Pa.s)
18	2479.0	–	–	–
30	702.5	1735.0	1207.0	1532.3
45	76.7	178.0	155.3	74.5
60	57.1	24.0	72.2	8.4
80	57.0	4.0	58.7	1.2

Comparing the curves of load–displacement under different temperatures for these two prepreg systems, the curves can be divided into two types. At lower temperatures, i.e. 18℃ for the 12500 prepreg or 30℃ for the X850 prepreg, the load increases almost linearly with the displacement at the initial stage. After the force reaches a maximum value, it gradually decreases to a steady level. These curves have ‘stick-slip’ peaks, which are similar to the characteristics of dry Coulomb friction. However, the state of ‘stick’ is a viscous behavior of one prepreg layer moving upon the other, rather than the quasi-static state. After the maximum force the prepreg plies ‘slip’, which means that the viscous action is weakened and the force condition gradually reaches the equilibrium. For higher temperatures, e.g. 30℃ for the12500 prepreg, the trends of load–displacement curves change. First, the load also increases almost linearly with the displacement, and then it develops at a gradually reducing rate. After the force goes to a maximum level, the curve levels off. There is not any ‘stick-slip’ peak in these curves. For the curve tested under 18℃ for the X850 prepreg, the load increases almost linearly with the displacement. However, the maximum force exceeded the range of the measurement for the tensile testing machine.

Moreover, as shown in Figure 7 and Table 2, it can be seen that for the two kinds of prepreg, the maximum frictional resistances significantly decrease when the temperature increases from 18℃ to 80℃. This decrease is because the viscosity of the resin decreases with the increasing temperature and the resin acts as a lubricant between the two layers of prepreg. This lubrication allows the prepreg plies to slip against each other easily and results in a low frictional resistance. As the frictional resistances are viscosity dependent and the load–displacement curves under low temperatures have similar characteristics of Coulomb friction, it can be concluded that the friction mechanisms for these two prepreg systems are the mixed friction. Under different processing conditions, the actual friction presents different slipping behaviors, a Coulomb- or a hydrodynamic-dominated friction. For the 12500 prepreg at 18℃, the friction between individual plies shows the Coulomb-dominated friction behavior, while for higher temperatures, the hydrodynamic friction dominates. For the X850 prepreg, when the temperature is 18℃ and 30℃, the frictions between the prepreg layers show the Coulomb-dominated friction. On the other hand, when the temperature goes up, the friction between the two prepreg surfaces in contact is hydrodynamic-dominated friction.

From the comparisons of the frictional resistances of the 12500 prepreg and the X850 prepreg listed in Table 2, it can be observed that the maximum frictional resistance of the 12500 prepreg is lower than that of the X850 prepreg at the same temperature from 30℃ to 80℃ and the friction mechanisms of the two prepreg systems are different at 30℃. To analyze the differences between the frictional resistances of these two prepreg systems and determine the governing mechanism controlling the interply friction, the surface topography and the surface roughness of the 12500 prepreg and the X850 prepreg were investigated by optical microscopy and surface roughness detector.

The micrographs of the uncured prepreg surfaces for the two kinds of prepreg at 18℃ are shown in Figure 8. From the photos, it can be observed that for the 12500 prepreg and the X850 prepreg, the prepreg surfaces are both covered with discrete resin. There are mico-valley regions (annotated by the red ellipse in Figure 8a and the red circle in Figure 8b) on the prepreg surfaces and the fibers can be distinguished from the resin in these areas. Moreover, the surface morphologies of the two kinds of prepreg are obviously different. For the 12500 prepreg, the valley regions are thin strips along the fiber direction, while the valley areas of the X850 prepreg are irregular polygons or circles.

Figure 8.

Micrographs of prepreg surface: (a) 12500 prepreg and (b) X850 prepreg.

