2018 Volume 59 Issue 3 Pages 475-481
Friction stir processing (FSP) was applied for the fabrication of carbon fiber (CF) reinforced magnesium (Mg) alloy, AZ91, metal matrix composite (MMC). The CFs were recycled ones extracted from long CF reinforced plastics. A narrow slit was introduced into AZ91 plate and then filled with the chopped CFs with the length about 1 mm. Subsequently, the rotating tool was plunged and traveled along the slit to disperse CFs into the matrix. Two kinds of FSP tools, namely conventional threaded pin tool and 3-flat pin tool were used. Microstructural observation revealed that chopped CFs were broken up into fine ones with the length less than 20 μm by severe stirring of material, and were dispersed in the stir zone (SZ). The 3-flat pin tool reduced the size and number of defects in the SZ compared with the conventional pin tool. The hardness of the SZ was higher than that of the as-cast material and FSPed one without CFs. Axial fatigue tests were conducted using the MMCs fabricated by the 3-flat pin tool to investigate fatigue properties. Fatigue strengths of the MMCs were comparable with those of the as-cast specimens, but lower than those of the FSPed ones without CFs. Fatigue cracks initiated at the agglomerations of CFs in the MMCs. The lower fatigue strengths of MMCs were attributed to the lower fatigue crack initiation resistance resulting from the inhomogeneous distribution of CFs.
Currently, magnesium (Mg) alloys are expected to be increasingly used for structural materials in automotive, shipping and aerospace industries since they have advantages in light weight and high specific strengths. To increase the reliability of components in those field, improvement of mechanical properties of Mg alloys is an important issue from the engineering point of view. Friction stir processing (FSP) is one of the material-strengthening procedures based on the microstructural modification induced by severe plastic deformation. FSP had been developed from the butt welding technique of friction stir welding (FSW). It has been reported that FSP could successfully increase the mechanical properties of light weight alloys such as cast aluminum (Al) alloys1–3) and Mg alloys4–6). Santena et al. applied FSP to cast Al alloys and successfully enhanced ductility and tensile strength by removing casting defects and refining grains1). Chan et al. improved hardness of cast Mg alloy mainly due to the grain refinement by FSP4). The authors had modified the microstructures of cast Al3) and Mg6) alloys by FSP, and concluded that the fatigue properties could be enhanced due to the microstructural modification.
Recently, FSP attracts attention as a new fabrication method for metal matrix composites (MMCs), in which strengthening particles are dispersed in the matrix by severe plastic deformation during FSP7–14). Morisada et al. fabricated SiC particles reinforced Mg alloys7,8) by FSP, and successfully obtained uniform distribution of SiC particles in the matrix and enhanced hardness. In those studies, hard particles such as carbon nanotube (CNT)7,12), SiC8,10,11,13), SiO29), Al2O311,12) and TiO214) are commonly used for the reinforcements of metal matrix. In automotive industries, recently, carbon fiber reinforcing plastics (CFRP) is the key material for the weight saving of car bodies. Therefore, recycling of CFRP became important matter because automobiles are typical mass products. A part of the authors had developed the recycling procedure of extracting continuous CFs from used CFRP with low energy consumption and low cost15). However, it is difficult to re-use the recycled CFs with heat history to form new CFRP, because virgin CFs generally underwent chemical surface treatment for better wettability with the matrix, and the recycled CFs exhibit about 80% of the strength of virgin CFs15). Therefore, it is important to develop the usage of recycled CFs. In the previous study16), recycled CFs were used as the reinforcements of AZ91 based MMCs, and FSP using a conventional pin tool was applied to disperse CFs into the matrix. It has been reported that the CFs in the matrix successfully increased the hardness compared with the FSPed material without reinforcements. However, the processed materials frequently contained worm-hole like defects along the path of FSP16). It was due to low deformability of Mg alloys. Yang et al. had reported that a 3-flat pin tool could give better stirring of material than a conventional pin tool in Mg alloys with low deformability, resulting in the better joint strengths of friction stir welds (FSWs)17). Sahraeinejad et al. also reported that a 3-flat pin tool could provide the microstructure of Mg alloy with enhanced strength and ductility11). In the present study, therefore, the fabrication of recycled CF reinforced MMC was tried by FSP using a 3-flat pin tool. In addition, it is known that the understanding of fatigue performances is important for the actual use of the developed materials, while fatigue researches of MMCs fabricated by FSP are quite limited18). Thus, axial fatigue tests were conducted to evaluate the fatigue performance of the fabricated MMC. The purposes of this study is not only to enhance the mechanical properties of Mg alloy by recycled CFs but also to develop the utilization of recycled CFs.
