Fabrication of Vapor-Grown Carbon Fiber-Reinforced Magnesium-Calcium Alloy Composites by Compo-Casting Process

Abstract

Magnesium-calcium alloy composites reinforced with nickel-coated vapor-grown carbon fibers (VGCFs) were fabricated using a compo-casting process. Then, the microstructures and mechanical properties of these composites were investigated. The Mg-5Al-3Ca (AX53) alloy exhibited a dendritic microstructure with a coarse lamellar (Mg, Al)₂Ca phase along the grain boundaries instead of the irregular β-Mg₁₇Al₁₂ phase found in the Mg-5Al alloy. For the 0.5% Ni-coated VGCF-reinforced AX53 alloy composite, the VGCFs were well dispersed in the matrix, with the nickel coating diffused into the metal. Al₃Ni compounds formed both inside the grains and on the grain boundaries. The ultimate tensile strength (UTS) and strain-hardening of the AX53 alloy, in comparison with the Mg-5Al alloy, were improved significantly to the point of fracture. Furthermore, an increase in the UTS of the composite was achieved with the addition of 0.5% VGCFs, along with an increase in the total elongation, which could mainly be attributed to the strain hardening during a larger strain. The 0.2% yield stress was slightly improved as a result of the small amount of introduced Ni-coated VGCFs. However, the elongation dropped for the 1.0% VGCF-reinforced AX53 alloy composites, which led to a low strength similar with that of the AX53 alloy.

1. Introduction

Vapor-grown carbon fibers (VGCFs) possess attractive mechanical, electrical, and thermal properties, with a carbon structure similar to that of carbon nanotubes.^1,2) VGCFs have attracted much attention in various fields, and have been particularly considered as an ideal candidate for the reinforcement of multifunctional composites and engineering applications. Recent studies on the use of carbon nanofibers in light metal matrix composites have mainly involved the enhancement of the mechanical properties,^3,4) along with the modification of the electrical conductivity,^5,6) thermal conductivity and coefficient of thermal expansion.^7,8)

A magnesium matrix composite has one of the lowest densities of metal matrix composites, has along with a high specific strength and specific stiffness, and excellent mechanical and physical properties. It has been given attention in the metal-matrix composite field, and applied in the aerospace, automotive, and military filed. Mg-Al-Ca alloys have been developed for elevated temperature applications.^9–12) The addition of calcium significantly improves their high-temperature strength, creep resistance, and oxidation resistances. To enhance the strength and rigidity of the matrix, VGCFs can be used to reinforce Mg-Al-Ca alloy, which are supposed to give them satisfactory mechanical properties at both room and elevated temperatures.

Among the variety of manufacturing processes available for discontinuous metal matrix composites, stir casting is easily adaptable and economically viable. Its advantages lie in its simplicity, flexibility, and applicability to large-volume production. However, there are some problems associated with the stir casting of metal matrix composites, including poor wettability and the heterogeneous distribution of the reinforcement material.

The poor wettability of the reinforcement in the melt prevents the reinforcement from infiltrating the molten matrix, with the result that it simply floats on the melt surface. This is due to the surface tension, very large specific surface area, and high interfacial energy of the reinforcements, along with the presence of oxide films on the melt surface. Compo-casting is a liquid state process in which the infiltration of reinforcement into a semi-solid metal (SSM) can be facilitated by means of agitation. In this process, after the material is completely melted, the melt is cooled to a semi-solid state, and the preheated reinforcements are added and mixed. Then, the slurry is again heated to a fully liquid state and mixed thoroughly. Reduced fluidity can be achieved in the SSM by means of shearing.¹³⁾ The primary solid particles already formed in the SSM can mechanically entrap the reinforcing particles, prevent their gravity segregation, and reduce their agglomeration,^14–16) which will result in better distribution of the reinforcement particles.

The uniform distribution of the reinforcement within the matrix and its bond strength with the matrix are essential structural requirements for a stronger metal matrix composite. The wettability and distribution of the reinforcement become more difficult because of the small size of VGCFs. This is due to the large surface area and surface energy of the particles, which cause an increasing tendency for agglomeration. Thus, an intermediate layer of nickel is coated on the carbon fibers to facilitate the wetting. Liquid metals almost always wet solid metals, and the highest wettability is found in the case of mutual solubility or the formation of inter-metallic compounds. Infiltration is thus made easier by the desorption of a metallic coating on the surface of the reinforcing solid.¹⁷⁾ Nickel is frequently used for coating reinforcement particles used with aluminum-based composites.^18–21) It has been found that nickel coatings can improve the wetting behavior of carbon fibers by molten aluminum and limit fiber segregation to obtain a homogeneous reinforcement distribution. Our previous study revealed that an improvement in the wettability of a magnesium alloy on a nickel-coated graphite sheet was achieved through the dissolution of the nickel into the liquid magnesium alloy.²²⁾

In this study, Ni-coated VGCF-reinforced Mg-Al-Ca alloy composites were fabricated using a compo-casting process. Then, the effects of the Ni-coated VGCFs on the microstructure and mechanical properties of the magnesium alloy were investigated, as well as the strengthening mechanism of the Ni-coated VGCF-reinforced Mg-Al-Ca alloy composites.

2. Experimental Methods

The Mg-5Al-3Ca (AX53) alloy was used as the matrix. This alloy was fabricated by permanent mold casting using the following chemical composition (mass%): 4.62 Al and 2.95 Ca, with the balance Mg, as described in Ref. 22). The Mg-5Al alloy was also fabricated for comparison. The carbon nanofibers used were commercially available VGCFs (Showa Denko Co. Ltd.). These VGCFs were typically 100–200 nm in diameter and 10–20 μm in length.

The VGCFs were coated with nickel using an electroless deposition process. Prior to this electroless plating, the VGCFs were pre-treated. To remove any rosin, 10 g dm⁻³ of VGCFs were placed in acetone and subjected to ultrasonic treatment. The VGCFs were dispersed homogeneously in the solution, filtered using a paper filter, and rinsed with pure water. Then, they were etched with 1.69 mol dm⁻³ HNO₃ for 5 min, sensitized with a 5.9 × 10⁻² mol dm⁻³ SnCl₂·2H₂O + 0.24 mol dm⁻³ HCl solution for 5 min, and activated with a 2.3 × 10⁻³ mol dm⁻³ PdCl₂ + 0.48 mol dm⁻³ HCl solution for 5 min. After filtration and rinsing, the VGCFs were immersed in a nickel electroless plating bath. The pH of the bath was adjusted to 6.5 by the addition of an NH₃ solution. Electroless plating was performed for 15 min at 385 K under ultrasonic treatment. All of the solution was pumped through the preform. After filtration and rinsing with methanol, the electroless-plated VGCFs were dried for 120 min at 363 K. The material deposited by the electroless process was then examined using an electron probe micro-analyzer (EPMA, JXA-8900RL) equipped with a wavelength-dispersive spectroscopy (WDS) detector. An X-ray diffraction (XRD) analysis was carried out using Cu-Kα radiation (k = 1.54056 Å) at a scanning speed of 1°/min.

The Ni-coated VGCF-reinforced AX53 alloy composites (Ni@VGCFs/AX53) were fabricated using a compo-casting process. A charge of 50 g of the AX53 alloy was placed in a mild steel crucible preheated to 673 K in an electric resistance furnace. Argon gas was allowed to pass through the furnace to avoid burning the Mg during melting. The furnace temperature was raised to 973 K and held for about 30 min. The melt was homogenized for 2 min by stirring. Then the temperature was brought down to around 878 K, and Ni-coated VGCFs (0.5, 1.0 mass%) wrapped in Al foil were added to the melt during stirring. After 10 min, it was again rapidly heated to 973 K rapidly. The composite melt was stirred for 5 min and poured into a permanent steel mold to form a 120 mm × 14 mm × 14 mm ingot.