The surface roughness of the two prepreg systems at different temperatures is shown in Figure 9. From the picture, it can be concluded that for each temperature, the average surface roughness of the X850 prepreg is larger than that of the 12500 prepreg. This difference in surface roughness may be the reason that the maximum frictional resistances of the X850 prepreg are higher than those of the 12500 prepreg at the same temperatures. In addition, for the X850 prepreg, the surface roughness significantly reduces with the increasing temperature, while for the 12500 prepreg, the influence of temperature on the surface roughness can be neglected. As is well known, the surface roughness has a strong dependency on the prepreg surface geometry.^38,39 When the viscosity of resin decreases with the increasing temperature, the resin on the surface of prepreg may flow into the micro-valley regions under the pressure. Therefore, the movement of resin changes its distribution on the prepreg surface and decreases the surface roughness of prepreg. From Figure 8, it can be observed that the surface roughness of the 12500 prepreg plies is more even than the one of X850 prepreg. Compared to the 12500 prepreg, the area of micro-valley regions of X850 prepreg is much larger and the flow of resin has greater influence on the distribution of resin on the prepreg surface. So the surface roughness of X850 prepreg is more sensitive to the temperature.

Figure 9.

Surface roughness of two kinds of prepreg at different temperatures.

The reason why the friction mechanisms of the two prepreg systems are different at 30℃ (Table 2) are as follows. As mentioned above, the actual friction will present different slipping behaviors under different processing conditions. For the mixed friction, when the asperity contact areas are larger than the lubricating film contact areas, the Coulomb friction dominates. For the inverse, the hydrodynamic friction dominates. At the temperature of 30℃, for the 12500 prepreg, because of small surface roughness, the resin on the prepreg surfaces easily results in separating the two prepreg layers in contact. This separation determines that the dominated friction mechanism is hydrodynamic friction. For the X850 prepreg, the surface roughness is larger and the interface of two prepreg plies may have more asperity contact areas. Therefore, the friction mechanism is the Coulomb-dominated friction. In a word, the surface roughness has a great effect on the slipping behavior controlling by the mixed friction.

The studies of the frictional resistances of prepreg under different temperatures indicate that increasing the temperature can make the friction mechanism of prepreg systems change from the Coulomb-dominated friction to the hydrodynamic-dominated friction and a higher temperature benefits the slipping behavior of prepreg. In our previous research on the hot diaphragm forming process,^6,7 as shown in Figure 10, the 12500 prepreg laminate with a stacking sequence of [45/-45]_4s cross-ply formed at 18 ℃ shows severe wrinkles on the inner surface of the corner region. The sliding behavior of prepreg is the Coulomb-dominated friction at 18℃ and the slipping ability of prepreg is low. There is not enough slippage during the deformation, and the wrinkles occur on the surface of the laminate. For higher forming temperatures, there is an obvious reduction in fiber wrinkles at 30℃ and the surfaces of the parts become smooth without wrinkling at 45℃, 60℃, and 80℃. At these temperatures, the friction is dominated by the hydrodynamic friction and the ability of slippage between prepreg plies improves.

Figure 10.

Photographs of surface quality of 12500 prepreg laminates fabricated at different temperatures using the hot diaphragm forming process.

This result demonstrates that the Coulomb-dominated friction is not conducive to the deformation of prepreg layers during the hot diaphragm forming process and results in wrinkles easily. Therefore, the understanding of the friction mechanism under different processing parameters is of great importance for controlling the slipping ability of prepreg systems and defining the composite manufacturing process to avoid defects.

Influence of pressure

The frictional resistances of the X850 prepreg at different pressures are shown in Figure 11. The temperature was 45℃ and the pulling rate was 1 mm/min. The maximum frictional resistances at 30℃ and 45℃ are listed in Table 3.

Figure 11.

Frictional resistances of X850 prepreg at different pressures (Temperature = 45℃, pulling rate = 1 mm/min, testing area = 3500 mm²).

Table 3.

Comparisons of frictional resistances of X850 prepreg at different pressures.