The material used in this study is a cast Mg alloy, AZ91, whose chemical composition (mass%) is as follows; Al: 8.29, Zn: 0.58, Mn: 0.30, Si: <0.021, Cu: <0.0093, Si: <0.021, Fe: 0.0035, Ni: <0.0006 and Mg: balance. The mechanical properties are as follows; 0.2% proof stress (σ0.2): 67 MPa, Tensile strength (σB): 135 MPa, Elongation (δ): 7%, Reduction of area (φ): 5%, Young's modulus (E): 35 GPa. The microstructures of the as-cast material is shown Fig. 1. Eutectic β phases (Mg17Al12) are seen in the large grains with the average size of 450 μm6). Plates with the length of 100 mm, width of 60 mm and thickness of 6 mm were sampled from the as-cast ingots by electrical discharge machining. Subsequently, narrow slits with the length of 70 mm, width of 1 mm and depth of 2.3 mm were introduced at the center of each plate. In the previous report16), the depth of the slit was set as 5 mm, while defects were frequently formed in SZ due to low deformability of Mg alloys. Thus the depths of the slit was reduced to 2.3 mm in the present case. The continuous long recycled CFs with the diameter of 8 μm are shown in Fig. 2. CFRP was carbonized at 500℃ in nitrogen gas, and then fired at 400℃ to extract CFs from CFRP. The detailed process of getting recycled fibers is described elsewhere15). Residual carbides of matrix are slightly seen on the surfaces of CFs (Fig. 2). The long CFs were chopped into the length of about 1 mm, and subsequently filled into the slits of AZ91 plates.
Microstructures of as-received material: (a) Overview, (b) Magnified view of β phases.
Appearance of recycled carbon fibers.
Figure 3 shows the conventional treaded pin and 3-flat pin tools. The conventional pin tool consists of a concave shoulder with the diameter of 18.5 mm and M8-threaded pin with the length of 4.7 mm as shown in Fig. 3(a), while the 3-flat pin tool has three flat surfaces in M8-threaded pin. Figure 4 indicates the schematic illustration of MMC fabrication procedure by FSP. Prior to FSP, CFs were filled in the slit of the plate as mentioned above. The weight of CFs in the slit was 40 mg. Firstly, the top of the M8-treaded pin or 3-flat pin was plunged at the depth of 0.3 mm and traveled along the slit. Consequently, the slit is sealed and the CFs could remain in the slit during the following FSP. Subsequently, the tool shoulder surface was plunged at the depth of 0.6 mm, and traveled along the slit. In this process, the threaded pin or 3-flat pin was completely inserted into the plate and the top of the pin reached near the bottom of the plate. The tool was rotated clockwise and tilted 3o opposite to the processing direction. The traveling speed of the tool was 100 mm/min and the rotational speed was 1400 rpm according to the low defect density condition proposed in the previous report16).
FSP tools: (a) Conventional threaded pin tool, (b) 3-flat pin tool.
Schematic illustration of fabrication of MMC by FSP.
The solution of picric acid 10 g, acetic anhydride 20 mL, ethanol 100 mL and distilled water 20 mL was used for the microstructural observation. Hardness profiles on the cross section were measured by a micro Vickers hardness tester with a load of 0.98 N and dwell time of 30 s. Figure 5 shows the configuration of the fatigue test specimen and the schematic illustration of specimen sampling from the FSPed plate. It should be noted that the gauge section is included in the microstructurally-modified zone. It is known that FSW sometimes induces defects in the SZ, known as warm-hole defects. Thus, in order to sample the fatigue specimens from defect-free area, both the upper and lower regions were removed where the removal depths were 2.5 mm from top and 1.5 mm from bottom. Prior to fatigue tests, the surface of the gauge section was polished using a #2000 grade emery paper and buff-finished to achieve a mirror surface. Fatigue tests were conducted using an electro-hydraulic fatigue testing machine with 98 kN capacity, using sinusoidal wave form with loading frequency f = 10 Hz and load ratio R = −1 in laboratory air at room temperature. All of the fracture surfaces after fatigue tests were observed by a scanning electron microscope (SEM).
