- Email: [email protected]
WC–Ni composite coatings prepared by plasma spraying
SCT-19922; No of Pages 9 Surface & Coatings Technology xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Surface & Coatings Technology journal homepage: www.elsevier.com/locate/surfcoat
Characterization of microstructure and rolling contact fatigue performance of NiCrBSi/WC–Ni composite coatings prepared by plasma spraying Z.Q. Zhang a, H.D. Wang b,⁎, B.S. Xu b, G.S. Zhang c a b c
School of Materials Science and Engineering, Tianjin University, Tianjin 300072, China National Key Lab for Remanufacturing, Academy of Armored Forces Engineering, Beijing 100072, China Institute of Remanufacture and Processing Material, Tianjin Research Institute of Construction Machinery, Tianjin 300409, China
a r t i c l e
i n f o
Article history: Received 6 September 2014 Accepted in revised form 24 November 2014 Available online xxxx Keywords: NiCrBSi/WC–Ni composite coating Microstructure Rolling contact fatigue Abrasion Spalling
a b s t r a c t In this paper, the properties of NiCrBSi/WC–Ni composite coatings deposited by plasma spraying were analyzed, including their microstructure, element distribution, phase composition, porosity, and microhardness. The rolling contact fatigue (RCF) life and failure modes were also investigated. The results showed that the NiCrBSi/WC–Ni composite coating exhibited a typical lamellar structure. Furthermore, evenly distributed W2C/WC phases were well wetted by other phases. In addition to WC and W2C phases, the NiCrBSi/WC–Ni composite coating consisted mainly of γ-Ni, Cr7C3, Cr23C6, FeNi3, Ni3Si, CrB, and NiC. The study also found that the NiCrBSi/WC–Ni composite coating (porosity of 1.84%) had a slightly lower density than the NiCrBSi coating (porosity of 1.62%). The microhardness of the NiCrBSi/WC–Ni composite coating was significantly enhanced by the addition of hard WC–Ni. Furthermore, the results indicated that the NiCrBSi/WC–Ni composite coating had a higher dispersion degree of the RCF life and a longer life than the NiCrBSi coating under 0.881 GPa. Therefore, under this stress level, the RCF performance of the NiCrBSi/WC–Ni composite coating was superior to that of the NiCrBSi coating. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Plasma spraying is an effective surface engineering technology that has been widely applied in various industrial fields [1–5]. Plasma spraying coating can improve the contact fatigue resistance, wear resistance, and corrosion resistance in metal components such as rollers in paper manufacturing, shafts and screws in petrochemical processes, and transmission shafts in construction machinery and cars. In addition, the plasma spraying technology can be applied to a wide range of materials, such as ceramics, pure metals or alloys, and polymeric materials [6–8]. Previous studies have showed that self-fluxing NiCrBSi alloys exhibit a low melting point, good deoxidizing and slagging properties, and great wettability. Based on their excellent processing properties, they are also particularly suited for the plasma spraying technology [9–11]. Theoretically, B is included in the NiCrBSi alloy to shift the alloy closer to the Ni–Ni3B eutectic composition and reduce the melting temperature of the alloy. In addition, B can form hard phases with Ni, such as Ni3B, thus improving the hardness of the coating. The addition of Si enhances the self-fluxing properties of the NiCrBSi alloy. Unless a large amount is added, Si will remain in solid solution with Ni. Si also slightly increases the hardness of the alloy. Furthermore, Cr can be added to the alloy to improve its corrosion and wear resistance. The ⁎ Corresponding author. Tel.: +86 10 66718475; fax: +86 10 66719325. E-mail address: [email protected] (H.D. Wang).
amount of Cr determines the structure of the alloy. If the content of Cr and C is sufficient, hard carbides such as Cr3C2, Cr7C3, and Cr23C6 can be formed. As the coating formation relies on the rapid solidification of the materials during the plasma spraying process, phase transformations are undoubtedly hindered to some degree, resulting in a unique microstructure. The addition of hard ceramic particles, such as tungsten carbide (WC), in a self-fluxing Ni-based alloy may increase the coating hardness, abrasive wear resistance, and contact fatigue resistance. Compared to other carbides, WC combines favorable properties such as high hardness, plasticity, and a good wettability by molten metals [9]. Therefore, WC is widely used as the hard phase in manufacturing metal composites. Moreover, previous studies have showed that WC can decarburize to W2C and result in a relatively soft W phase. However, WC-coated Ni can prevent and reduce the decarburization during the plasma spraying process [12–16]. Some studies have been performed on the effects of the addition of WC on the performance of the NiCrBSi alloys. The microstructure of a NiCrBSi/WC composite coating prepared by electric arc spraying technology consisted of NiCr, NiCrW, and WC/W2C as major phases [9]. In addition, NiCrBSi coatings and NiCrBSi/WC–Ni composite coatings were prepared on stainless steel by laser cladding technology. The results showed that the NiCrBSi/WC–Ni composite coating had a higher wear resistance at high temperature than the NiCrBSi coating, and the difference was attributed to the formation of a hard WC phase after laser cladding [17]. Several types of mixtures of self-fluxing NiCrBSi
http://dx.doi.org/10.1016/j.surfcoat.2014.11.061 0257-8972/© 2014 Elsevier B.V. All rights reserved.
Please cite this article as: Z.Q. Zhang, et al., Surf. Coat. Technol. (2014), http://dx.doi.org/10.1016/j.surfcoat.2014.11.061
2
Z.Q. Zhang et al. / Surface & Coatings Technology xxx (2014) xxx–xxx
alloy powder and WC–Ni powder (10 wt.% Ni and balance WC) were laser cladded on stainless steel substrates. The study found that most of the clad layer properties, such as its porosity and microhardness, were affected by the percentage of the WC particles present in the mixture. Dense and pore-free layers could be obtained as long as the WC content in the mixture was kept below 50 wt.% [18]. However, the properties of the plasma-sprayed NiCrBSi/WC–Ni composite coatings, such as the microstructure, element distribution, phase composition, porosity, and microhardness, have not been systematically studied yet. Moreover, since the rolling contact fatigue (RCF) failure of the majority of the components originates from the surface or subsurface [19,20], plasma spraying can play an important role in repairing the surfaces of failed parts, thus prolonging their lives [21–23]. Interestingly, the RCF resistance of the NiCrBSi/WC–Ni composite coatings has not been investigated yet. The RCF can be defined as a damage process on the contact surface of the rotating components bearing alternating stress [24]. In this work, we focus on the preparation of a protective NiCrBSi/WC–Ni composite coating on tempered 1045 steel by plasma spraying. The properties of the NiCrBSi/WC–Ni composite coating, including the microstructure, element distribution, phase composition, porosity, and microhardness were analyzed. In addition, the RCF life performance and failure mode were also examined. A NiCrBSi coating was prepared to provide a comparison with the NiCrBSi/WC–Ni composite coating.
2. Experimental procedures 2.1. Preparation of the coatings A GP-80 plasma spraying system manufactured by Taixing Spraying Corporation of China was used to prepare the NiCrBSi/WC–Ni composite coating and NiAl undercoating. Hydrogen, argon, and nitrogen were used as working gas, protective gas, and feeding gas, respectively. Commercial tempered 1045 steel was used as a substrate. Fig. 1 shows two-dimensional and three-dimensional diagrams of both the coated and RCF-tested specimens. The size of the specimens is shown in Fig. 1(a). The length of the line contact in the RCF test is 6 mm. The red surface area shown in Fig. 1(b) is planed to prepare for the coating. Fig. 2 shows NiCrBSi/WC–Ni composite powders with 15 wt.% WC–Ni as sprayed material. The composition of the NiCrBSi alloy powder is Cr 16, B 7.5, Si 4.7, C 0.9, Fe ≤ 4.5, Ni balance (wt.%), with particles of 40–80 μm size and globular shape, as indicated by the arrow in Fig. 2. The composition of WC–Ni is Ni 5, WC balance (wt.%), with particles of 40–100 μm size and irregular shape, as indicated by the arrow in Fig. 2. A NiAl alloy powder with a composition of Ni 80, Al 20 (wt.%) and particle size of
Fig. 2. Morphology of the NiCrBSi/WC–Ni composite powders.
50–100 μm was used as the undercoating material. The bond strength between the coating and the substrate can be significantly increased as a result of the heat generated by the reaction between the melting nickel and aluminum arriving on the substrate. The optimized plasma spraying parameters are listed in Table 1. Fig. 3 illustrates the preparation process of plasma spraying coating. The thicknesses of the NiCrBSi/WC–Ni composite coating and NiAl undercoating were adjusted in the range of 500–600 μm and 200–250 μm, respectively. In addition, a NiCrBSi coating was also prepared to provide a comparison with the NiCrBSi/WC–Ni composite coating. The composition of the NiCrBSi powder was identical to that of NiCrBSi in the NiCrBSi/WC–Ni composite powder. The spraying parameters and the thickness were the same as those of the NiCrBSi/WC–Ni composite.
2.2. Characterization of the coatings The microstructure and RCF failure morphology were observed by scanning electron microscopy (SEM, Philips Quant 200), while the composition and distribution of the elements were analyzed by energy dispersive spectroscopy (EDS, Philips Quant 200). The phase composition was identified using a D8 X-ray diffractometer (40 kV, 30 mA, Cu Kα radiation, λKα = 0.154 nm, 2θ scanning step of 0.03°, scanning in the range of 30° b 2θ b 100°). The microhardness distribution along the depth direction of the coating was determined using a HV-1000A Vickers microhardness tester with a load of 0.98 N and dwell time of 5 s. For each hardness profile along the depth direction of the coating, three tests were performed (the spacing of each press mark was less than 50 μm) and the averaged results of the three repeated tests were used in this article. The porosity of the coating was measured by image processing software based on the gray analysis method. Ten cross-sectional micrographs of the coating at 500 × magnification
Table 1 Plasma spraying parameters.