The as-cast magnesium alloys and composites were cut into various specimens for the following analyses. The microstructures of the magnesium alloys and composites were observed using the EPMA and a scanning electron microscope (FEI Siron200 SEM). The specimens of as-cast alloy were etched in a hydrochloric acid solution (5% HCl + 95% alcohol). The average grain size was measured using Image-Pro Plus 5.0 software. The values reported for Vickers hardness represented the average of five separate measurements taken at randomly selected points using a load of 100 g for 10 s. Tensile tests were carried out using a universal testing machine with a strain rate of 0.5 mm/min at room temperature. Plate tensile specimens with a thickness of 2 mm and gauge length of 18 mm were used in accordance with ASTM test method E8M-11.

3. Results and Discussion

3.1 Characterization of VGCFs and Ni-coated VGCFs

Figure 1 shows SEM micrographs of the raw VGCFs and electroless nickel-plated VGCFs. The nickel was homogeneously deposited on the VGCFs, resulting in nickel-coated VGCF powder (Fig. 1(b)) with a larger diameter than the raw VGCFs (Fig. 1(a)). A qualitative analysis using WDS showed the presence of about 70 mass% nickel in the Ni-coated VGCFs. Phosphorus, at about 1–2 mass%, was present in the nickel coatings. Figure 2 shows XRD patterns of the VGCFs and Ni-coated VGCFs. The Ni-coated VGCFs exhibit broad humps corresponding to nanocrystals or some amorphous nickel.

Fig. 1

Microstructures of (a) raw VGCFs and (b) nickel-coated VGCFs obtained by electroless deposition with WDS analysis.

Fig. 2

X-ray diffraction patterns of (a) raw VGCFs and (b) nickel-coated VGCFs.

3.2 Microstructures

The microstructures of the as-cast Mg-5Al, AX53, and Ni@VGCFs/AX53 composites are shown in Fig. 3. The morphology of the as-cast Mg-5Al alloy included α-Mg and an irregular β-Mg₁₇Al₁₂ intermetallic compound along the grain boundaries (Fig. 3(a)). The AX53 alloy exhibited a dendritic microstructure with a Ca-containing phase along the grain boundaries (Fig. 3(b)). A coarse lamellar phase with bright contrast formed along the grain boundaries instead of the irregular β phase in the Mg-5Al alloy (Fig. 3(c)). With an Al:Ca ratio of 5:3, the intermetallic compounds mainly consisted of the (Mg, Al)₂Ca phase with a dihexagonal C36 crystal structure.²³⁾ The (Mg, Al)₂Ca phase is a ternary Laves compound and is different from other Laves phases such as Mg₂Ca (C14, hexagonal) and Al₂Ca (C15, cubic). Figure 3(d)–(f) shows SEM images of the 0.5% Ni@VGCFs/AX53 composites. There are also some (Mg, Al)₂Ca compounds distributed along the grain boundaries. In addition, some cube-shaped Al₃Ni particles with an orthorhombic structure were generated both inside grains and on grain boundaries, because Al-Ni intermetallic compounds prefer to form in the Mg-Al-Ni binary system.²⁴⁾ However, the Al₃Ni particles were not evenly distributed, especially in the 1.0% VGCF composites (Fig. 3(g)–(i)). There were also areas in which the VGCFs were not evenly dispersed and formed clusters, as shown in Fig. 3(d) and (g). Even with the nickel coating on the VGCFs, it was difficult to perfectly disperse the VGCFs perfectly using mechanical agitation for 5 min.

Fig. 3

Microstructures of as-cast (a) Mg-5Al, (b), (c) AX53, (d), (e), (f) 0.5%Ni@VGCFs/AX53, and (g), (h), (i) 1.0%Ni@VGCFs/AX53.

Figure 4 shows the distributions of the grain sizes in the composites, as well as the average grain sizes. With an increase in the amount of Ni-coated VGCFs, the grain size of the composite decreased, along with the range of the grain distribution. This indicated that Ni-coated VGCFs have an effect on the grain refinement of an AX53 alloy. On one hand, this was because, as heterogeneous nuclei, the VGCFs promoted nucleation during the magnesium alloy crystallization. On the other hand, a large amount of dispersively distributed VGCFs impeded the grain growth.

Fig. 4

Distributions of grain sizes and average grain sizes of (a) AX53, (b) 0.5%Ni@VGCFs/AX53, and (c) 1.0%Ni@VGCFs/AX53.

Figure 5 shows the wetting and dispersion of the Ni-coated VGCFs in metal melts. The preliminary mix of the Ni-coated VGCFs into the SSM was facilitated by compo-casting process. The primary solid particles already formed in the SSM increased the melt viscosity and mechanically entrapped the reinforcements in the SSM. During this process, the Ni-coated VGCFs existed mostly in the form of clusters. Then, the slurry was again heated to a fully liquid state and mechanically mixed, which further dispersed the Ni-coated VGCF clusters. Moreover, according to our previous work,²²⁾ a nickel coating on graphite improves the wettability of a magnesium alloy through the dissolution of nickel into the liquid magnesium alloy. With nickel coatings diffused into the metal, a homogeneous VGCF distribution and high bond strength with the matrix were obtained. In addition, Al₃Ni intermetallic compounds formed during casting.

Fig. 5

Schematic of wetting and dispersion of Ni-coated VGCFs in metal melts. (a) Preliminary mix of Ni-coated VGCFs into semi-solid metal, (b) Dispersion of VGCFs with nickel coating diffusing into the melt during agitation, (c) Al₃Ni phase formation during casting.

3.3 Micro-Vickers hardness

The micro-Vickers hardness values of the Mg-5Al, AX53 and Ni@VGCFs/AX53 composites are shown in Fig. 6. The hardness of the AX53 alloy was obviously improved by the addition of calcium. This was attributed to the formation of (Mg, Al)₂Ca compounds, which have higher hardness values than Mg₁₇Al₁₂. Compared to AX53, the average hardness values of the 0.5% and 1.0% Ni-coated VGCFs/AX53 increased by approximately 18% and 24%, respectively. The hardness of the magnesium composites increased with the increasing content of Ni-coated VGCFs in AX53. The dependence of the hardness on the grain size is generally well-established, and can be described by the Hall–Petch relation²⁵⁾ for hardness:   

\[ {\rm Hv} = H_0 + k_Hd^{- 1/2} \](1)

where $H_0$ and $k_H$ are the appropriate constants associated with the hardness measurements. However, the $k_H$ slope is 226, which is much larger than that of some magnesium alloys. The improved hardness of the composites was attributed not only to the refined grain, but also to the introduction of Al₃Ni particles and VGCFs. Because Al₃Ni is inherently much harder than magnesium,²⁶⁾ the presence of Al₃Ni intermetallic compounds caused a higher resistance to the localized deformation of the matrix during indentation.

Fig. 6

Microhardness values of as-cast Mg-5Al, AX53, 0.5%Ni@VGCFs/AX53, and 1.0%Ni@VGCFs/AX53 (each value is the average of at least five tests).

3.4 Tensile properties

The stress–strain curves for the Mg-5Al, AX53, and Ni@VGCFs/AX53 composites are shown in Fig. 7. Their tensile properties are summarized in Table 1. It can be observed that the ultimate tensile strength (UTS) of the AX53 alloy was improved by the addition of calcium to the Mg-5Al alloy, while the total elongation was significantly reduced. The addition of 0.5% VGCFs improved the 0.2% yield stress, UTS, and elongation of the AX53 matrix by 6%, 14% and 43%, respectively. The increase in the tensile property was attributed to the overall effect derived from the refined grain and introduction of both VGCFs and Al₃Ni intermetallic compounds. The 0.2% yield stress was slightly improved as a result of the small amount of introduced Ni-coated VGCFs. Moreover, the increase in the UTS of the composite was achieved with increased total elongation, which could mainly be attributed to the strain hardening during a larger strain. However, with the addition of 1.0% VGCFs, the elongation of the composite dropped even lower than that of the AX53 matrix alloy with a similar UTS.

Fig. 7

Tensile stress–strain curves of as-cast Mg-5Al, AX53, 0.5%Ni@VGCFs/AX53, and 1.0%Ni@VGCFs/AX53.

Table 1 Tensile properties of as-cast Mg-5Al, AX53, 0.5% Ni@VGCFs/AX53, and 1.0% Ni@VGCFs/AX53.

Material	0.2% yield stress (MPa)	Ultimate tensile strength (MPa)	Fracture strain (%)
Mg-5Al	51.4	132.5	5.2
AX53	101.5	141.0	1.4
0.5%Ni@VGCFs/AX53	107.7	161.8	2.0
1.0%Ni@VGCFs/AX53	111.8	144.7	1.3

The fracture surfaces of the magnesium alloys are shown in Fig. 8. The fracture surface of the Mg-5Al alloy shows the presence of many dimples (Fig. 8(a), (b)). This indicates that the fracture mechanism is ductile fracture. For the AX53 alloy, a fracture analysis revealed that the mechanism was brittle fracture. As shown in Fig. 8(c) and (d), the fracture surface of the AX53 alloy shows typical intergranular fracture features. It is characterized by equiaxed facets that are correlated with the grain size, and no dimple is observed. This is attributed to the continuous network-shaped distribution of the brittle (Mg, Al)₂Ca phase along the grain boundaries. Figure 8(e) shows the presence of shrinkage cavities, which could form from solidification shrinkage in the AX53 alloy. Figure 8(f) shows the dendritic arms associated to shrinkage cavities on the fracture surface. Micro-cracks near the defects produce and grow along the grain boundaries, which leads to the failure of the material. The tensile strength and elongation of the AZ91 magnesium alloy at ambient temperature were reduced by the Ca addition, but this produced elevated temperature strengthening.²⁷⁾ Masoumi et al. reveal that the presence of the (Mg, Al)₂Ca phase was the main cause of the intergranular fracture in the Mg–Al–Ca alloy.²⁸⁾ The coarse (Mg, Al)₂Ca compound is more brittle and easily broken than the fine Mg₂Ca compound.²⁹⁾ The application of hot extrusion cracked the secondary phase along the grain boundaries in the as-cast alloy, which were dispersedly distributed by the fine spherical secondary phases along the extrusion direction.³⁰⁾ The ductility of the as-extruded alloys at room temperature was high compared with other magnesium alloys containing Ca. However, the ductility was enhanced at elevated temperature, and climb-controlled dislocation creep could be a dominant deformation process.³¹⁾

Fig. 8

SEM images of fracture surfaces of (a), (b) Mg-5Al alloy and (c), (d) AX53 alloy, along with (e) shrinkage cavities in AX53 alloy and (f) dendritic arms associated to shrinkage cavities on fracture surface of AX53 alloy.

The fracture surfaces of the composites essentially revealed brittle fracture, as shown in Fig. 9(a) and (e). With the addition of Ni-coated VGCFs to the matrix alloy, the grains were refined, and the defects decreased, as shown in Fig. 9(b). The increased total elongation of the composite could be attributed to the decrease in the density of shrinkage cavities and presence of VGCFs and Al₃Ni intermetallic compounds on the grain boundaries, which could increase the resistance to crack propagation. Moreover, according to the load transfer reinforcement mechanism,³²⁾ VGCFs well bonded with the magnesium matrix could generate elastic deformation to fit the deformation of the matrix during stretching. Because of the high interface bonding strength, their interface could seldom separate, and there were few pulled-out VGCFs, as shown in Fig. 9(c). In addition, the improved mechanical properties of the composite could be attributed to grain refinement, which could be explained by the Hall–Petch model.³³⁾ However, as shown in Fig. 9(d), clusters of Ni-coated VGCFs were clearly observed. These clusters probably increased the micro-cracks density for their generation and propagation of the specimens, which led to the deterioration of the mechanical properties of the composites. For the 1.0% Ni@VGCFs/AX53, when a relatively larger amount of VGCFs was added to the Mg matrix, it contained more VGCF clusters (Fig. 9(f)). These clusters prevented effective bonding between the Mg and VGCFs and led to minute cracks in the matrix. These cracks inevitably led to the failure of the material with low strength.

Fig. 9

SEM images of (a), (c), (d) fracture surface of 0.5%Ni@VGCFs/AX53, along with (b) shrinkage cavities in 0.5%Ni@VGCFs/AX53 and (e), (f) fracture surface of 1.0%Ni@VGCFs/AX53.

4. Conclusions

(1) Nickel-coated VGCFs were prepared using an electroless plating process, with the nickel homogeneously deposited on the VGCFs.

(2) Ni-coated VGCF-reinforced Mg-5Al-3Ca composites were fabricated using the compo-casting method. The Mg-5Al-3Ca alloy exhibited a dendritic microstructure with a coarse lamellar (Mg, Al)₂Ca phase in the shape of a continuous network along the grain boundaries. The addition of Ni-coated VGCFs could refine the grain of the AX53 alloy. With the nickel coating diffused into the metal, the VGCFs were well dispersed. Al₃Ni compounds formed both inside the grains and on the grain boundaries.

(3) The Mg-5Al-3Ca alloy exhibited a higher UTS and strain-hardening than the Mg-5Al alloy, while the total elongation decreased significantly. With a 0.5 mass% addition of Ni-coated VGCFs, an increase in the UTS of the composite was achieved, along with an increase in the total elongation, which could mainly be attributed to the strain hardening during a larger strain. However, for the 1.0% VGCF-reinforced AX53 alloy composites, the elongation dropped, which led to a low strength similar to that of the AX53 matrix alloy.

Acknowledgements

This study was supported by the Light Metal Educational Foundation of Japan.

REFERENCES

• 1)  Show Denko Co., Ltd., Material Safety Data Sheet, (2007) p. 1.
• 2)  M. Endo, Y.A. Kim, T. Hayashi, K. Nishimura, K. Miyashita and M.S. Dresselhaus: Carbon 39 (2001) 1287–1297. 10.1016/S0008-6223(00)00295-5
• 3)  L. Wang, H. Choi, J.M. Myoung and W. Lee: Carbon 47 (2009) 3427–3433. 10.1016/j.carbon.2009.08.007
• 4)  A.M.K. Esawi, K. Morsi, A. Sayed, M. Taher and S. Lanka: Composites Part A 42 (2011) 234–243. 10.1016/j.compositesa.2010.11.008
• 5)  G. Sasaki, F. Kondo, K. Matsugi and O. Yanagisawa: Trans Tech Publications. 561 (2007) 729–732.
• 6)  Z.F. Xu, Y.B. Choi, K. Matsugi, D.C. Li and G. Sasaki: Mater. Trans. 50 (2009) 2160–2164. 10.2320/matertrans.MAW200917
• 7)  J.M. Ting and M.L. Lake: J. Mater. Res. 10 (1995) 247–250. 10.1557/JMR.1995.0247
• 8)  K. Fukuchi, K. Sasaki, K. Katagiri, T. Imanishi and A. Kakitsuji: Procedia Eng. 10 (2011) 912–917. 10.1016/j.proeng.2011.04.150
• 9)  R. Ninomiya, T. Ojiro and K. Kubota: Acta Metall. Mater. 43 (1995) 669–674. 10.1016/0956-7151(94)00269-N
• 10)  A.A. Luo, M.P. Balogh and B.R. Powell: Metall. Mater. Trans., A 33 (2002) 567–574. 10.1007/s11661-002-0118-1
• 11)  A. Suzuki, N.D. Saddock, J.W. Jones and T.M. Pollock: Scr. Mater. 51 (2004) 1005–1010. 10.1016/j.scriptamat.2004.07.011
• 12)  L. Han, H. Hu and D.O. Northwood: Mater. Lett. 62 (2008) 381–384. 10.1016/j.matlet.2007.05.047
• 13)  S. Naher, D. Brabazon and L. Looney: J. Mater. Process. Technol. 143–144 (2003) 567–571. 10.1016/S0924-0136(03)00368-6
• 14)  M.C. Flemings: Metall. Trans. 22 (1991) 957–981. 10.1007/BF02661090
• 15)  S. Naher, D. Brabazon and L. Looney: J. Mater. Process. Technol. 166 (2005) 430–439. 10.1016/j.jmatprotec.2004.09.043
• 16)  Z. Fan: Int. Mater. Rev. 47 (2002) 49–85. 10.1179/095066001225001076
• 17)  F. Delannay, L. Froyen and A. Deruyttere: J. Mater. Sci. 22 (1987) 1–16. 10.1007/BF01160545
• 18)  H. Chen and A.T. Alpas: Wear 192 (1996) 186–198. 10.1016/0043-1648(95)06795-7
• 19)  T. Rajan, R. Pillai and B. Pai: J. Mater. Sci. 33 (1998) 3491–3503. 10.1023/A:1004674822751
• 20)  Y.M. Ryu, E.P. Yoon and M.H. Rhee: J. Mater. Sci. Lett. 19 (2000) 1103–1105. 10.1023/A:1006780212533
• 21)  J. Rams, A. Urena and M.D. Escalera: Compos. 38 (2007) 566–575. 10.1016/j.compositesa.2006.02.010
• 22)  Y. Yao, Z. Xu, K. Sugio, Y. Choi, S. Kang, R. Fu and G. Sasaki: Mater. Trans. 56 (2015) 1693–1697. 10.2320/matertrans.MAW201506
• 23)  A. Suzuki, N.D. Saddock, J.W. Jones and T.M. Pollock: Acta Mater. 53 (2005) 2823–2834. 10.1016/j.actamat.2005.03.001
• 24)  H. Wang, L. Liu and F. Liu: Mater. Des. 50 (2013) 463–466. 10.1016/j.matdes.2013.02.085
• 25)  Y.S. Sato, M. Urata, H. Kokawa and K. Ikeda: Mater. Sci. Eng. A 354 (2003) 298–305. 10.1016/S0921-5093(03)00008-X
• 26)  H. Chen and A.T. Alpas: Wear 192 (1996) 186–198. 10.1016/0043-1648(95)06795-7
• 27)  W. Qudong, C. Wenzhou, Z. Xiaoqin, L. Yizhen, D. Wenjiang and M. Mabuchi: J. Mater. Sci. 36 (2001) 3035–3040. 10.1023/A:1017927109291
• 28)  S.W. Xu, N. Matsumoto, K. Yamamoto, S. Kamado, T. Honma and Y. Kojima: Mater. Sci. Eng. A 509 (2009) 105–110. 10.1016/j.msea.2009.02.024
• 29)  M. Masoumi and H. Hu: Mater. Sci. Eng. A 528 (2011) 3589–3593. 10.1016/j.msea.2011.01.032
• 30)  L. Zhang, K.K. Deng, K.B. Nie, F.J. Xu, K. Su and W. Liang: Mater. Sci. Eng. A 636 (2015) 279–288. 10.1016/j.msea.2015.03.100
• 31)  H. Watanabe, M. Yamaguchi, Y. Takigawa and K. Higashi: Mater. Sci. Eng. A 454–455 (2007) 384–388. 10.1016/j.msea.2006.12.130
• 32)  J.L. Tsai and T.C. Lu: Compos. Struct. 90 (2009) 172–179. 10.1016/j.compstruct.2009.03.004
• 33)  W.S. Miller and F.J. Humphreys: Scr. Metall. Mater. 25 (1991) 33–38. 10.1016/0956-716X(91)90349-6