Fatigue test specimen: (a) Specimen configuration, (b) Schematic illustration of specimen sampling from the FSPed plate.
The microstructures of the FSPed AZ91 without any reinforcements are described elsewhere6). The processing conditions had been the tool rotational speed of 800 rpm and the traveling speed of 500 mm/min. The microstructure could be characterized by the breakup of large β phases, grain refinement and the formation of ring-like structure called onion rings. The average grain size in the SZ was 9 μm6). Figure 6 shows the appearance of transverse cross section of the MMC fabricated using the conventional pin tool. The transverse cross section was obtained by cutting the FSPed plate at the center. The ring-like structure is formed in the SZ as seen in Fig. 6(a). The magnified views of the rectangular areas (b) and (c) in Fig. 6(a) are revealed in Figs. 6(b) and 6(c), respectively. The CFs are found in those figures as small black particles dispersed in the Mg matrix. It should be noted that the size of CFs filled in the narrow silt was 1 mm before FSP, while the size of CFs in Figs. 6(b) and 6(c) was less than 20 μm. It should be noted that some very fine CFs are also seen as shown by the arrows in Fig. 6(c). That indicates that the CFs were broken up into small pieces by severe plastic deformation during FSP. The average grain size of the as-cast material is 450 μm, while the size in the SZ is 12 μm, revealing that the grain refinement occurred by FSP6). The longitudinal cross section along FSP path is shown in Fig. 7(a). Some defects are found especially in the upper half region. Figures 7(b), 7(c) and 7(d) are the magnified views at the areas (b), (c) and (d) in Fig. 7(a), respectively. As shown in Figs. 7(b) and 7(c), rather large hole defects are recognized, while CFs are uniformly distributed in the area (d) (Fig. 7(d)). The longitudinal cross section of Fig. 7(a) corresponds to the broken line in the macroscopic view of Fig. 7(e). It reveals that the hole defects tend to locate near the starting area of FSP.
Cross section of MMC fabricated using a pin tool: (a) Macroscopic view, (b) Magnified view of region (b) in Fig. (a), (c) Magnified view of region (c) in Fig. (a). Arrows in Fig. (c) indicate very fine CFs.
Longitudinal section of MMC fabricated using a pin tool: (a) Macroscopic view, (b) Magnified view of region (b) in Fig. (a), (c) Magnified view of region (c) in Fig. (a), (d) Magnified view of region (d) in Fig. (a), (e) Observation are in Fig. (a).
Figures 8 and 9 are the microstructural appearances of the MMC fabricated by the 3-flat pin tool observed on the transverse cross section (Fig. 8) and the longitudinal cross section (Fig. 9), respectively. From those figures, it is found that CFs are dispersed throughout the SZ. It should be noted that the size of CFs is less than 20 μm and some very fine CFs shown by the arrows are seen in Figs. 8(b) and 8(c) similar to the MMC fabricated by the conventional pin tool (Fig. 6(c)). Furthermore, the average grain size in the SZ is 12 μm, and the same with that of the MMC fabricated by the conventional pin tool due to the same processing parameters. On the longitudinal cross section (Fig. 9), small defects are still present as shown in Fig. 9(b). However, defects observed in the MMC fabricated by the 3-flat pin tool are much smaller and the density of defects is lower than those in the MMC fabricated by the conventional pin tool. The observation area of longitudinal cross section is the same with Fig. 7(e). Sahraeinejad et al. used a 3-flat pin tool for the fabrication of Mg alloy based MMC to get better mixing of material11). It is considered that the 3-flat pin tool in the present study also resulted in the better mixing, which induced the reduction of defects.
Cross section of MMC fabricated using a 3-flat pin tool: (a) Macroscopic view, (b) Magnified view of region (b) in Fig. (a), (c) Magnified view of region (c) in Fig. (a). White dotted line in Fig. (a) represents the cross section of fatigue specimen.
Longitudinal section of MMC fabricated using a 3-flat pin tool: (a) Macroscopic view, (b) Magnified view of region (b) in Fig. (a), (c) Magnified view of region (c) in Fig. (a). White dotted lines in Fig. (a) represent the cross section of fatigue specimen.
Figure 10 shows the hardness profiles of the MMC fabricated using the 3-flat pin tool. The hardness profiles of the as-cast material and the FSPed one without CFs in the previous study6) are also exhibited in Fig. 10. Hardness profiles of the FSPed material without CFs and MMC were measured along three lines on the transverse cross section of the mid-thickness, 1.3 mm above and below the mid-thickness. Figure 10 indicates that the hardness of the MMC is higher than that of the as-cast material and FSPed one without CFs.