Fig. 1. Diagram of the coated and RCF tested specimens.
Processing parameters
NiCrBSi/WC–Ni
NiAl
Spraying voltage (V) Spraying current (A) Argon gas flow (L/h) Hydrogen gas flow (L/h) Nitrogen gas flow (L/min) Spraying distance (mm) Coating thickness (μm)
60 500 30 130 6 90–110 500–600
56 500 30 140 5 90–110 200–250
Please cite this article as: Z.Q. Zhang, et al., Surf. Coat. Technol. (2014), http://dx.doi.org/10.1016/j.surfcoat.2014.11.061
Z.Q. Zhang et al. / Surface & Coatings Technology xxx (2014) xxx–xxx
3
Fig. 3. Schematic of the preparation process of the plasma spraying coating.
were randomly collected by SEM, and an average value was calculated from the ten different porosity test results. 2.3. The RCF test As shown in Fig. 4, a double-roller RCF test rig manufactured by National Key Lab for Remanufacturing of China was used to evaluate the RCF performance of the NiCrBSi/WC–Ni composite coating and NiCrBSi coating [25]. A tested roller with the coating and a paired contact roller were fixed on the rotation axis by a clamp and driven by two servo motors. The type of rolling contact is a line contact. The paired contact roller was machined using AISI 52100 steel with a surface roughness of 0.012 μm and a Rockwell hardness of 62 HRC. In addition, the thickness, external diameter, and internal diameter of the paired contact roller were kept to 20 mm, 60 mm, and 32 mm, respectively. The rotation speed can be stepless adjusted in the range of 0–2000 rpm by setting up the software parameters, and the slip ratio can be set in the range of 0–100%. To assure a pure rolling contact form in this experiment, the rotation speeds of both the tested roller and the paired contact roller
were both kept at a constant value of 600 rpm, namely, with a slip ratio of 0%. All the RCF tests were conducted under oil lubrication. The load was applied by hydraulic pressure and, here, set to a constant value of 2000 N. The maximum contact stress, σmax, was calculated using Hertz's equation, which is expressed as:
σ max
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi PE R1 þ R2 ¼ 0:418 l R1 R2
where P (with a unit of N) is the magnitude of the applied load at the line between the tested roller and the paired contact roller, R1 = R2 = 30 mm are the radii of the tested and paired contact roller, respectively, l = 6 mm is the length of the line contact, and E denotes the elastic modulus. The elastic moduli of the tested and paired contact roller are both set to 200 GPa. Thus, the maximum calculated contact stress is 0.881 GPa under a load of 2000 N. Vibration and AE sensors were used to monitor the RCF failure process. In this study, when the vibration and AE signals exceeded the pre-setting threshold and sharply increased as a result of a failure, the test was stopped and the fatigue life of the specimen calculated. As the experimental data from the RCF tests are considerably scattered, ten rolling contact tests were performed in the same conditions to obtain statistically significant results of the RCF life. A new pair of contact rollers was used for each test. 3. Results and discussion 3.1. Characterization of the coatings
Fig. 4. Schematic of the double-roller RCF test rig.
As the surfaces of the NiCrBSi/WC–Ni composite coating and NiCrBSi coating were severely roughened after the plasma spraying process, the specimens were initially ground and polished. The final thickness of the coating was approximately 400–500 μm (including the working coating and undercoating). The surface roughness of the polished coating was measured by a Talysurf 5P-120 surface contour mapping system. The value of the roughness average (Ra) after five measurements was 0.276 μm. The melting NiCrBSi/WC–Ni composite powders, after being heated and accelerated by the plasma arc, were deposited on the substrate. The spraying particles were firmly bonded by the mechanical interlock structure. Shrinkage holes were formed because of an uneven contraction of the melting particles during the cooling process, as
Please cite this article as: Z.Q. Zhang, et al., Surf. Coat. Technol. (2014), http://dx.doi.org/10.1016/j.surfcoat.2014.11.061
4
Z.Q. Zhang et al. / Surface & Coatings Technology xxx (2014) xxx–xxx
Fig. 5. Scanning electron microscopy (SEM) surface micrograph of the NiCrBSi/WC–Ni composite coating.
indicated by the arrows in Fig. 5. The holes could not be completely filled by the subsequent molten particles. A surface SEM micrograph of the NiCrBSi/WC–Ni composite coating after plasma spraying is shown in Fig. 6(a). The area in Fig. 6(a) was scanned and analyzed by EDS. The results show that the NiCrBSi/WC– Ni composite coating mainly comprises W, C, Ni, Cr, Fe, and Si elements, in addition to B, as shown in Fig. 7. However, microscale oxygen element was detected after plasma spraying, mainly because the molten particles were oxidized during the flying and deposition processes. Fig. 6(b–f) shows the elemental distribution obtained by EDS analysis. The white area in Fig. 6(a) consists essentially of W and C elements, while Ni, Cr, and Si elements are mainly distributed in different areas. Therefore, the white area was mainly enriched with the WC/W2C phases. Moreover, even though the C element is considered a lightweight element, it can be still detected by EDS, as illustrated by the red scattered points in Fig. 6(c). The reason may be attributed to the precipitated carbon and the distribution of the WC particles.
Fig. 6. Energy dispersive spectroscopy (EDS) showing the element distribution of the NiCrBSi/WC–Ni composite coating.
Please cite this article as: Z.Q. Zhang, et al., Surf. Coat. Technol. (2014), http://dx.doi.org/10.1016/j.surfcoat.2014.11.061
Z.Q. Zhang et al. / Surface & Coatings Technology xxx (2014) xxx–xxx
5
Fig. 9. High-magnification cross-sectional scanning electron microscopy (SEM) micrograph of the NiCrBSi/WC–Ni composite coating. Fig. 7. Energy dispersive spectroscopy (EDS) element analysis of the NiCrBSi/WC–Ni composite coating.
In general, in plasma spraying coating, micro-defects can be classified as pores and cracks. Although the majority of the pores typically presented circular or elliptical cavities and had small dimensions, some irregular pores were identified. The pores were mainly distributed among the sprayed particles; few existed in the intragranular areas. Presumably, the pores were formed as a result of the uneven cooling and the shrinking of the sprayed particles during the spraying process. In addition, cracks, which are often distributed in the interlaminar structure of the coating, presented a long gap. However, some intergranular cracks can be observed among deposited particles. Fig. 8 shows a crosssectional SEM micrograph of the NiCrBSi/WC–Ni composite. The typical lamellar structure can be clearly observed. Sprayed particles were fully flattened and laminated. In addition, the results indicated that the coatings exhibited excellent adherence with each other, showing a dense structure with low porosity. A large number of small pores and a few interlaminar cracks were clearly observed in the coatings. The white W2C/WC phases were homogeneously distributed within the coating. Transverse interlaminar cracks were mainly distributed within a range of 200 μm from the coating surface. The particles deposited within a range of 200 μm from the coating surface could bear high tensile stress. Presumably, this was due to the uneven contraction during the cooling step. The tensile stress could generate transverse interlaminar cracks. We believe that this adverse effect may hinder the heat dissipation because of the increase of the coating thickness. However, there were few visible cracks in the interface between the NiAl undercoating and the substrate. Owing to the heat generated by the reaction between melting Ni and Al metal during the plasma spraying, the substrate material
adjacent to the interface between the NiAl undercoating and the substrate could melt and form metallurgical bonding. As a result, using an undercoating would greatly improve the coating adhesive strength. Moreover, the interface between the NiCrBSi/WC–Ni composite coating and the NiAl undercoating was quite flat and did not present any cracks. Fig. 9 shows a high-magnification cross-sectional SEM micrograph of the NiCrBSi/WC–Ni composite coating. The results showed that the W2C/WC phases were well wetted by the other phases. Moreover, a small amount of intragranular pores and cracks, generally due to the uneven cooling and shrinking of the sprayed particles during the spraying process, can be clearly observed. To improve the performance and reduce the amount of the micro-defects of the plasma spraying coating, a re-melting step by post-heat treatment is often adopted to eliminate the micro-defects [26]. Fig. 10 shows a cross-sectional SEM micrograph of the NiCrBSi coating. The coating presented a typical lamellar structure with several small pores and cracks. In addition, the interface between the NiCrBSi coating and NiAl undercoating was also quite flat compared with that of the NiCrBSi/WC–Ni composite coating. Similar to the case of the NiCrBSi/WC–Ni composite coating, transverse interlaminar cracks were mainly distributed within a range of 200 μm from the surface of the coating. Transverse interlaminar cracks are likely to grow under alternating shear stress, finally leading to the RCF failure. Therefore, the layers with transverse interlaminar cracks should be removed during the grinding and polishing process. Fig. 11 shows a surface SEM image of the microstructure of the NiCrBSi/WC–Ni composite coating after being corroded by aqua regia.
Fig. 8. Cross-sectional scanning electron microscopy (SEM) micrograph of the NiCrBSi/ WC–Ni composite coating.
Fig. 10. Cross-sectional scanning electron microscopy (SEM) micrograph of the NiCrBSi coating.
Please cite this article as: Z.Q. Zhang, et al., Surf. Coat. Technol. (2014), http://dx.doi.org/10.1016/j.surfcoat.2014.11.061
6
Z.Q. Zhang et al. / Surface & Coatings Technology xxx (2014) xxx–xxx
Fig. 11. Microstructure of the NiCrBSi/WC–Ni composite coating after corrosion.
The interface between the WC–Ni and NiCrBSi particles, indicated by the dashed line, can be clearly seen. However, as microcracks were not present in the interface, we can conclude that the melting WC–Ni particles and NiCrBSi particles have a good reciprocal wettability. In addition, the micrograph clearly shows that the W2C/WC phase in the WC–Ni particles was irregularly granular, as illustrated by the top left rectangle in Fig. 11. To prevent oxidation and decarburization, WC was coated by Ni metal in the WC–Ni particle. The reason for the presence of the Ni metal-clad structure in WC is that the W2C/WC phase can be easily wetted and infiltrated by melting NiCrBSi particles. The γ-Ni solid solution in the NiCrBSi particles presented a continuous dimple structure, as illustrated by the top right rectangle in Fig. 11. The XRD spectrum (Cu-Kα radiation) performed on the polished surface of the NiCrBSi/WC–Ni composite coating is shown in Fig. 12. The XRD analysis shows a considerable amount of possible phases due to the complexity of the NiCrBSi/WC–Ni composite coating after plasma spraying. The XRD pattern revealed that the coating consists mainly of γ-Ni, Cr7C3, Cr23C6, FeNi3, Ni3Si, CrB, and NiC in addition to the representative peaks of tungsten carbide WC and W2C. Si and B elements mainly dissolve into the γ-Ni phase with face-centered cubic structure. During the plasma spraying process, excessive heating by hightemperature plasma arc and molten particles resulted in the dissolution of the WC particles and the precipitation of carbon as graphite: 2WC → W2C + C. This reaction occurs because of the kinetically driven decarburization process. The precipitated C possibly reacts with Cr and Ni to form carbides [9]. The peak of the diffuse scattering at 2θ = 45°, which is a typical XRD feature of an amorphous material, can be clearly
Fig. 12. X-ray diffraction (XRD) spectrum of the NiCrBSi/WC–Ni composite coating.
observed. The reason is that the hyperthermal sprayed particles were rapidly deposited and cooled on the substrate. The coating porosity is defined as the percentage of micro-defects in the coating. It is an important parameter for the characterization of the coating density. A comparison of the micrographs of the NiCrBSi/ WC–Ni composite coating before and after the gray method processing is shown in Fig. 13. Micro-defects, such as pores and cracks, were highlighted by the gray method processing, as shown in Fig. 13(b). The average porosity of the NiCrBSi/WC–Ni composite coating was estimated as 1.84% by measurements and calculations. In addition, the average porosity of the NiCrBSi coating was 1.64%. Therefore, the density of NiCrBSi coating is slightly higher than that of the NiCrBSi/WC–Ni composite coating. The addition of the WC–Ni melting particles further enhances the uneven contraction, leading to an increase in the number of pores and cracks. However, the two coatings exhibited a dense structure with low porosity, which is beneficial to improve the corrosion resistance, wear resistance, and RCF resistance. Although completely removing the pores and cracks from a coating is very difficult, the porosity can be reduced by optimizing the plasma spraying parameters. Fig. 14 shows the microhardness of the NiCrBSi/WC–Ni composite coating and NiCrBSi coating. The microhardness values were distributed with a gradient form. The microhardness of the NiCrBSi/WC–Ni composite coating exceeded 800 HV, while the microhardness of the NiCrBSi coating was in the range of 700–750 HV. Therefore, the microhardness of the NiCrBSi/WC–Ni composite coating was approximately 100 HV higher than that of the NiCrBSi coating because of the addition of hard WC–Ni. In general, high hardness is useful to improve the surface abrasion resistance. However, the microhardness values of the NiAl undercoating and substrate were below 400 HV. A relatively high hardness on the surface and a relatively low hardness within the coating are beneficial, as a high surface hardness can provide a significant wear resistance, and a soft ductile layer can absorb energy from shocks to prevent crack initiation and growth. 3.2. RCF failure mode and life Ten RCF tests for the NiCrBSi/WC–Ni composite coating were performed under the maximum contact stress of 0.881 GPa to obtain statistically significant results. As a comparison, ten RCF tests for the NiCrBSi coating were conducted under identical experimental conditions. The results of the RCF life and failure mode tests are shown in Fig. 15. Based on the failed surface morphology, the NiCrBSi/WC–Ni composite coating and NiCrBSi coating exhibited two distinct failure modes: abrasion and spalling. For the NiCrBSi/WC–Ni composite coating, abrasion was the main failure mode, as seven out of ten samples failed in abrasion, whereas the remaining three failed in spalling. Conversely, the main failure mode for the NiCrBSi coating was spalling, as seven out of ten samples failed in spalling and the others failed in abrasion. Therefore, the damage in the NiCrBSi coating was more serious than that in the NiCrBSi/WC–Ni composite coating. As in relation to the morphologies of abrasion and spalling failure of the two types of coating are similar, the failure characteristics were analyzed only by considering the NiCrBSi/WC–Ni composite coating. The abrasion is generally defined as a removal of surface material leading to a large number of micro-pits. Although the RCF tests were performed under oil lubrication, there were also some sharp asperities on the coating surface that probably generated local plastic deformations and fractures, forming wear debris. In addition, extremely hard WC particles could also fall off to form abrasive debris. The wear debris would accelerate the abrasion failure. Fig. 16(a) shows the typical abrasion morphology of the failed NiCrBSi/WC–Ni composite coating. Several pits with small dimension and shallow depth were distributed in the overall contact damage zone. Fig. 16(b) shows a high-magnification image of the morphology of the pit. The pit presented a semicircular shape with a lateral size of 40–60 μm, which is slightly smaller than the size of the spraying powder. This may result from the phase transformation during the melting
Please cite this article as: Z.Q. Zhang, et al., Surf. Coat. Technol. (2014), http://dx.doi.org/10.1016/j.surfcoat.2014.11.061
Z.Q. Zhang et al. / Surface & Coatings Technology xxx (2014) xxx–xxx
7
Fig. 13. Compared scanning electron microscopy (SEM) micrographs of the NiCrBSi/WC–Ni composite coating: (a) before gray processing; and (b) after gray processing.
and solidification in the plasma spraying process. That is to say, the spraying particles were peeled off to form the pits. The spalling failure is generally considered as a removal of shallow material from the surface of the coating. Fig. 17(a) shows a typical spalling morphology of the failed NiCrBSi/WC–Ni composite coating. The spalling with an irregular shape appeared in the rolling contact region. In addition, the maximum depth of the spalling can be estimated to be approximately 100 μm. The area of the spalling is obviously larger than that of the abrasion. Fig. 17(b) shows a high-magnification image of the morphology at the bottom of the spalling. The figure shows that the fracture plane at the bottom of spalling was quite rough, and crack growth marks can be observed, as highlighted by the dotted line in Fig. 17(b). We believe that the spalling was related to crack growth within the coating. Cracks, especially transverse interlaminar cracks in the subsurface, existed in adjacent lamellae and through the lamellae in the coating. In addition, a large number of pores with small dimensions generally appeared in the coating. The subsurface cracks may initiate and grow under alternating stresses because of the existence of such micro-defects leading to stress concentrations. Therefore, the intra-lamellar and inter-lamellar cracks in the coating may play an important role in the initiation of unstable growth cracks, as these cracks may grow and join with each other or with the newly initiated cracks. The joined cracks may then grow toward the contact surface, resulting in a final spalling failure. In addition, the oil lubrication may produce instantaneous shocks under alternating contact stress. Thus, when the crack opened or the oil lubrication was
introduced in the crack, the crack growth would accelerate and finally lead to the spalling failure. Generally, the higher the surface microhardness of the coating, the better its abrasion resistance. In addition, extremely hard WC particles could fall off to form abrasive debris. The wear debris could accelerate the abrasion failure. Moreover, the spalling was related to microdefects within the coating, such as cracks and pores. Although the density of the NiCrBSi coating is slightly higher than that of the NiCrBSi/ WC–Ni composite coating, the hard WC particles could prevent the crack growth. Therefore, abrasion was the main failure mode for the NiCrBSi/WC–Ni composite coating, whereas spalling was the main failure mode for the NiCrBSi coating. Fig. 15 shows that the RCF lives of the two coatings were quite discrete. Kang et al. also reached the same conclusion for the RCF life of the AT40 ceramic coating using this RCF test rig [25]. There was an overlap zone between the minimum life of the NiCrBSi/WC–Ni composite coating and the maximum life of the NiCrBSi coating. Thus, evaluating the RCF life only by the average life, maximum life, or minimum life does not provide a reliable estimate. The Weibull distribution is a type of probabilistic method widely used to process scattering data of the fatigue life in reliability analyses and calculations. Based on the number of the Weibull parameters, there are three different types of distributions: one-parameter, two-parameter, and three-parameter Weibull distribution. Typically, it is difficult to process statistical data accurately by one-parameter Weibull distribution, and, although a deeper analysis
Fig. 14. Microhardness of the NiCrBSi/WC–Ni composite coating and NiCrBSi coating.
Fig. 15. Results of the RCF life and failure mode of the NiCrBSi/WC–Ni composite coating and NiCrBSi coating.
Please cite this article as: Z.Q. Zhang, et al., Surf. Coat. Technol. (2014), http://dx.doi.org/10.1016/j.surfcoat.2014.11.061
8
Z.Q. Zhang et al. / Surface & Coatings Technology xxx (2014) xxx–xxx
Fig. 16. Abrasion morphology of the failed NiCrBSi/WC–Ni composite coating: (a) overall observation; and (b) high-magnification observation.