Pressure (MPa)	0.1	0.37	0.45	0.6
30℃ maximum frictional resistance (N)	1207.0	–	–	3238.0
45℃ maximum frictional resistance (N)	155.3	228.2	234.7	254.6

From the results, it can be observed that only increasing pressure does not change the trends of load–displacement curves of the X850 prepreg system. In addition, at 30℃ and 45℃, the maximum frictional resistances both increase when the pressure increases from 0.1 to 0.6 MPa (Table 3). According to Martin et al.,⁸ the friction coefficient under different parameters can be calculated. As the pressure increases from 0.1 to 0.6 MPa, the friction coefficient at 30℃ decreases from 1.72 to 0.77, while the friction coefficient at 45℃ decreases from 0.22 to 0.06. These results are consistent with some previous works.^5,8,10,11

When the temperature is 30℃, it can be noticed that the increase in the maximum frictional resistance is 170% from 0.1 to 0.6 MPa and the testing curve has the ‘stick-slip’ peak at the pressure of 0.1 MPa (Figure 7). If the friction is purely dry Coulomb friction and the surface roughness of prepreg is constant at different pressures, the friction should be linearly dependent on the normal force, as shown in equation (2). However, the predicted value from equation (2) is higher than the measured one at 30℃, which is attributed to the change of the surface roughness of prepreg resulting from the deformation of resin under pressure. This outcome indicates that the friction between individual plies shows the mixed friction and is Coulomb dominated.

When the temperature is 45℃, the increases of the maximum frictional resistances at 0.37, 0.45, and 0.6 MPa are about 50–60%, compared to the maximum frictional resistance of prepreg at 0.1 MPa, and the testing curves have no ‘stick-slip’ peak. If the friction is purely hydrodynamic friction, the frictional resistance should be independent of the normal force according to equation (3). However, the normal force influences the friction behavior of prepreg at 45℃. This may be because of the flowing resin on the prepreg layers. Therefore, it can be concluded that the friction between the prepreg layers is the mixed friction and hydrodynamic dominated. In addition, when the pressure is 0.1 MPa, the curve flattens out at the end of the test. However, for higher pressures, the force seems to increase slightly with the displacement. This phenomenon indicates that the force condition cannot reach the equilibrium and the X850 prepreg plies are difficult to ‘slip’ completely at higher pressures.

The results above demonstrate that the friction mechanism of prepreg shows the mixed friction under different pressures and is not be changed by only increasing pressure within this study. At different temperature, the pressure has influences on the Coulomb- or the hydrodynamic-dominated friction.

Influence of pulling rate

The pulling rate during the testing represents the slipping velocity of individual prepreg plies and Figure 12 shows the frictional resistances of the X850 prepreg at different pulling rates. The maximum frictional resistances are listed in Table 4. The temperature was 30℃ and the pressure was 0.1 MPa.

Figure 12.

Frictional resistances of X850 prepreg at different pulling rates (Temperature = 30℃, pressure = 0.1 MPa, testing area = 3500 mm²).

Table 4.

Maximum frictional resistances of X850 prepreg at different pulling rates.

Pulling rate (mm/min)	0.1	0.2	1
Maximum frictional resistance (N)	278.6	339.7	1207.0

From these results, it can be noticed that for the X850 prepreg system at 30℃, the load–displacement curve at a pulling rate of 1 mm/min has a similar ‘stick-slip’ peak of the dry Coulomb friction, whereas the peak disappears for lower pulling rates. It obviously shows that a decreasing pulling rate can change the friction mechanism of the X850 prepreg system.

Moreover, the maximum frictional resistance increases when the pulling rate goes up from 0.1 to 1 mm/min. This phenomenon may be attributed to the resin on the surface of prepreg, which requires higher forces to deform at higher pulling rates. In addition, when the pulling rates are 0.1 and 0.2 mm/min, the friction between prepreg layers is hydrodynamic-dominated friction. According to equation (3), when the film thickness and the resin viscosity are constant, the frictional resistance increases with the increasing pulling rate.

In a word, the decreasing pulling rate can change the sliding condition from a Coulomb-dominated friction to a hydrodynamic-dominated one, and a lower pulling rate benefits the slipping behavior of the prepreg.