Hardness profiles.
Figure 11 shows the S-N curves of the MMC fabricated using the 3-flat pin tool, the as-cast material and FSPed ones without CFs6). The fatigue strengths of the as-cast specimens have large scatter, but tend to have a clear knee at around 105 cycles, leading to the fatigue strength at 107 cycles (hereafter, fatigue limit) of 40 MPa. On the other hand, the fatigue strengths of the FSPed specimens without CFs are considerably increased (fatigue limit 80 MPa) compared with the as-cast ones. However, the fatigue strengths of the MMC specimens are similar with those of the as-cast specimens and are lower than those of the FSPed specimens without CFs, especially at the stress levels lower than 80 MPa.
S-N diagram.
Lee et al. revealed that the sound AZ91D friction stir welds could be fabricated at the tool rotational speeds lower than 1432 rpm and traveling speeds ranging from 41 to 180 mm/min19). The FSP condition in the present study was in the range of sound processing parameters, while defects with the size of approximately 200～300 μm were found in the microstructure of the MMC fabricated using the conventional pin tool (Fig. 7). It is considered that the material flow was insufficient to re-fill the narrow slit installed in the initial plate. It could be attributed to the low deformability of Mg alloys. By using the 3-flat pin tool to increase plastic flow, the size and number of defects decreased (Fig. 9). It indicates that the 3-flat pin tool could improve the microstructure more efficiently than the conventional threaded pin tool11) and enhance the hardness as described in the previous section. Hardness enhancement shown in Fig. 10 is reasonably in terms of grain refinement, namely Hall-Petch relationship. It is assumed that the dispersion of CFs also could enhance the hardness. Figure 12 shows the hardness contour map on the transverse cross section of the MMC fabricated using 3-flat pin tool. In the figure, the peak hardness (120 HV) corresponds to the area with high CF density. This tendency is similar as reported by Asadi et al. that the peak hardness in the micro hardness profile appears at the agglomeration of the reinforcement particles in the particular region of the SZ20). It should be noted that the hardness profiles of MMC in Fig. 10 exhibit larger scatter than those of the FSPed material without CFs. Consequently, the mechanical property, namely hardness, could be successfully improved by the combination of grain refinement and CFs dispersion, while the counter map and hardness profiles revealed that the distribution of CFs was not uniform, resulting in the agglomeration of CFs.
Hardness contour map in the transverse cross section of MMC. White dotted line in the figure represents the cross section of fatigue specimen.
The fatigue strengths of the as-cast specimens drastically increased by the simple FSP without CFs. The increase had been attributed to the change of the crack initiation mechanism. Fatigue cracks of the as-cast specimens initiated due to the cracking of large intermetallic compounds in the matrix. In the FSPed specimens, however, fatigue cracks initiated due to cyclic slip deformation of the matrix, because large IMCs were broken up into small pieces by severe stirring of material6). Figure 13 shows the typical example of the fatigue fracture surfaces of the MMC specimens. The magnified view at the fatigue crack initiation site is indicated in Fig. 13(b). It reveals that the fatigue crack initiated at the agglomeration of CFs. From the analyses of all the fracture surfaces of the MMC specimens, it is confirmed that fatigue cracks mainly initiated at the agglomerations of CFs. It has been reported by Sun and Fujii that the agglomeration of reinforced particles can occur during FSW, which deteriorates microstructure and mechanical properties21). In the present study, the agglomerations of CFs are observed in the microstructure of MMC, which could be attributed to the low deformability of Mg alloys. The 3-flat pin tool could have reduced the size and number of macroscopic void defects by the enhanced plastic flow, while the agglomerations of CFs could not be fully removed. However, the hardness was highly improved by the dispersed CFs in the matrix, showing the way to the utilization of recycled CFs.
SEM micrographs showing fracture surfaces of MMC specimen (σa = 80 MPa, Nf = 9.0 × 103): (a) Macroscopic appearance, (b) Magnified view at the fatigue crack initiation site.
The MMC of AZ91 Mg alloy reinforced by recycled CFs was fabricated by FSP using conventional pin and 3-flated pin tools. The microstructures, hardness and fatigue properties of the MMCs were investigated. The main conclusions can be made as follows:
