Tribol Lett (2016)63:14 DOI 10.1007/s11249-016-0699-2

ORIGINAL PAPER

Investigation of Post-deposition Annealing Effects on Microstructure, Mechanical and Tribological Properties of WC/a-C Nanocomposite Coatings Dongqing He1,3 • Jibin Pu2 • Liping Wang1,2 • Guangan Zhang1 • Yongxin Wang2 Qunji Xue1

•

Received: 18 March 2016 / Accepted: 21 May 2016 Ó Springer Science+Business Media New York 2016

Abstract Nanocomposite WC/a-C coatings were successfully fabricated using a magnetron sputtering process, and post-deposition annealing was conducted in vacuum for 1 h at the annealing temperatures ranging from 100 to 500 °C. The changes in coating structure, internal stress, hardness, toughness, friction coefficient and wear have been investigated to assess the effects of annealing on microstructure, mechanical and tribological properties of the WC/a-C coatings. The results show that the nanocrystalline WC1-x partially decays to metastable W2C when annealing at 300–500 °C and no graphitization of amorphous carbon matrix starts up to 500 °C. This structural change results in a slightly increased hardness and an improved toughness as well as a gradually decreased internal stress. In addition, the time for the annealed coatings to achieve a low steady friction coefficient decreases with the increase of annealing temperature. An optimized tribological property with low friction coefficient of about 0.06 and enhanced wear resistance of the WC/a-C coating is obtained by annealing at 400 °C. Friction reduction and wear resistance caused by & Liping Wang [email protected] & Guangan Zhang [email protected] 1

State Key Laboratory of Solid Lubrication, Lanzhou Institute of Chemical Physics, Chinese Academy of Sciences, Lanzhou 730000, China

2

Key Laboratory of Marine Materials and Related Technologies, Key Laboratory of Marine Materials and Protective Technologies of Zhejiang Province, Ningbo Institute of Materials Technology and Engineering, Chinese Academy of Sciences, Ningbo 315201, China

3

University of Chinese Academy of Sciences, Beijing 100039, China

annealing can be attributed to the friction-induced WO3rich tribofilm which slides against a thin carbon-rich layer on the coating surface resulting in a low friction, and the partition effect of the stationary WO3-rich tribofilm combining with the improved mechanical properties generates a high wear resistance. Keywords Nanocomposite  WC/a-C coatings  Vacuum annealing  Microstructure  Tribological properties

1 Introduction Tungsten carbide (WC) is a well-known refractory material used for hard and damage tolerant coatings due to its high hardness, wear resistance, high melting point (2870 °C), chemical inertness and oxidation resistance [1–4]. However, the relatively high friction coefficient (e.g. 0.5–0.6) of WC can be a drawback and limits overall performance [5]. The combination of hard WC with an amorphous carbon in the form of nanostructured coatings (WC/a-C), such as nanocomposite or multilayered structures, can combine the high hardness and thermal stability of carbides with the low friction coefficient of diamond-like carbon (DLC) to create a high-performance coating [6–9]. So, the carbide-doped DLC coatings (WC/a-C) have been used successfully in several engineering applications, e.g., bearings, pumps, compressors, gears and tools [10, 11]. In previous work, we fabricated nanostructured WC/a-C coatings and reported an improved tribological property with low friction coefficients of 0.05 at 25 °C and 0.28 at 200 °C due to formation of the ‘‘superlattice’’ microstructure and continuously compacted tribofilms [12]. However, a relatively high intrinsic compressive stress [13, 14] and poor toughness [15] of WC/a-C coatings tend to

123

14

Page 2 of 14

catastrophically fail early due to delamination and fracture, which restricts its practical application to the machine components, especially the gear surface which suffers instantaneous impact load. For hard coatings, good toughness is an important attribute for improving wear resistance, which can dissipate strain energy caused by normal and tangential loads in sliding or rolling contact [16, 17]. Therefore, it is important to take more attention to internal stress and toughness of WC/a-C coatings. Several approaches have been developed to reduce the intrinsic stresses and improve toughness, such as formation of the multilayer coatings consisting of soft-hard layers [18], incorporating ductile phases [19] or post-deposition annealing [20]. Post-deposition annealing has been widely used as a convenient and effective method to decrease the internal stress of the coatings without sacrificing its superior properties [20–22], but so far no works have been reported on the improving toughness and tribological properties of WC/a-C coatings by vacuum annealing. In the present work, we have prepared nanocomposite WC/a-C coatings by magnetron sputtering and investigated the evolution of their microstructure, mechanical and tribological properties with annealing temperature in vacuum ranging from 100 to 500 °C and attempt to reveal the toughening, frictionreducing and wear-resistant mechanism of as-annealed WC/a-C coatings. In order to retain full mechanical properties and dimensional stability of the substrates, the highest annealing temperature should be below the recrystallization temperature of itself [23]. Thus, the critical annealing temperature of 500 °C in this study is just only suitable for the high melting point metal substrates which has a relatively higher recrystallization temperature, such as stainless steel, tool steel and titanium alloy.

2 Experimental Details 2.1 Coating Deposition and Annealing Treatment Nanocomposite WC/a-C coatings about 4.46 lm thickness were deposited on AISI 304 stainless steel by a closed field unbalanced magnetron sputtering (CFUBMS) technique with two graphite targets, one WC target and one Cr target. The substrates received a final polish to a surface roughness of Ra B 0.03 lm. All specimens were ultrasonically degreased with acetone and cleaned with Ar? ions at a substrate pulsed bias of -500 V in sequence before coating deposition. Subsequently, a thin Cr metal interlayer with a thickness of about 210 nm was deposited on substrates to improve the adhesion of the coatings to the substrates. Then nanocomposite WC/a-C coatings were deposited at WC target power of 240 W and the graphite target power

123

Tribol Lett (2016)63:14

of 2500 W. Each coating process started at a base pressure of approximately 3 9 10-3 Pa. Magnetron sputtering was conducted at a substrate pulsed bias of -70 V and a process pressure of 0.2 Pa under Ar flow. The substrate temperature was maintained at 160–180 °C during coating deposition. Thermal annealing was performed at 100, 200, 300, 400 and 500 °C in a vacuum heat treating furnace for 1 h, with a chamber pressure of about 2 9 10-3 Pa and a heating rate of 5 °C/min. After the annealing process, samples were cooled inside the furnace to room temperature. 2.2 Coating Characterization The coatings were phase identified by X-ray diffraction (XRD) with a grazing angle of 1° on a D/Max-2400X diffractometer using Cu Ka radiation. The cross-sectional morphologies of as-deposited and annealed coatings were investigated using field emission scanning electron microscopy (FESEM) on a FEI Quanta FEG 250 instrument. A further investigation on microstructure characterization was performed by high-resolution transmission electron microscope (HRTEM) using JEOL 3010 TEM. The coating composition and chemical bonding were investigated by X-ray photoelectron spectroscopy (XPS) on a PerkinElmer PHI-5702 multi-functional photoelectron spectrometer with a monochromatic source of Al Ka (1486.68 eV) radiation. Sample charging is a common problem when using XPS to analyze insulating or only partially conducting materials such as these WC/a-C coatings due to the incomplete neutralization of the photoemitted electrons. Thus, an electron flood gun was used for the neutralization of these excess electrons with the energy varying from 0 to 5 eV depending on the degree of charging. Samples were previously cleaned using Ar? ion bombardment with an energy of 2.5 keV to remove the largest possible amount of native oxide surface, and the sputtering was kept for 60 s with a nominal drain current of 1.7 lA/cm2 and an Ar partial pressure of 2.5 9 10-3 Pa. The area of the analyzed zone was approximately 200 9 200 lm2. All the peaks were calibrated by taking the C1 s peak (284.8 eV) as a reference. Peak synthesis was performed after a Shirley background subtraction. XPS Peak4.1 fitting software [9] was used to fit the XPS peaks by maintaining the appropriate peak doublet area ratio and binding energies as well as the appropriate full width half maximum (FWHM) value. A Gaussian–Lorentzian (10 %) product function component line shape was found to provide a satisfactory fitting result. Micro-Raman measurements were taken to identify the bonding structure within the coatings using a LabRAM Jobin–Yvon spectrometer with a laser excitation wavelength of 532 nm and a spot size of 10 lm in diameter.

Tribol Lett (2016)63:14

Hardness and Young’s moduli were measured using a Nanoindenter II microprobe with a diamond Berkovich (three-sided pyramid) indenter tip. The maximum load was selected in such a way that the maximum indentation depth did not exceed 10–15 % of the coating thickness in order to minimize the substrate contribution. For each sample, six indents were measured and then averaged. The toughness of as-annealed WC/a-C coatings was evaluated by Rockwell C test with the crack density (divide cracks’ number by the perimeter of the indent) and average crack length around the indents. The coatings’ internal stress was measured by measuring the beam bending using a profilometer. The deformation of the substrate due to the stress in the WC/a-C coatings was measured by using a thin silicon wafer beam as a substrate and calculating the radius of curvature of the beam, and hence the internal stress was calculated by the Stoney’s equation [24].   Es ts2 1 1 rðx; yÞ ¼  6ð1  ms Þ tf Rs Rf where Es, vs, ts and tf are the Young’s modulus, Poisson ratio, thickness of the substrate and thickness of the film, respectively. Rs is the radii of curvature of the substrate before coating deposition, and Rf is the radii of curvature after coating deposition. 2.3 Tribological Tests A ball-on-disk tribo-meter (CSM Instrument) was used for friction measurement and qualitative comparison of the wear resistance of the WC/a-C coatings at room temperature in ambient air with a relative humidity of 35 %. The friction tests were performed on a reciprocating mode with amplitude of 5 mm, a normal load of 10 N and sliding frequency of 5 Hz. An AISI52100 steel ball of 6 mm diameter with hardness of HV725 was selected as the counterpart. The initial Hertzian contact stress in the friction tests was about 1.28 GPa. Before each test, the frictional pairs were ultrasonically cleaned with acetone and alcohol, respectively. After the friction tests, the wear rate was calculated by measuring the volume of the wear track using an optical profiler (Micro XAM 3D). The morphologies and composition distribution of the wear scars were studied using FESEM with energy dispersive X-ray spectroscopy (EDS). Then, a comprehensive structural information of transfer film on the steel balls was collected by Raman spectroscopy from several different zones. The chemical compositions and chemical bonding in the wear tracks were immediately analyzed by XPS after the friction tests in order to avoid natural oxidation as much as possible. The XPS date acquired from the wear tracks is an average result of the analyzed zone, which is randomly selected in the wear tracks.