Conclusions

A testing device was designed to characterize the slipping ability between individual prepreg plies, and the friction mechanisms of carbon fiber/epoxy prepreg systems at different temperatures, pressures, and pulling rates were discussed with the surface morphology observation and the surface roughness measurements. The results show that the friction mechanisms for the two kinds of prepreg systems are the mixed friction. Under different processing conditions, the actual slipping behaviors present different states: a Coulomb- or a hydrodynamic-dominated friction.

The temperatures, pressures, and pulling rates have different impacts on the interply friction mechanisms of the two studied prepreg systems. Increasing the temperature or decreasing the pulling rate can make the slipping behavior of prepreg plies change from the Coulomb-dominated friction to the hydrodynamic-dominated friction. The friction mechanisms cannot be changed only by increasing pressure within this study. Furthermore, the studies of the surface morphology and the surface roughness of prepreg indicate that the surface roughness has a great effect on the slipping behavior controlled by the mixed friction. The Coulomb-dominated friction means higher frictional resistance between prepreg plies and is not conducive to the deformation of prepreg layers during the hot diaphragm forming process, easily resulting in fiber wrinkles on the laminates. The hydrodynamic-dominated friction should be achieved by adjusting the processing conditions to avoid manufacturing defects. These results are significant for controlling the slipping ability of prepreg systems and optimizing the composite manufacturing process to avoid defects.

Footnotes

Funding

This study was supported by funding from the Fund of National Engineering and Research Center for Commercial Aircraft Manufacturing (Project no. SAMC11-JS-07-220) and the National 973 Program of China (Project no. 2010CB631104).

Conflict of interest

None declared.

References

Mills

. Automation of carbon fiber preform manufacture for affordable aerospace applications. Compos Part A 2001; 32: 955–962.

Crossley

Schubel

Warrior

. Experimental determination and control of prepreg tack for automated manufacture. Plast Rubber Compos 2011; 40: 363–368.

. Hot drape forming of composites. Aeronaut Manuf Technol 2009; S2: 113–116.

Chen

. Application of composite forming technique in military freighter A400M. Aeronaut Manuf Technol 2008; 10: 32–35.

Larberg

Akermo

. On the interply friction of different generations of carbon/epoxy prepreg systems. Compos Part A 2011; 42: 1067–1074.

Sun

. Effect of forming temperature on the quality of hot diaphragm formed C-shaped thermosetting composite laminates. J Reinf Plast Compos 2012; 21: 1629–1635.

Yao

. Experimental research on hot diaphragm forming and defects cause of composite materials, ME thesis, Beijing, 2012.

Martin

Seferis

Wilhelm

. Frictional resistance of thermaset prepregs and its influence on honeycomb composite processing. Compos Part A 1996; 27: 943–951.

Ersoy

Potter

Wisnom

. An experimental method to study the frictional processes during composites manufacturing. Compos Part A 2005; 36: 1536–1544.

10.

Gorczyca-Cole

Sherwood

Chen

. A friction model for thermostamping comingled glass-polypropylene woven fabric. Compos Part A 2007; 38: 393–406.

11.

Kalebec

Babaarslan

. Effect of weight and applied force on the friction coefﬁcient of the spunlace nonwoven fabrics. Fibre Polym 2010; 11: 277–284.

12.

Chauhan

Kumar

Singh

. Study on friction and sliding wear behavior of woven S-glass fiber reinforced vinylester composites manufactured with different comonomers. J Mater Sci 2009; 44: 6338–6347.

13.

Fetfatsidis

Sherwood

Chen

. Characterization of the fabric/tool and fabric/fabric friction during the thermostamping process. Int J Mater Form 2009; 2: 165–168.

14.

Fetfatsidis

Gamache

Gorczyca

. Design of an apparatus for measuring tool/fabric and fabric/fabric friction of woven-fabric composites during the thermostamping process. Int J Mater Form 2011, pp. 1–11.

15.

ten Thije

RHW

Akkerman

. Design of an experimental setup to measure tool-ply and ply-ply friction in thermoplastic laminates. Int J Mater Form 2009; 2: 197–200.

16.

Vanclooster

Lomov

Verpoest

. Inter-ply and tool-ply friction measurement of woven reinforced thermoplastic composites at forming conditions. In: Proceedings of the 10th international conference on textile composites, Lille, France, 2010, 2010.

17.

Johnston

Vaziri

Poursartip

. A plane strain model for process-induced deformation of laminated composite structures. J Compos Mater 2001; 35: 1435–1469.

18.

Fernlund

Rahman

Courdji

. Experimental and numerical study of the effect of cure cycle, tool surface, geometry, and lay-up on the dimensional fidelity of autoclave-processed composite parts. Compos Part A 2002; 33: 341–351.

19.

Twigg

Poursatip

Ferlund

. Tool-part interaction in composite processing. Part I: experimental investigation and analytical model. Compos Part A 2004; 35: 121–133.

20.

Twigg

Poursatip

Ferlund

. Tool-part interaction in composite processing. Part II: numerical modeling. Compos Part A 2004; 35: 135–141.

21.

Khoun

Hubert

. Investigation of the dimensional stability of carbon epoxy cylinders manufactured by resin transfer moulding. Compos Part A 2010; 41: 116–124.

22.

Kaushik

Raghavan

. Experimental study of tool-part interaction during autoclave processing of thermoset polymer composite structures. Compos Part A 2010; 41: 1210–1218.

23.

Zeng

Raghavan

. Role of tool-part interaction in process-induced warpage of autoclave-manufactured composite structures. Compos Part A 2010; 41: 1174–1183.

24.

Sunderland

Manson

. A thermoviscoelastic analysis of process-induced internal stresses in thermoplastic matrix composites. Polym Compos 2001; 22: 579–592.

25.

Parnas

Walsh

. Vacuum-assisted resin transfer molding model. Polym Compos 2005; 26: 477–485.

26.

Sun

. Numerical study on effects of interaction between rubber mould and lay-up on consolidation of l-shaped laminates in autoclave process. Polym Polym Compos 2011; 19: 271–278.

27.

Sun

. Role of tool-part interaction in consolidation of L-shaped laminates during autoclave process. Appl Compos Mater 2012; 19: 583–597.

28.

Mallon

O'Bradaigh

. Development of a pilot autoclave for polymeric diaphragm forming of continuous fiber reinforced thermoplastics. Composites 1988; 19: 37–47.

29.

Mallon

O'Bradaigh

Pipes

. Polymeric diaphragm forming of complex-curvature thermoplastic composite parts. Composites 1989; 20: 48–56.

30.

O'Bradaigh

Mallon

. Effect of forming temperature on the properties of polymeric diaphragm formed thermoplastic composites. Compos Sci Technol 1989; 35: 235–255.

31.

Smiley

Pipes

. Analysis of the diaphragm forming of continuous fiber reinforced thermoplastics. J Thermoplast Compos Mater 1988; 1: 298–321.

32.

Labeas

Watiti

Katsiropoulos

ChV

. Thermomechanical simulation of infrared heating diaphragm forming process for thermoplastic parts. J Thermoplast Compos Mater 2008; 21: 353–370.

33.

Bersee

HEN

Beukers

. Diaphragm forming of continuous fibre reinforced thermoplastics: influence of temperature, pressure and forming velocity on the forming of Upilex-Rw diaphragms. Compos Part A 2002; 33: 949–958.

34.

Keane

Mulhern

Mallon

. Investigation of the effects of varying the processing parameters in diaphragm forming of advanced thermoplastic composite laminates. Compos Manuf 1995; 6: 145–152.

35.

Wang

. Introduction to tribology, Beijing: Water Power Press, 1990, pp. 18–26.

36.

Stachowiak

Batchelor

. Engineering tribology, Boston: Butterwoth-Heinemann, 2000.

37.

GB1031-83. Parameters and values in surface roughness testing, China, 1983, 1983.

38.

Yang

Pitchumani

. A fractal cantor set based description of interlaminar contact evolution during thermoplastic composites processing. J Mater Sci 2001; 36: 4661–4671.

39.

Yang

Pitchumani

. Interlaminar contact development during thermoplastic fusion bonding. Polym Eng Sci 2002; 42: 424–439.