Wear Behavior of PAN- and Pitch-Based Carbon Fiber Reinforced Aluminum Alloy Composites under Dry Sliding Condition
2017 Volume 58 Issue 6 Pages 898-905
Details
2017 Volume 58 Issue 6 Pages 898-905
The effects of carbon fiber-reinforcement on the wear behavior of aluminum alloy under a dry sliding conditions have been investigated. Two kinds of carbon fibers, PAN-based and pitch-based short carbon fibers, were used as the reinforcements. The composites were fabricated by squeeze casting, and wear testing was carried out using the pin-on-disk method. The wear loss of the alloy and counterpart decreased due to the fiber-reinforcement. The change in the coefficient of friction during the wear test and the scatter in the roughness of the worn surface also decreased by the reinforcement. Examination of the worn surfaces and temperature change of the specimens during the wear revealed that these results were mainly attributed to the crumbled fibers forming a solid lubricant film on the worn surfaces thus preventing seizure of the matrix with the counterpart. Under a high load and sliding speed, the wear loss of the pitch-based fiber composite was lower than that of the PAN-based fiber one. The examination described above revealed that the improvement in the wear and seizure resistance was mainly attributed to the higher thermal conductivity of the pitch-based fiber composite.
The aluminum alloy is attractive as a lightweight and oxidation-resistant material. The drawbacks of the alloy are its low strength and rigidity at elevated temperature and poor wear resistance. Although some aluminum alloys, such as those that contain nickel and a high silicon percent can be made, it is still difficult to satisfy these properties. Therefore, the application of the alloy as a heat and wear resistant material is still limited. The reinforcement of aluminum alloys with non-metallic fibers has been proposed in order to improve these properties. The carbon fiber is one of the most widely used reinforcements because it not only has a high strength, high rigidity and low thermal expansion which can also be seen in many ceramic fibers, but also has a low density and self-lubricant properties. Carbon fibers are roughly divided into two types; the PAN-based one and the pitch-based one. The thermal conductivity of the pitch-based carbon fiber is generally greater than that of the PAN-based carbon fiber. Some researchers1–4) have reported that the PAN-based carbon fiber easily reacted with the aluminum alloy melt to produce a mechanically and chemically unstable Al4C3 and this reaction led to the degradation of the strength of the composite, while the pitch-based carbon fiber was more stable than the PAN-based carbon fiber in the aluminum alloy melt. These findings suggest that the PAN-based and pitch-based carbon fibers dispersed in the aluminum alloy would have a different effect on the wear behavior of the alloy. There are some reports about the wear behavior of the carbon fiber-reinforced aluminum alloy composites. Daoud5) and Ramesh et al.6) reported the sliding wear behavior of aluminum alloy composites reinforced with Ni-coated continuous carbon fibers. Liu et al.7,8) investigated the wear behavior of the aluminum alloy composites reinforced with the PAN-based short carbon fiber. Investigations of the wear behavior of hybrid aluminum alloy composites reinforced with the alumina fiber and graphite fiber (or flake)9) and those reinforced with SiC particles and PAN-based carbon fiber10) have also been reported. Based on these investigations, no research has been found regarding the wear behavior of the aluminum alloy composites reinforced with the pitch-based carbon fibers. Therefore, the effects of the reinforcement with the carbon fibers having different properties on the wear behavior of the aluminum alloy remain unclear.
In the present study, the PAN-based and pitch-based carbon fiber reinforced aluminum alloy composites were fabricated by squeeze casting, and the wear properties of the two composites under dry sliding conditions were compared, then the effects of the fiber-reinforcement on the wear behavior of the aluminum alloy were investigated.
The JIS-AC8A aluminum alloy (Al-12Si-1Mg-1Cu-1Ni) was used as the matrix metal.