Fig. 17. Spalling morphology of the failed NiCrBSi/WC–Ni composite coating: (a) overall observation; and (b) high-magnification observation.
can be performed using the three-parameter Weibull distribution, this is not effectively applied because of the tedious data processing. Therefore, the two-parameter Weibull distribution is frequently employed to process the fatigue life:
NiCrBSi/WC–Ni composite coating was similar to that of the AT40 ceramic coating. Fig. 18 shows the Weibull distribution plots of the RCF life of the NiCrBSi/WC–Ni composite coating and NiCrBSi coating.
" # N b F ðN Þ ¼ 1− exp − Na where F(N) is the failure probability and N are the cycles of the coating specimen. Na and b are two parameters of the Weibull distribution. Na represents the characteristic life when the failure probability is 63.2%, and b stands for the Weibull plot slope, which can indicate the discrete degree of the RCF life. In general, the RCF life becomes more discrete as the Weibull plot slope b decreases. The values of Na and b can be evaluated using the theory of maximum likelihood estimation (MLE) based on the obtained life. By applying the MLE method on the RCF life of the NiCrBSi/WC–Ni composite coating and NiCrBSi coating with a 90% confidence level, the values of Na were estimated to be 2.1649 and 1.6041, respectively, and the values of b were estimated to be 6.0805 and 5.3149, respectively. The RCF Life of the NiCrBSi/WC–Ni composite coating had a higher dispersion degree than that of the NiCrBSi coating under a contact stress of 0.881 GPa. Additionally, the Weibull plot slope b of the RCF life of the AT40 ceramic coating exhibited a value of 6.2849 under a contact stress of 0.75 GPa [25]. Therefore, the discrete degree of the RCF life of the
Fig. 18. Weibull distribution plot of the RCF life of the NiCrBSi/WC–Ni composite coating and NiCrBSi coating.
Please cite this article as: Z.Q. Zhang, et al., Surf. Coat. Technol. (2014), http://dx.doi.org/10.1016/j.surfcoat.2014.11.061
Z.Q. Zhang et al. / Surface & Coatings Technology xxx (2014) xxx–xxx
The figure depicts the failure percentage of the specimens (ordinate) versus the number of stress cycles for RCF failure (abscissa). The Weibull plot can indicate the statistical percentage of the failed coating in the RCF tests and the RCF life when designating the failure percentage. In Fig. 18, the confidence limits of the NiCrBSi/WC–Ni composite coating and NiCrBSi coating are shown by dotted and dashed lines, respectively. The estimate of the RCF life is reliable because the data points fall within the confidence limits with a confidence level of 90%. Owing to the addition of WC–Ni, the RCF life of the NiCrBSi/WC–Ni composite coating was longer than that of the NiCrBSi coating when designating the failure percentage under this contact level. In addition, The RCF life of the NiCrBSi/ WC–Ni composite coating was easier to predict than that of the NiCrBSi coating, as the increase of the Weibull plot slope b indicates. Therefore, the RCF performance of the NiCrBSi/WC–Ni composite coating was superior to that of the NiCrBSi coating under a contact stress of 0.881 GPa. 4. Conclusions This paper systematically investigated the microstructure, element distribution, phase composition, porosity, microhardness, and RCF performance of NiCrBSi/WC–Ni composite coatings. On the basis of the above results and discussion, it can be concluded that: (1) The NiCrBSi/WC–Ni composite coating presented typical lamellar structures. Sprayed particles were fully flattened and laminated. White WC–Ni particles were homogeneously distributed in the NiCrBSi/WC–Ni composite coating, and W2C/WC phases could be well wetted by other phases. Moreover, shrinkage holes were formed due to the uneven contraction of the melting particles during the cooling. Transverse interlaminar cracks were distributed within a range of 200 μm from the coating surface. The NiCrBSi/ WC–Ni composite coating consists mainly of γ-Ni, Cr7C3, Cr23C6, FeNi3, Ni3Si, CrB, and NiC, in addition to WC and W2C. However, microscale oxygen inevitably exists in the coatings. The NiCrBSi/ WC–Ni composite coating (porosity of 1.84%) had slightly lower density than the NiCrBSi coating (porosity of 1.62%). Furthermore, as a result of the addition of hard WC–Ni, the microhardness of the NiCrBSi/WC–Ni composite coating exceeded 800 HV, which is significantly higher than that shown by the NiCrBSi coating. (2) The NiCrBSi/WC–Ni composite coating and NiCrBSi coating exhibited two distinct failure modes, abrasion and spalling, based on failed surface morphology under contact stress of 0.881 GPa. Abrasion was the main failure mode for the NiCrBSi/WC–Ni composite coating, whereas the main failure mode for the NiCrBSi coating was spalling. That is to say, the damage of the NiCrBSi coating was more serious than that of the NiCrBSi/WC–Ni composite coating.
9
(3) The NiCrBSi/WC–Ni composite coating had a higher dispersion degree of the RCF life and longer life than the NiCrBSi coating under a contact stress of 0.881 GPa. Therefore, the RCF performance of the NiCrBSi/WC–Ni composite coating was superior to that of the NiCrBSi coating.
Acknowledgments The paper was financially supported by 973 Project (2011CB013405), Distinguished Young Scholars of NSFC (51125023), and “Twelfth Five-Year” National Science and Technology Support Program (2011BAF11B08). References [1] L. Wang, J.C. Fang, Z.Y. Zhao, H.P. Zeng, Surf. Coat. Technol. 201 (2007) 5085–5089. [2] H.L. Yu, W. Zhang, H.M. Wang, Tribol. Int. 66 (2013) 105–113. [3] Y.L. An, G.L. Hou, J. Chen, X.Q. Zhao, G. Liu, H.D. Zhou, J.M. Chen, Vacuum 107 (2014) 132–140. [4] J. Chen, Y.L. An, J. Yang, X.Q. Zhao, F.Y. Yan, H.D. Zhou, J.M. Chen, Surf. Coat. Technol. 235 (2013) 521–528. [5] R.G. Song, C. Wang, Y. Jiang, H. Li, G. Lu, Z.X. Wang, J. Alloys Compd. 544 (2012) 13–18. [6] A. Cattini, D. Bellucci, A. Sola, L. Pawłowski, V. Cannillo, Surf. Coat. Technol. 236 (2013) 118–126. [7] Y.R. Niu, X.B. Zheng, H. Ji, L.J. Qi, C.X. Ding, J.L. Chen, G.N. Luo, Fusion Eng. Des. 85 (2013) 1521–1526. [8] R. Mušálek, O. Kovářík, T. Skiba, P. Haušild, M. Karlík, J. Colmenares-Angulo, Intermetallics 18 (2010) 1415–1418. [9] P. Niranatlumpong, H. Koiprasert, Surf. Coat. Technol. 206 (2011) 440–445. [10] T. Liyanage, G. Fisher, A.P. Gerlich, Surf. Coat. Technol. 205 (2010) 759–765. [11] N. Serres, F. Hlawka, S. Costil, C. Langlade, F. Machi, Surf. Coat. Technol. 205 (2010) 1039–1046. [12] L.G. Yu, K.A. Khor, H. Li, K.C. Pay, T.H. Yip, P. Cheang, Surf. Coat. Technol. 182 (2004) 308–317. [13] C.J. Li, G.J. Yang, Int. J. Refract. Met. Hard Mater. 39 (2013) 2–17. [14] Q. Zhan, L.G. Yu, F.X. Ye, X.J. Xue, H. Li, Surf. Coat. Technol. 206 (2012) 4068–4074. [15] G.Z. Xie, J.X. Zhang, Y.J. Lu, K.Y. Wang, X.Y. Mo, P.H. Lin, Mater. Sci. Eng. A 460–461 (2007) 351–356. [16] P. Suresh Babu, Bikramjit Basu, G. Sundararajan, Wear 268 (2010) 1387–1399. [17] C. Guo, J.S. Zhou, J.M. Chen, J.R. Zhao, Y.J. Yu, H.D. Zhou, Wear 270 (2011) 492–498. [18] M.J. Tobar, C. Álvarez, J.M. Amado, G. Rodríguez, A. Yáñez, Surf. Coat. Technol. 200 (2006) 6313–6317. [19] Anup S. Pandkar, Nagaraj Arakere, Ghatu Subhash, Int. J. Fatigue 63 (2014) 191–202. [20] H. Matsunaga, H. Komata, J. Yamabe, Y. Fukushima, S. Matsuoka, Procedia Mater. Sci. 3 (2014) 1663–1668. [21] G.L. Li, Z.Q. Zhang, H.D. Wang, B.S. Xu, Z.Y. Piao, L.N. Zhu, Tribol. Int. 61 (2013) 129–137. [22] J.J. Kang, B.S. Xu, H.D. Wang, C.B. Wang, Tribol. Int. 73 (2014) 128–137. [23] X.C. Zhang, B.S. Xu, F.Z. Xuan, H.D. Wang, Y.X. Wu, Appl. Surf. Sci. 254 (2008) 3734–3744. [24] S. Stewart, R. Ahmed, Wear 253 (2002) 1132–1144. [25] J.J. Kang, B.S. Xu, H.D. Wang, C.B. Wang, Tribol. Int. 66 (2013) 249–258. [26] S. Stewart, R. Ahmed, T. Itsukaichi, Surf. Coat. Technol. 190 (2005) 171–189.
Please cite this article as: Z.Q. Zhang, et al., Surf. Coat. Technol. (2014), http://dx.doi.org/10.1016/j.surfcoat.2014.11.061
Contents lists available at ScienceDirect
Surface & Coatings Technology journal homepage: www.elsevier.com/locate/surfcoat
Characterization of microstructure and rolling contact fatigue performance of NiCrBSi/WC–Ni composite coatings prepared by plasma spraying Z.Q. Zhang a, H.D. Wang b,⁎, B.S. Xu b, G.S. Zhang c a b c
School of Materials Science and Engineering, Tianjin University, Tianjin 300072, China National Key Lab for Remanufacturing, Academy of Armored Forces Engineering, Beijing 100072, China Institute of Remanufacture and Processing Material, Tianjin Research Institute of Construction Machinery, Tianjin 300409, China
a r t i c l e
i n f o
Article history: Received 6 September 2014 Accepted in revised form 24 November 2014 Available online xxxx Keywords: NiCrBSi/WC–Ni composite coating Microstructure Rolling contact fatigue Abrasion Spalling
a b s t r a c t In this paper, the properties of NiCrBSi/WC–Ni composite coatings deposited by plasma spraying were analyzed, including their microstructure, element distribution, phase composition, porosity, and microhardness. The rolling contact fatigue (RCF) life and failure modes were also investigated. The results showed that the NiCrBSi/WC–Ni composite coating exhibited a typical lamellar structure. Furthermore, evenly distributed W2C/WC phases were well wetted by other phases. In addition to WC and W2C phases, the NiCrBSi/WC–Ni composite coating consisted mainly of γ-Ni, Cr7C3, Cr23C6, FeNi3, Ni3Si, CrB, and NiC. The study also found that the NiCrBSi/WC–Ni composite coating (porosity of 1.84%) had a slightly lower density than the NiCrBSi coating (porosity of 1.62%). The microhardness of the NiCrBSi/WC–Ni composite coating was significantly enhanced by the addition of hard WC–Ni. Furthermore, the results indicated that the NiCrBSi/WC–Ni composite coating had a higher dispersion degree of the RCF life and a longer life than the NiCrBSi coating under 0.881 GPa. Therefore, under this stress level, the RCF performance of the NiCrBSi/WC–Ni composite coating was superior to that of the NiCrBSi coating. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Plasma spraying is an effective surface engineering technology that has been widely applied in various industrial fields [1–5]. Plasma spraying coating can improve the contact fatigue resistance, wear resistance, and corrosion resistance in metal components such as rollers in paper manufacturing, shafts and screws in petrochemical processes, and transmission shafts in construction machinery and cars. In addition, the plasma spraying technology can be applied to a wide range of materials, such as ceramics, pure metals or alloys, and polymeric materials [6–8]. Previous studies have showed that self-fluxing NiCrBSi alloys exhibit a low melting point, good deoxidizing and slagging properties, and great wettability. Based on their excellent processing properties, they are also particularly suited for the plasma spraying technology [9–11]. Theoretically, B is included in the NiCrBSi alloy to shift the alloy closer to the Ni–Ni3B eutectic composition and reduce the melting temperature of the alloy. In addition, B can form hard phases with Ni, such as Ni3B, thus improving the hardness of the coating. The addition of Si enhances the self-fluxing properties of the NiCrBSi alloy. Unless a large amount is added, Si will remain in solid solution with Ni. Si also slightly increases the hardness of the alloy. Furthermore, Cr can be added to the alloy to improve its corrosion and wear resistance. The ⁎ Corresponding author. Tel.: +86 10 66718475; fax: +86 10 66719325. E-mail address: [email protected] (H.D. Wang).
amount of Cr determines the structure of the alloy. If the content of Cr and C is sufficient, hard carbides such as Cr3C2, Cr7C3, and Cr23C6 can be formed. As the coating formation relies on the rapid solidification of the materials during the plasma spraying process, phase transformations are undoubtedly hindered to some degree, resulting in a unique microstructure. The addition of hard ceramic particles, such as tungsten carbide (WC), in a self-fluxing Ni-based alloy may increase the coating hardness, abrasive wear resistance, and contact fatigue resistance. Compared to other carbides, WC combines favorable properties such as high hardness, plasticity, and a good wettability by molten metals [9]. Therefore, WC is widely used as the hard phase in manufacturing metal composites. Moreover, previous studies have showed that WC can decarburize to W2C and result in a relatively soft W phase. However, WC-coated Ni can prevent and reduce the decarburization during the plasma spraying process [12–16]. Some studies have been performed on the effects of the addition of WC on the performance of the NiCrBSi alloys. The microstructure of a NiCrBSi/WC composite coating prepared by electric arc spraying technology consisted of NiCr, NiCrW, and WC/W2C as major phases [9]. In addition, NiCrBSi coatings and NiCrBSi/WC–Ni composite coatings were prepared on stainless steel by laser cladding technology. The results showed that the NiCrBSi/WC–Ni composite coating had a higher wear resistance at high temperature than the NiCrBSi coating, and the difference was attributed to the formation of a hard WC phase after laser cladding [17]. Several types of mixtures of self-fluxing NiCrBSi
http://dx.doi.org/10.1016/j.surfcoat.2014.11.061 0257-8972/© 2014 Elsevier B.V. All rights reserved.
Please cite this article as: Z.Q. Zhang, et al., Surf. Coat. Technol. (2014), http://dx.doi.org/10.1016/j.surfcoat.2014.11.061
2
Z.Q. Zhang et al. / Surface & Coatings Technology xxx (2014) xxx–xxx
alloy powder and WC–Ni powder (10 wt.% Ni and balance WC) were laser cladded on stainless steel substrates. The study found that most of the clad layer properties, such as its porosity and microhardness, were affected by the percentage of the WC particles present in the mixture. Dense and pore-free layers could be obtained as long as the WC content in the mixture was kept below 50 wt.% [18]. However, the properties of the plasma-sprayed NiCrBSi/WC–Ni composite coatings, such as the microstructure, element distribution, phase composition, porosity, and microhardness, have not been systematically studied yet. Moreover, since the rolling contact fatigue (RCF) failure of the majority of the components originates from the surface or subsurface [19,20], plasma spraying can play an important role in repairing the surfaces of failed parts, thus prolonging their lives [21–23]. Interestingly, the RCF resistance of the NiCrBSi/WC–Ni composite coatings has not been investigated yet. The RCF can be defined as a damage process on the contact surface of the rotating components bearing alternating stress [24]. In this work, we focus on the preparation of a protective NiCrBSi/WC–Ni composite coating on tempered 1045 steel by plasma spraying. The properties of the NiCrBSi/WC–Ni composite coating, including the microstructure, element distribution, phase composition, porosity, and microhardness were analyzed. In addition, the RCF life performance and failure mode were also examined. A NiCrBSi coating was prepared to provide a comparison with the NiCrBSi/WC–Ni composite coating.
2. Experimental procedures 2.1. Preparation of the coatings A GP-80 plasma spraying system manufactured by Taixing Spraying Corporation of China was used to prepare the NiCrBSi/WC–Ni composite coating and NiAl undercoating. Hydrogen, argon, and nitrogen were used as working gas, protective gas, and feeding gas, respectively. Commercial tempered 1045 steel was used as a substrate. Fig. 1 shows two-dimensional and three-dimensional diagrams of both the coated and RCF-tested specimens. The size of the specimens is shown in Fig. 1(a). The length of the line contact in the RCF test is 6 mm. The red surface area shown in Fig. 1(b) is planed to prepare for the coating. Fig. 2 shows NiCrBSi/WC–Ni composite powders with 15 wt.% WC–Ni as sprayed material. The composition of the NiCrBSi alloy powder is Cr 16, B 7.5, Si 4.7, C 0.9, Fe ≤ 4.5, Ni balance (wt.%), with particles of 40–80 μm size and globular shape, as indicated by the arrow in Fig. 2. The composition of WC–Ni is Ni 5, WC balance (wt.%), with particles of 40–100 μm size and irregular shape, as indicated by the arrow in Fig. 2. A NiAl alloy powder with a composition of Ni 80, Al 20 (wt.%) and particle size of
Fig. 2. Morphology of the NiCrBSi/WC–Ni composite powders.
50–100 μm was used as the undercoating material. The bond strength between the coating and the substrate can be significantly increased as a result of the heat generated by the reaction between the melting nickel and aluminum arriving on the substrate. The optimized plasma spraying parameters are listed in Table 1. Fig. 3 illustrates the preparation process of plasma spraying coating. The thicknesses of the NiCrBSi/WC–Ni composite coating and NiAl undercoating were adjusted in the range of 500–600 μm and 200–250 μm, respectively. In addition, a NiCrBSi coating was also prepared to provide a comparison with the NiCrBSi/WC–Ni composite coating. The composition of the NiCrBSi powder was identical to that of NiCrBSi in the NiCrBSi/WC–Ni composite powder. The spraying parameters and the thickness were the same as those of the NiCrBSi/WC–Ni composite.
2.2. Characterization of the coatings The microstructure and RCF failure morphology were observed by scanning electron microscopy (SEM, Philips Quant 200), while the composition and distribution of the elements were analyzed by energy dispersive spectroscopy (EDS, Philips Quant 200). The phase composition was identified using a D8 X-ray diffractometer (40 kV, 30 mA, Cu Kα radiation, λKα = 0.154 nm, 2θ scanning step of 0.03°, scanning in the range of 30° b 2θ b 100°). The microhardness distribution along the depth direction of the coating was determined using a HV-1000A Vickers microhardness tester with a load of 0.98 N and dwell time of 5 s. For each hardness profile along the depth direction of the coating, three tests were performed (the spacing of each press mark was less than 50 μm) and the averaged results of the three repeated tests were used in this article. The porosity of the coating was measured by image processing software based on the gray analysis method. Ten cross-sectional micrographs of the coating at 500 × magnification
Table 1 Plasma spraying parameters.
Fig. 1. Diagram of the coated and RCF tested specimens.
Processing parameters
NiCrBSi/WC–Ni
NiAl
Spraying voltage (V) Spraying current (A) Argon gas flow (L/h) Hydrogen gas flow (L/h) Nitrogen gas flow (L/min) Spraying distance (mm) Coating thickness (μm)
60 500 30 130 6 90–110 500–600
56 500 30 140 5 90–110 200–250
Please cite this article as: Z.Q. Zhang, et al., Surf. Coat. Technol. (2014), http://dx.doi.org/10.1016/j.surfcoat.2014.11.061
Z.Q. Zhang et al. / Surface & Coatings Technology xxx (2014) xxx–xxx
3
Fig. 3. Schematic of the preparation process of the plasma spraying coating.
were randomly collected by SEM, and an average value was calculated from the ten different porosity test results. 2.3. The RCF test As shown in Fig. 4, a double-roller RCF test rig manufactured by National Key Lab for Remanufacturing of China was used to evaluate the RCF performance of the NiCrBSi/WC–Ni composite coating and NiCrBSi coating [25]. A tested roller with the coating and a paired contact roller were fixed on the rotation axis by a clamp and driven by two servo motors. The type of rolling contact is a line contact. The paired contact roller was machined using AISI 52100 steel with a surface roughness of 0.012 μm and a Rockwell hardness of 62 HRC. In addition, the thickness, external diameter, and internal diameter of the paired contact roller were kept to 20 mm, 60 mm, and 32 mm, respectively. The rotation speed can be stepless adjusted in the range of 0–2000 rpm by setting up the software parameters, and the slip ratio can be set in the range of 0–100%. To assure a pure rolling contact form in this experiment, the rotation speeds of both the tested roller and the paired contact roller
were both kept at a constant value of 600 rpm, namely, with a slip ratio of 0%. All the RCF tests were conducted under oil lubrication. The load was applied by hydraulic pressure and, here, set to a constant value of 2000 N. The maximum contact stress, σmax, was calculated using Hertz's equation, which is expressed as:
σ max
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi PE R1 þ R2 ¼ 0:418 l R1 R2
where P (with a unit of N) is the magnitude of the applied load at the line between the tested roller and the paired contact roller, R1 = R2 = 30 mm are the radii of the tested and paired contact roller, respectively, l = 6 mm is the length of the line contact, and E denotes the elastic modulus. The elastic moduli of the tested and paired contact roller are both set to 200 GPa. Thus, the maximum calculated contact stress is 0.881 GPa under a load of 2000 N. Vibration and AE sensors were used to monitor the RCF failure process. In this study, when the vibration and AE signals exceeded the pre-setting threshold and sharply increased as a result of a failure, the test was stopped and the fatigue life of the specimen calculated. As the experimental data from the RCF tests are considerably scattered, ten rolling contact tests were performed in the same conditions to obtain statistically significant results of the RCF life. A new pair of contact rollers was used for each test. 3. Results and discussion 3.1. Characterization of the coatings
Fig. 4. Schematic of the double-roller RCF test rig.
As the surfaces of the NiCrBSi/WC–Ni composite coating and NiCrBSi coating were severely roughened after the plasma spraying process, the specimens were initially ground and polished. The final thickness of the coating was approximately 400–500 μm (including the working coating and undercoating). The surface roughness of the polished coating was measured by a Talysurf 5P-120 surface contour mapping system. The value of the roughness average (Ra) after five measurements was 0.276 μm. The melting NiCrBSi/WC–Ni composite powders, after being heated and accelerated by the plasma arc, were deposited on the substrate. The spraying particles were firmly bonded by the mechanical interlock structure. Shrinkage holes were formed because of an uneven contraction of the melting particles during the cooling process, as
Please cite this article as: Z.Q. Zhang, et al., Surf. Coat. Technol. (2014), http://dx.doi.org/10.1016/j.surfcoat.2014.11.061
4
Z.Q. Zhang et al. / Surface & Coatings Technology xxx (2014) xxx–xxx
Fig. 5. Scanning electron microscopy (SEM) surface micrograph of the NiCrBSi/WC–Ni composite coating.
indicated by the arrows in Fig. 5. The holes could not be completely filled by the subsequent molten particles. A surface SEM micrograph of the NiCrBSi/WC–Ni composite coating after plasma spraying is shown in Fig. 6(a). The area in Fig. 6(a) was scanned and analyzed by EDS. The results show that the NiCrBSi/WC– Ni composite coating mainly comprises W, C, Ni, Cr, Fe, and Si elements, in addition to B, as shown in Fig. 7. However, microscale oxygen element was detected after plasma spraying, mainly because the molten particles were oxidized during the flying and deposition processes. Fig. 6(b–f) shows the elemental distribution obtained by EDS analysis. The white area in Fig. 6(a) consists essentially of W and C elements, while Ni, Cr, and Si elements are mainly distributed in different areas. Therefore, the white area was mainly enriched with the WC/W2C phases. Moreover, even though the C element is considered a lightweight element, it can be still detected by EDS, as illustrated by the red scattered points in Fig. 6(c). The reason may be attributed to the precipitated carbon and the distribution of the WC particles.
Fig. 6. Energy dispersive spectroscopy (EDS) showing the element distribution of the NiCrBSi/WC–Ni composite coating.
Please cite this article as: Z.Q. Zhang, et al., Surf. Coat. Technol. (2014), http://dx.doi.org/10.1016/j.surfcoat.2014.11.061
Z.Q. Zhang et al. / Surface & Coatings Technology xxx (2014) xxx–xxx
5
Fig. 9. High-magnification cross-sectional scanning electron microscopy (SEM) micrograph of the NiCrBSi/WC–Ni composite coating. Fig. 7. Energy dispersive spectroscopy (EDS) element analysis of the NiCrBSi/WC–Ni composite coating.
In general, in plasma spraying coating, micro-defects can be classified as pores and cracks. Although the majority of the pores typically presented circular or elliptical cavities and had small dimensions, some irregular pores were identified. The pores were mainly distributed among the sprayed particles; few existed in the intragranular areas. Presumably, the pores were formed as a result of the uneven cooling and the shrinking of the sprayed particles during the spraying process. In addition, cracks, which are often distributed in the interlaminar structure of the coating, presented a long gap. However, some intergranular cracks can be observed among deposited particles. Fig. 8 shows a crosssectional SEM micrograph of the NiCrBSi/WC–Ni composite. The typical lamellar structure can be clearly observed. Sprayed particles were fully flattened and laminated. In addition, the results indicated that the coatings exhibited excellent adherence with each other, showing a dense structure with low porosity. A large number of small pores and a few interlaminar cracks were clearly observed in the coatings. The white W2C/WC phases were homogeneously distributed within the coating. Transverse interlaminar cracks were mainly distributed within a range of 200 μm from the coating surface. The particles deposited within a range of 200 μm from the coating surface could bear high tensile stress. Presumably, this was due to the uneven contraction during the cooling step. The tensile stress could generate transverse interlaminar cracks. We believe that this adverse effect may hinder the heat dissipation because of the increase of the coating thickness. However, there were few visible cracks in the interface between the NiAl undercoating and the substrate. Owing to the heat generated by the reaction between melting Ni and Al metal during the plasma spraying, the substrate material
adjacent to the interface between the NiAl undercoating and the substrate could melt and form metallurgical bonding. As a result, using an undercoating would greatly improve the coating adhesive strength. Moreover, the interface between the NiCrBSi/WC–Ni composite coating and the NiAl undercoating was quite flat and did not present any cracks. Fig. 9 shows a high-magnification cross-sectional SEM micrograph of the NiCrBSi/WC–Ni composite coating. The results showed that the W2C/WC phases were well wetted by the other phases. Moreover, a small amount of intragranular pores and cracks, generally due to the uneven cooling and shrinking of the sprayed particles during the spraying process, can be clearly observed. To improve the performance and reduce the amount of the micro-defects of the plasma spraying coating, a re-melting step by post-heat treatment is often adopted to eliminate the micro-defects [26]. Fig. 10 shows a cross-sectional SEM micrograph of the NiCrBSi coating. The coating presented a typical lamellar structure with several small pores and cracks. In addition, the interface between the NiCrBSi coating and NiAl undercoating was also quite flat compared with that of the NiCrBSi/WC–Ni composite coating. Similar to the case of the NiCrBSi/WC–Ni composite coating, transverse interlaminar cracks were mainly distributed within a range of 200 μm from the surface of the coating. Transverse interlaminar cracks are likely to grow under alternating shear stress, finally leading to the RCF failure. Therefore, the layers with transverse interlaminar cracks should be removed during the grinding and polishing process. Fig. 11 shows a surface SEM image of the microstructure of the NiCrBSi/WC–Ni composite coating after being corroded by aqua regia.
Fig. 8. Cross-sectional scanning electron microscopy (SEM) micrograph of the NiCrBSi/ WC–Ni composite coating.
Fig. 10. Cross-sectional scanning electron microscopy (SEM) micrograph of the NiCrBSi coating.
Please cite this article as: Z.Q. Zhang, et al., Surf. Coat. Technol. (2014), http://dx.doi.org/10.1016/j.surfcoat.2014.11.061
6
Z.Q. Zhang et al. / Surface & Coatings Technology xxx (2014) xxx–xxx
Fig. 11. Microstructure of the NiCrBSi/WC–Ni composite coating after corrosion.
The interface between the WC–Ni and NiCrBSi particles, indicated by the dashed line, can be clearly seen. However, as microcracks were not present in the interface, we can conclude that the melting WC–Ni particles and NiCrBSi particles have a good reciprocal wettability. In addition, the micrograph clearly shows that the W2C/WC phase in the WC–Ni particles was irregularly granular, as illustrated by the top left rectangle in Fig. 11. To prevent oxidation and decarburization, WC was coated by Ni metal in the WC–Ni particle. The reason for the presence of the Ni metal-clad structure in WC is that the W2C/WC phase can be easily wetted and infiltrated by melting NiCrBSi particles. The γ-Ni solid solution in the NiCrBSi particles presented a continuous dimple structure, as illustrated by the top right rectangle in Fig. 11. The XRD spectrum (Cu-Kα radiation) performed on the polished surface of the NiCrBSi/WC–Ni composite coating is shown in Fig. 12. The XRD analysis shows a considerable amount of possible phases due to the complexity of the NiCrBSi/WC–Ni composite coating after plasma spraying. The XRD pattern revealed that the coating consists mainly of γ-Ni, Cr7C3, Cr23C6, FeNi3, Ni3Si, CrB, and NiC in addition to the representative peaks of tungsten carbide WC and W2C. Si and B elements mainly dissolve into the γ-Ni phase with face-centered cubic structure. During the plasma spraying process, excessive heating by hightemperature plasma arc and molten particles resulted in the dissolution of the WC particles and the precipitation of carbon as graphite: 2WC → W2C + C. This reaction occurs because of the kinetically driven decarburization process. The precipitated C possibly reacts with Cr and Ni to form carbides [9]. The peak of the diffuse scattering at 2θ = 45°, which is a typical XRD feature of an amorphous material, can be clearly
Fig. 12. X-ray diffraction (XRD) spectrum of the NiCrBSi/WC–Ni composite coating.
observed. The reason is that the hyperthermal sprayed particles were rapidly deposited and cooled on the substrate. The coating porosity is defined as the percentage of micro-defects in the coating. It is an important parameter for the characterization of the coating density. A comparison of the micrographs of the NiCrBSi/ WC–Ni composite coating before and after the gray method processing is shown in Fig. 13. Micro-defects, such as pores and cracks, were highlighted by the gray method processing, as shown in Fig. 13(b). The average porosity of the NiCrBSi/WC–Ni composite coating was estimated as 1.84% by measurements and calculations. In addition, the average porosity of the NiCrBSi coating was 1.64%. Therefore, the density of NiCrBSi coating is slightly higher than that of the NiCrBSi/WC–Ni composite coating. The addition of the WC–Ni melting particles further enhances the uneven contraction, leading to an increase in the number of pores and cracks. However, the two coatings exhibited a dense structure with low porosity, which is beneficial to improve the corrosion resistance, wear resistance, and RCF resistance. Although completely removing the pores and cracks from a coating is very difficult, the porosity can be reduced by optimizing the plasma spraying parameters. Fig. 14 shows the microhardness of the NiCrBSi/WC–Ni composite coating and NiCrBSi coating. The microhardness values were distributed with a gradient form. The microhardness of the NiCrBSi/WC–Ni composite coating exceeded 800 HV, while the microhardness of the NiCrBSi coating was in the range of 700–750 HV. Therefore, the microhardness of the NiCrBSi/WC–Ni composite coating was approximately 100 HV higher than that of the NiCrBSi coating because of the addition of hard WC–Ni. In general, high hardness is useful to improve the surface abrasion resistance. However, the microhardness values of the NiAl undercoating and substrate were below 400 HV. A relatively high hardness on the surface and a relatively low hardness within the coating are beneficial, as a high surface hardness can provide a significant wear resistance, and a soft ductile layer can absorb energy from shocks to prevent crack initiation and growth. 3.2. RCF failure mode and life Ten RCF tests for the NiCrBSi/WC–Ni composite coating were performed under the maximum contact stress of 0.881 GPa to obtain statistically significant results. As a comparison, ten RCF tests for the NiCrBSi coating were conducted under identical experimental conditions. The results of the RCF life and failure mode tests are shown in Fig. 15. Based on the failed surface morphology, the NiCrBSi/WC–Ni composite coating and NiCrBSi coating exhibited two distinct failure modes: abrasion and spalling. For the NiCrBSi/WC–Ni composite coating, abrasion was the main failure mode, as seven out of ten samples failed in abrasion, whereas the remaining three failed in spalling. Conversely, the main failure mode for the NiCrBSi coating was spalling, as seven out of ten samples failed in spalling and the others failed in abrasion. Therefore, the damage in the NiCrBSi coating was more serious than that in the NiCrBSi/WC–Ni composite coating. As in relation to the morphologies of abrasion and spalling failure of the two types of coating are similar, the failure characteristics were analyzed only by considering the NiCrBSi/WC–Ni composite coating. The abrasion is generally defined as a removal of surface material leading to a large number of micro-pits. Although the RCF tests were performed under oil lubrication, there were also some sharp asperities on the coating surface that probably generated local plastic deformations and fractures, forming wear debris. In addition, extremely hard WC particles could also fall off to form abrasive debris. The wear debris would accelerate the abrasion failure. Fig. 16(a) shows the typical abrasion morphology of the failed NiCrBSi/WC–Ni composite coating. Several pits with small dimension and shallow depth were distributed in the overall contact damage zone. Fig. 16(b) shows a high-magnification image of the morphology of the pit. The pit presented a semicircular shape with a lateral size of 40–60 μm, which is slightly smaller than the size of the spraying powder. This may result from the phase transformation during the melting
Please cite this article as: Z.Q. Zhang, et al., Surf. Coat. Technol. (2014), http://dx.doi.org/10.1016/j.surfcoat.2014.11.061
Z.Q. Zhang et al. / Surface & Coatings Technology xxx (2014) xxx–xxx
7
Fig. 13. Compared scanning electron microscopy (SEM) micrographs of the NiCrBSi/WC–Ni composite coating: (a) before gray processing; and (b) after gray processing.
and solidification in the plasma spraying process. That is to say, the spraying particles were peeled off to form the pits. The spalling failure is generally considered as a removal of shallow material from the surface of the coating. Fig. 17(a) shows a typical spalling morphology of the failed NiCrBSi/WC–Ni composite coating. The spalling with an irregular shape appeared in the rolling contact region. In addition, the maximum depth of the spalling can be estimated to be approximately 100 μm. The area of the spalling is obviously larger than that of the abrasion. Fig. 17(b) shows a high-magnification image of the morphology at the bottom of the spalling. The figure shows that the fracture plane at the bottom of spalling was quite rough, and crack growth marks can be observed, as highlighted by the dotted line in Fig. 17(b). We believe that the spalling was related to crack growth within the coating. Cracks, especially transverse interlaminar cracks in the subsurface, existed in adjacent lamellae and through the lamellae in the coating. In addition, a large number of pores with small dimensions generally appeared in the coating. The subsurface cracks may initiate and grow under alternating stresses because of the existence of such micro-defects leading to stress concentrations. Therefore, the intra-lamellar and inter-lamellar cracks in the coating may play an important role in the initiation of unstable growth cracks, as these cracks may grow and join with each other or with the newly initiated cracks. The joined cracks may then grow toward the contact surface, resulting in a final spalling failure. In addition, the oil lubrication may produce instantaneous shocks under alternating contact stress. Thus, when the crack opened or the oil lubrication was
introduced in the crack, the crack growth would accelerate and finally lead to the spalling failure. Generally, the higher the surface microhardness of the coating, the better its abrasion resistance. In addition, extremely hard WC particles could fall off to form abrasive debris. The wear debris could accelerate the abrasion failure. Moreover, the spalling was related to microdefects within the coating, such as cracks and pores. Although the density of the NiCrBSi coating is slightly higher than that of the NiCrBSi/ WC–Ni composite coating, the hard WC particles could prevent the crack growth. Therefore, abrasion was the main failure mode for the NiCrBSi/WC–Ni composite coating, whereas spalling was the main failure mode for the NiCrBSi coating. Fig. 15 shows that the RCF lives of the two coatings were quite discrete. Kang et al. also reached the same conclusion for the RCF life of the AT40 ceramic coating using this RCF test rig [25]. There was an overlap zone between the minimum life of the NiCrBSi/WC–Ni composite coating and the maximum life of the NiCrBSi coating. Thus, evaluating the RCF life only by the average life, maximum life, or minimum life does not provide a reliable estimate. The Weibull distribution is a type of probabilistic method widely used to process scattering data of the fatigue life in reliability analyses and calculations. Based on the number of the Weibull parameters, there are three different types of distributions: one-parameter, two-parameter, and three-parameter Weibull distribution. Typically, it is difficult to process statistical data accurately by one-parameter Weibull distribution, and, although a deeper analysis
Fig. 14. Microhardness of the NiCrBSi/WC–Ni composite coating and NiCrBSi coating.
Fig. 15. Results of the RCF life and failure mode of the NiCrBSi/WC–Ni composite coating and NiCrBSi coating.
Please cite this article as: Z.Q. Zhang, et al., Surf. Coat. Technol. (2014), http://dx.doi.org/10.1016/j.surfcoat.2014.11.061
8
Z.Q. Zhang et al. / Surface & Coatings Technology xxx (2014) xxx–xxx
Fig. 16. Abrasion morphology of the failed NiCrBSi/WC–Ni composite coating: (a) overall observation; and (b) high-magnification observation.