Two kinds of short carbon fibers, the PAN-based and the pitch-based milled fibers were used as the reinforcements. Table 1 lists the properties of the fibers11–13). The PAN-based fiber of the type MLD-300 was supplied by Toray Industries, Inc., and the pitch-based fiber of the type XN-100-15M was supplied by the Nippon Graphite Fiber Corporation. As shown in Table 1, the most significant difference between these fibers can be seen in the thermal conductivity and elastic modulus. Although the size (diameter and length) of these two fibers are different because the fibers are from different companies, we have selected the two fibers as close to the same size as possible in order to reduce the size effect on the wear properties of the alloy.
| Density (Mg/m3) |
Length (μm) |
Diameter (μm) |
Thermal Conductivity (W/(m・K)) |
CTE (× 10−6 K−1) |
Elastic Modulus (GPa) |
|
|---|---|---|---|---|---|---|
| PAN | 1.76 | 130 | 7 | 10 | −0.4 | 230 |
| Pitch | 2.20 | 150 | 9.7 | 900 | −1.2～0 | 940 |
The fiber preforms were fabricated as follows. The fibers were dispersed in an aqueous medium containing polyvinyl alcohol (PVA) as the organic binder and SiO2 sol as the inorganic binder. Dewatering was carried out by press-forming, then dried at 373 K for 3 hours to evaporate any residual water and to obtain the green strength due to the PVA. The preform had a 50-mm diameter and was 15-mm thick. The fiber volume fraction in the preform was controlled at 30 vol%.
Squeeze casting was used to fabricate the composite. The preform heated at 673 K was placed in a permanent mold, and the AC8A alloy melt was quickly poured into the mold at 973 K. Due to the heating, the PVA was burned off and the preform strength due to the SiO2 binder was generated. A pressure of 40 MPa for 1 min was then applied by a plunger to accomplish the melt infiltration. Observation of the vertical cross-section of the composites revealed that the preform was perfectly infiltrated with the melt. As a result of the fiber volume fraction measurement in the composites using the Archimedian principle, it was 30 vol%, which is the same as the fiber volume fraction in the preforms.
Figure 1 shows optical micrographs of the composites in the section parallel to the melt-infiltrated surface. The fibers were observed as a dark phase in the micrographs. No agglomeration of the fibers or porosity was observed in the composite, indicating that the melt infiltration into the preform was perfectly accomplished. Due to the press-forming during the preform fabrication process, the fibers were in a planar-dimensionally random arrangement in the composites. No significant difference in the arrangement feature between the two composites was observed.
Optical micrographs in parallel section of composites.
Figure 2 shows the transmission electron microstructure in the vicinity of the fiber-matrix interface of the composites. The reaction products, such as the aluminum carbide (Al4C3), which is generally considered an undesirable phase in the composites primarily due to its brittleness and hygroscopic properties, was not observed near the interface between the aluminum alloy matrix and carbon fibers. As already described, it was reported that the pitch-based carbon fiber was more stable than the PAN-based carbon fiber in the aluminum alloy melt1–4). It has also been reported that the addition of silicon to the alloy suppresses the interfacial reaction3,4). The silicon content of the AC8A alloy used as the matrix metal in the present study was 12 mass%. In addition, the residual SiO2 binder was detected on the fiber surface in the preform. These facts would lead to suppressing the reaction between the aluminum alloy melt and not only the pitch-based fiber, but also the PAN-based fiber.
Transmission electron micrographs in the vicinity of the interface between matrix and fiber in the composites.
Table 2 lists the properties of the unreinforced AC8A alloy and composites. These properties were obtained as described in previous reports14,15).
| Density (Mg/m3) |
Hardness (HV) |
Compressive 0.2% proof stress (MPa)*1 |
Compressive elastic modulus (GPa)*1 |
Thermal conductivity (W/(mK))*2 |
CTE (× 10−6 K−1)*3 |
|
|---|---|---|---|---|---|---|
| AC8A | 2.8 | 118 | 184 | 88 | 125 | 20.0 |
| Composite (PAN) |
2.2 | 61 | 260 | 119 | 52 | 15.8 |
| Composite (pitch) |
2.5 | 52 | 137 | 79 | 184 | 15.3 |
*1 Measured at room temperature (load applied perpendicular to the section shown in Fig. 1)
*2 Measured at room temperature (heat flows parallel to the section shown in Fig. 1)
*3 Measured between RT and 373 K
The composites were subjected to a wear test using the pin-on-disk method, as shown in Fig. 3. The unreinforced AC8A alloy and composites were cut into the pin specimen of 5-mm diameter and 15-mm length that was taken from a direction parallel to the melt-infiltrated direction; the wear surface is parallel to the section observed in Fig. 1. High carbon chromium bearing steel (JIS-SUJ2) (hardness: 188 HV) was formed into a disk shape for the counterpart. Before the wear test, the specimen surfaces were polished with No. 800 abrasive paper and washed with acetone. The wear tests were carried out under contact loads (P) of 17, 27 and 37 N and at the sliding speeds (v) of 0.3, 1.5 and 3.0 m/s. It was carried out under a dry sliding condition and no lubrication was used. The mass loss was measured after cleaning the specimen in acetone using a precision balance with an accuracy of 0.001 mg. The coefficient of friction (μ) during the wear test was also measured using a pin-on-disk wear test machine equipped with a load cell. The temperature of the pin specimen during the wear test was measured by a type K-thermocouple. The tip of the thermocouple with the wire diameter of 0.3 mm was centrally embedded in the radial direction and 5 mm from the worn surface of the specimen.
Schematic illustration of wear test.
The worn surfaces of the pin specimen and the counterpart were observed by optical and scanning electron microscopies. The distribution of each element on the worn surfaces were examined using X-ray maps obtained by the electron probe microanalysis (EPMA). The roughness values (arithmetical mean roughness, Ra) of the worn surfaces of the specimens were measured using a laser microscope. By measuring 3 points for each sample, the average values and scatter due to the measured points were obtained.
Figure 4 shows the relationship between the sliding distance and the amount of wear loss of the pin specimens (unreinforced AC8A alloy and composites) (P = 17 N). By reinforcing the alloy with the carbon fibers, the wear loss of the AC8A alloy was reduced (Fig. 4(a)). When v was 3.0 m/s, the wear loss of the AC8A alloy was severe and the test could not be continued at the distance of more than 400 m (Fig. 4(b)). On the other hand, the composites even withstood longer sliding distances. At the high sliding speed, the decrease in the wear loss by the fiber-reinforcement was more pronounced than that at low speed.
Relationship between the sliding distance and the amount of wear loss of the unreinforced AC8A alloy and composites (P = 17 N).
The relationship between sliding distance and the amount of wear loss of the counterparts is shown in Fig. 5. The wear of the counterpart combined with the unreinforced alloy approximately increased in a linear fashion in all the ranges of wear distance, while that combined with the composite was negligible at every sliding speed.
Relationship between the sliding distance and the amount of wear loss of the counterparts (P = 17 N).
Figure 6 represents the effects of the sliding speed on the wear rate of the unreinforced alloy and composites. The wear rate of the unreinforced alloy drastically increased when the sliding speed exceeded 1.5 m/s under a load of 37 N. The fiber-reinforcement drastically reduced the wear, especially under a high load. Of particular note is that the reinforcement with the pitch-based fiber has a substantial effect on decreasing the wear loss.
Effects of sliding speed on wear rate of the unreinforced alloy and composites.
The coefficient of friction (μ) of the specimens during the wear test is shown in Fig. 7. The error bars in the figure represent the range between the maximum and minimum values, namely, the scatter during the wear. The large range represents a greater change in values of the μ during the wear test. For every specimen, the scatter of μ was greater under a low load, while it was independent of the sliding speed. The average values, which are represented as open or closed symbols in the figure, slightly increased along with the load. The scatter slightly decreased by the fiber-reinforcement under every sliding speed and load, while the reinforcement did not have a significant role in changing the average values of μ. Although the difference in the average values in μ between the two composites was low, the scatter for the pitch-based fiber composite seems to be slightly lower than those for the PAN-based fiber one.
The coefficient of friction (μ) of the specimens during the wear test. Error bars in the figure represent the range between the maximum and minimum values.
Figure 8 shows the change in the temperature of the pin specimens (unreinforced AC8A alloy and composites) during the wear test. There was only a slight difference in the behavior among the curves of the unreinforced alloy and those of the composites under the low v and P; v and P were 0.3 m/s and 17 N, respectively (Fig. 8(a)). A significant difference in the behavior was observed when v was increased to 3.0 m/s (Fig. 8(b)); the temperature of the unreinforced alloy sharply rose, while that of the composites slightly rose. The difference between the curves of the two composites was still low. In contrast, under the high load of 37 N, the difference between the two composites was clearly observed; the temperature of the PAN-based fiber composite sharply rose, while that of the pitch-based fiber composite had only slightly increased (Fig. 8(c)). Under this condition, the temperature of the unreinforced alloy was not able to be accurately measured due to the severe wear of the pin specimen.
Change in temperature of the pin specimens during the wear test.
Figure 9 shows the surface roughness values (Ra) of the specimens after the wear test. Although the change in the average values was not apparent by the fiber-reinforcement at every sliding speed, the scatter of the values due to the measured points decreased by the reinforcement.
Surface roughness values (Ra) of the specimens after the wear test.
Scanning electron micrographs and the X-ray maps of the worn surfaces of the counterparts are shown in Fig. 10 (v = 3.0 m/s, P = 17 N). Figure 10(a) indicates that the surface of the counterpart combined with the unreinforced alloy mainly consisted of aluminum. On the other hand, iron, carbon and only a slight amount of aluminum were detected on the worn surfaces of the counterparts combined with the composites, suggesting that the carbon fiber has a strong effect on suppressing the seizure of the aluminum alloy with the counterpart. The distribution of the carbon suggests that most of fibers near the worn surface had crumbled into small particles. The comparison between Figs. 10(b) and (c) revealed that a wide area of the surface was covered with the crumbled fibers when combined with the pitch-based fiber composite, while the covered area was limited when combined with the PAN-based fiber composite. A similar tendency was observed when P was 37 N, as shown in Fig. 11.
Scanning electron micrographs and the X-ray maps of worn surfaces of the counterparts (v = 3.0 m/s, P = 17 N).
Scanning electron micrographs and the X-ray maps of worn surfaces of the counterparts (v = 3.0 m/s, P = 37 N).
Cross-sectional micrographs of the worn surfaces of the unreinforced alloy and composites (v = 3.0 m/s, P = 37 N) are shown in Fig. 12. The mechanically milled layer (MML) with a thickness of about 45 μm was obviously observed near the worn surface of the unreinforced alloy (Fig. 12(a)), while it cannot be clearly observed for the composites (Figs. 12(b) and (c)). Near the worn surfaces of both composites, neither the extraction nor the delamination of the fibers from the matrix were observed, while the crumbling of the fibers near the surfaces was observed. Plastic flow was hardly observed for the pitch-based fiber composite, while a slight plastic flow with a thickness of about 10 μm can be observed near the worn surface of the PAN-based fiber composite.
Cross-sectional micrographs of worn surfaces of unreinforced alloy and composites (v = 3.0 m/s, P = 37 N). MML in the Fig. (a) represents the mechanically milled layer.
The effects of the carbon fiber-reinforcement on the wear behavior of the AC8A alloy were initially discussed. As shown in Fig. 12(a), the MML was observed near the worn surface of the unreinforced alloy. The MML is usually formed by undergoing repetitive plastic deformation. The plastic flow near the surface causes the scraping and removal of the layer that forms the wear debris, which leads to the wear loss of the alloy. The seizure, which is generated by the frictional heat in the contact portion between the metals during the wear, would promote the plastic flow near the surface and increase the scatter of the coefficient of friction during the wear. At every sliding speed and load, the wear loss, the scatter of the coefficient of friction during the wear test and the scatter of the surface roughness of the AC8A alloy decreased due to the fiber-reinforcement (Figs. 4, 6, 7 and 9). In addition, based on Figs. 7, 8(b) and 10, the carbon fiber-reinforcement has a strong effect on suppressing the seizure of the aluminum alloy with the counterpart. Although the MML was not clearly observed in the cross-sectional microstructures of the composites (Figs. 12(b) and (c)), crumbled fibers were observed on the worn surfaces (Figs. 10 and 11). These results indicated that the carbon fibers formed a solid lubricant film on the worn surfaces thus preventing seizure of the matrix with the counterpart, as previously reported regarding the wear behavior of the aluminum composite reinforced with the PAN-based carbon fiber, graphite fiber and graphite flake5–7,9,10). Due to the microscopic roughness, the direct contact area on the contact face between the pin (aluminum) and disk (counterpart) specimens would be limited. The increase in the load and sliding speed would increase the frictional heat on the contact part between the pin and disk during the wear, then the aluminum matrix (and the steel counterpart) would be softened and the direct contact area would increase. The increase in the contact area would clarify the effect of the fiber as a solid lubricant. This would lead to the fact that the effect of the fiber-reinforcement was pronounced under a high load and high sliding speed, as shown in Fig. 6.
Of particular note is that the wear loss and the scatter of the coefficient of friction of the pitch-based fiber composite were lower than those for the PAN-based fiber one. As shown in Table 1, since the differences in the length and diameter of the two fibers were low, the influence of these factors on the wear behavior of the composites would also be low. Table 1 also shows that the thermal conductivities of the PAN-based and the pitch-based fiber are 10 W/(m・K) and 900 W/(m・K), respectively. The difference in these values led to the higher thermal conductivity of the pitch-based fiber composite, as shown in Table 2. The high thermal conductivity can improve the radiation of heat which was generated by the friction in the pin specimen, leading to the suppression of seizure between the aluminum alloy matrix and the counterpart. The decrease in the seizure would lead to a decrease in the wear of both the pin and disk, and the scatter in the coefficient of friction value during the wear test. In the present study, at the low sliding speed and applied load (0.3 m/s and 17 N, respectively), there was a slight difference in the wear rate (Fig. 6(a)) and temperature behavior during the wear (Fig. 8(a)) between the unreinforced alloy and composites. A significant difference in the wear rate and temperature behavior was observed when v was increased to 3.0 m/s (Figs. 6(a) and 8(b)); the wear rate and temperature of the unreinforced alloy were greater than those of the two composites. Under these conditions, the difference in the wear rate and temperature between the two composites was still very slight. Under the high load of 37 N, however, the wear rate and temperature of the PAN-based fiber composite were significantly greater than those of the pitch-based fiber composite. These results indicate that there is an obvious relation between the wear rate and the temperature behavior; the pitch-based fibers in the composite would act as a heat radiator near the worn surface during the wear, leading to the suppression of the frictional heat and seizure, and subsequent decrease in the wear loss.
Another significant difference in the properties of these two fibers is the elastic modulus. Although a catalogue value of the elastic modulus of the pitch-based fiber (940 GPa) was about 4 times greater than that of the PAN-based fiber (230 GPa) (Table 1), the elastic modulus of the pitch-based fiber composite (79 GPa) was lower than that of the unreinforced AC8A alloy (88 GPa) and PAN-based fiber composite (119 GPa), as shown in Table 2. The elastic moduli of the fibers shown in Table 1 are the ones measured in the longitudinal direction of the fiber. It is generally considered that the high elastic modulus of the pitch-based fiber is attributed to the graphite crystals highly-aligned in the longitudinal direction. This leads to the low elastic modulus in the transverse direction of the fiber and the value in the transverse direction of the pitch-based fiber would be much lower than that of the PAN-based fiber. Therefore, when the fibers are unidirectionally aligned in the composites, the increase in the elastic modulus of the alloy by the fiber-reinforcement is expected in the longitudinal direction. However, when the fibers were randomly aligned in the composites such as those in the present study, the reinforcement effect would not be fully obtained. This would be why the elastic modulus of the pitch-based fiber composite was lower than that of the PAN-based fiber composite. Naplocha et al.9) reported that the wear resistance of the graphite fiber-reinforced aluminum composite was better than that of the graphite flake-reinforced composite, because the weak layers of the matrix near the worn surface easily broke and delaminated for the composite with the graphite flake. This result indicates that the fiber-reinforcement restricts the plastic flow or delamination of the surface layer. Since the elastic modulus of the PAN-based fiber composite was greater than that of the pitch-based fiber composite (Table 2), the PAN-based fiber composite can be expected to have a greater effect on restricting the plastic flow. However, in the present study, the plastic flow near the worn surface of the PAN-based fiber composite was more remarkable than that of the pitch-based fiber composite (Figs. 12 (b) and (c)). This indicated that the reinforcement with the PAN-based fiber, which increased the elastic modulus of the alloy, had a low effect on suppressing the formation of the plastic flow; the plastic flow observed in the PAN-based fiber composite would be mainly due to its low thermal conductivity.
These results lead to the conclusion that the elastic modulus does not have a significant effect on the wear, but the thermal conductivity has a greater effect on decreasing the wear in the present study.
The effects of the reinforcement with the carbon fibers on the wear behavior of the AC8A aluminum alloy were investigated. The following conclusions were obtained.