Fig. 17. Spalling morphology of the failed NiCrBSi/WC–Ni composite coating: (a) overall observation; and (b) high-magnification observation.
can be performed using the three-parameter Weibull distribution, this is not effectively applied because of the tedious data processing. Therefore, the two-parameter Weibull distribution is frequently employed to process the fatigue life:
NiCrBSi/WC–Ni composite coating was similar to that of the AT40 ceramic coating. Fig. 18 shows the Weibull distribution plots of the RCF life of the NiCrBSi/WC–Ni composite coating and NiCrBSi coating.
" # N b F ðN Þ ¼ 1− exp − Na where F(N) is the failure probability and N are the cycles of the coating specimen. Na and b are two parameters of the Weibull distribution. Na represents the characteristic life when the failure probability is 63.2%, and b stands for the Weibull plot slope, which can indicate the discrete degree of the RCF life. In general, the RCF life becomes more discrete as the Weibull plot slope b decreases. The values of Na and b can be evaluated using the theory of maximum likelihood estimation (MLE) based on the obtained life. By applying the MLE method on the RCF life of the NiCrBSi/WC–Ni composite coating and NiCrBSi coating with a 90% confidence level, the values of Na were estimated to be 2.1649 and 1.6041, respectively, and the values of b were estimated to be 6.0805 and 5.3149, respectively. The RCF Life of the NiCrBSi/WC–Ni composite coating had a higher dispersion degree than that of the NiCrBSi coating under a contact stress of 0.881 GPa. Additionally, the Weibull plot slope b of the RCF life of the AT40 ceramic coating exhibited a value of 6.2849 under a contact stress of 0.75 GPa [25]. Therefore, the discrete degree of the RCF life of the
Fig. 18. Weibull distribution plot of the RCF life of the NiCrBSi/WC–Ni composite coating and NiCrBSi coating.
Please cite this article as: Z.Q. Zhang, et al., Surf. Coat. Technol. (2014), http://dx.doi.org/10.1016/j.surfcoat.2014.11.061
Z.Q. Zhang et al. / Surface & Coatings Technology xxx (2014) xxx–xxx
The figure depicts the failure percentage of the specimens (ordinate) versus the number of stress cycles for RCF failure (abscissa). The Weibull plot can indicate the statistical percentage of the failed coating in the RCF tests and the RCF life when designating the failure percentage. In Fig. 18, the confidence limits of the NiCrBSi/WC–Ni composite coating and NiCrBSi coating are shown by dotted and dashed lines, respectively. The estimate of the RCF life is reliable because the data points fall within the confidence limits with a confidence level of 90%. Owing to the addition of WC–Ni, the RCF life of the NiCrBSi/WC–Ni composite coating was longer than that of the NiCrBSi coating when designating the failure percentage under this contact level. In addition, The RCF life of the NiCrBSi/ WC–Ni composite coating was easier to predict than that of the NiCrBSi coating, as the increase of the Weibull plot slope b indicates. Therefore, the RCF performance of the NiCrBSi/WC–Ni composite coating was superior to that of the NiCrBSi coating under a contact stress of 0.881 GPa. 4. Conclusions This paper systematically investigated the microstructure, element distribution, phase composition, porosity, microhardness, and RCF performance of NiCrBSi/WC–Ni composite coatings. On the basis of the above results and discussion, it can be concluded that: (1) The NiCrBSi/WC–Ni composite coating presented typical lamellar structures. Sprayed particles were fully flattened and laminated. White WC–Ni particles were homogeneously distributed in the NiCrBSi/WC–Ni composite coating, and W2C/WC phases could be well wetted by other phases. Moreover, shrinkage holes were formed due to the uneven contraction of the melting particles during the cooling. Transverse interlaminar cracks were distributed within a range of 200 μm from the coating surface. The NiCrBSi/ WC–Ni composite coating consists mainly of γ-Ni, Cr7C3, Cr23C6, FeNi3, Ni3Si, CrB, and NiC, in addition to WC and W2C. However, microscale oxygen inevitably exists in the coatings. The NiCrBSi/ WC–Ni composite coating (porosity of 1.84%) had slightly lower density than the NiCrBSi coating (porosity of 1.62%). Furthermore, as a result of the addition of hard WC–Ni, the microhardness of the NiCrBSi/WC–Ni composite coating exceeded 800 HV, which is significantly higher than that shown by the NiCrBSi coating. (2) The NiCrBSi/WC–Ni composite coating and NiCrBSi coating exhibited two distinct failure modes, abrasion and spalling, based on failed surface morphology under contact stress of 0.881 GPa. Abrasion was the main failure mode for the NiCrBSi/WC–Ni composite coating, whereas the main failure mode for the NiCrBSi coating was spalling. That is to say, the damage of the NiCrBSi coating was more serious than that of the NiCrBSi/WC–Ni composite coating.
9
(3) The NiCrBSi/WC–Ni composite coating had a higher dispersion degree of the RCF life and longer life than the NiCrBSi coating under a contact stress of 0.881 GPa. Therefore, the RCF performance of the NiCrBSi/WC–Ni composite coating was superior to that of the NiCrBSi coating.
Acknowledgments The paper was financially supported by 973 Project (2011CB013405), Distinguished Young Scholars of NSFC (51125023), and “Twelfth Five-Year” National Science and Technology Support Program (2011BAF11B08). References [1] L. Wang, J.C. Fang, Z.Y. Zhao, H.P. Zeng, Surf. Coat. Technol. 201 (2007) 5085–5089. [2] H.L. Yu, W. Zhang, H.M. Wang, Tribol. Int. 66 (2013) 105–113. [3] Y.L. An, G.L. Hou, J. Chen, X.Q. Zhao, G. Liu, H.D. Zhou, J.M. Chen, Vacuum 107 (2014) 132–140. [4] J. Chen, Y.L. An, J. Yang, X.Q. Zhao, F.Y. Yan, H.D. Zhou, J.M. Chen, Surf. Coat. Technol. 235 (2013) 521–528. [5] R.G. Song, C. Wang, Y. Jiang, H. Li, G. Lu, Z.X. Wang, J. Alloys Compd. 544 (2012) 13–18. [6] A. Cattini, D. Bellucci, A. Sola, L. Pawłowski, V. Cannillo, Surf. Coat. Technol. 236 (2013) 118–126. [7] Y.R. Niu, X.B. Zheng, H. Ji, L.J. Qi, C.X. Ding, J.L. Chen, G.N. Luo, Fusion Eng. Des. 85 (2013) 1521–1526. [8] R. Mušálek, O. Kovářík, T. Skiba, P. Haušild, M. Karlík, J. Colmenares-Angulo, Intermetallics 18 (2010) 1415–1418. [9] P. Niranatlumpong, H. Koiprasert, Surf. Coat. Technol. 206 (2011) 440–445. [10] T. Liyanage, G. Fisher, A.P. Gerlich, Surf. Coat. Technol. 205 (2010) 759–765. [11] N. Serres, F. Hlawka, S. Costil, C. Langlade, F. Machi, Surf. Coat. Technol. 205 (2010) 1039–1046. [12] L.G. Yu, K.A. Khor, H. Li, K.C. Pay, T.H. Yip, P. Cheang, Surf. Coat. Technol. 182 (2004) 308–317. [13] C.J. Li, G.J. Yang, Int. J. Refract. Met. Hard Mater. 39 (2013) 2–17. [14] Q. Zhan, L.G. Yu, F.X. Ye, X.J. Xue, H. Li, Surf. Coat. Technol. 206 (2012) 4068–4074. [15] G.Z. Xie, J.X. Zhang, Y.J. Lu, K.Y. Wang, X.Y. Mo, P.H. Lin, Mater. Sci. Eng. A 460–461 (2007) 351–356. [16] P. Suresh Babu, Bikramjit Basu, G. Sundararajan, Wear 268 (2010) 1387–1399. [17] C. Guo, J.S. Zhou, J.M. Chen, J.R. Zhao, Y.J. Yu, H.D. Zhou, Wear 270 (2011) 492–498. [18] M.J. Tobar, C. Álvarez, J.M. Amado, G. Rodríguez, A. Yáñez, Surf. Coat. Technol. 200 (2006) 6313–6317. [19] Anup S. Pandkar, Nagaraj Arakere, Ghatu Subhash, Int. J. Fatigue 63 (2014) 191–202. [20] H. Matsunaga, H. Komata, J. Yamabe, Y. Fukushima, S. Matsuoka, Procedia Mater. Sci. 3 (2014) 1663–1668. [21] G.L. Li, Z.Q. Zhang, H.D. Wang, B.S. Xu, Z.Y. Piao, L.N. Zhu, Tribol. Int. 61 (2013) 129–137. [22] J.J. Kang, B.S. Xu, H.D. Wang, C.B. Wang, Tribol. Int. 73 (2014) 128–137. [23] X.C. Zhang, B.S. Xu, F.Z. Xuan, H.D. Wang, Y.X. Wu, Appl. Surf. Sci. 254 (2008) 3734–3744. [24] S. Stewart, R. Ahmed, Wear 253 (2002) 1132–1144. [25] J.J. Kang, B.S. Xu, H.D. Wang, C.B. Wang, Tribol. Int. 66 (2013) 249–258. [26] S. Stewart, R. Ahmed, T. Itsukaichi, Surf. Coat. Technol. 190 (2005) 171–189.
Please cite this article as: Z.Q. Zhang, et al., Surf. Coat. Technol. (2014), http://dx.doi.org/10.1016/j.surfcoat.2014.11.061