Page 3 of 14

14

3 Results and Discussion 3.1 Composition and Microstructure The WC/a-C coating consists of 91.62 % C, 5.43 % W and 2.95 % O, and the content of C, W and O is almost constant with coating depth according to the elemental depthprofiling analysis by glow discharge spectrometry (GDS; data not shown). The XRD patterns of the as-deposited and annealed coatings are shown in Fig. 1. Although a pure sintered WC target is used, and a-WC is the only thermodynamically stable carbide phase in the W–C system at room temperature, no peaks from the hexagonal a-WC are observed for the as-deposited coating. Similar results have been reported in the literatures for WC coatings prepared by magnetron sputtering [25, 26]. All the samples show only one asymmetric and broad peak at 37.5°, which is from the diffraction of the (111) plane of nanocrystalline bWC1-x and indicates lattice defects in the coatings. For the annealed coatings, no change in the diffraction patterns up to a temperature of 500 °C. However, it is hard to deduce whether there is a phase transformation, because the broad peak cannot be clearly attributed to WC1-x, W2C or mixture of both. In fact, it has been reported that b-WC1-x phase can decay to b-W2C and a-WC at around 430 °C, and the transformation temperature is estimated to be at 470 °C [27]. Normally, the precipitated new carbide phase (b-W2C or a-WC) has very small grain size and is even amorphous at primary stage of phase transformation; thus, the exact determination of the individual new phase by XRD is difficult due to overlapping peaks and peaks broadening. With the aim of obtaining a deeper insight into the electronic state of the annealed coatings, XPS analysis was carried out on the as-deposited coating and three specific

Fig. 1 Glancing incidence XRD patterns of the as-deposited WC/a-C coating and the coatings annealed at different temperatures

123

14

Page 4 of 14

samples, which were annealed at 100, 300 and 500 °C. As shown in Fig. 2, the W 4f spectrum of the as-deposited coating surface is mainly characterized by binding energies of *32.0 and *34.1 eV, which is associated with the presence of WC1-x [12]. It is well consisted with the XRD analysis results that the coatings are dominated by the WC1-x phase. Though the coating surface was cleaned by Ar? ion bombardment, a small amount of tungsten oxides was detected in almost all coatings. W 4f region presents two pairs of oxide peaks corresponding to WO2 (W 4f 7/2 at *32.9 eV and W 4f 5/2 at *35.1 eV) and WO3 (W 4f 7/2 at *35.5 eV and W 4f 5/2 at *37.5 eV) for the asdeposited coating and the coating annealed at 100 °C. By increasing the annealing temperature up to 300 °C, apart from the peaks of WC1-x and WO3, two additional peaks occur at *31.6 and *33.7 eV, corresponding to W2C or WC [28, 29]. It means that the structural transformations between different phases of the WC systems may take place in a relatively low temperature of 300 °C:

Fig. 2 XPS spectra of W 4f energy region for the as-deposited WC/ a-C coating and the coatings annealed at 100, 300 and 500 °C

123

Tribol Lett (2016)63:14

WC1-x ? W2C/WC. Further annealing at 500 °C, it is observed that the relative intensity of the W 4f peaks from W2C or WC are markedly increased, which indicates the formation of these phases become more pronounced when the annealing temperature is increased from 300 to 500 °C. Additionally, the WO3 peaks of annealed coatings are obviously stronger than the one of as-deposited coating. This phenomenon can be attributed to the presence of absorbed oxygen and water vapor on the coating surface which result in the surface oxidation during vacuum annealing [30]. Therefore, it can be concluded that the sputtered WC1-x in this study may not be stable as described in the phase diagram of W–C system [31]. Several previous studies [27, 32] reported analogous results, the WC1-x may first decay to metastable W2C at a relatively low temperature (300–700 °C), and then the W2C transforms into stable phase of WC at a higher temperature (700–1000 °C). Thus, the precipitated carbide phase according to the XRD and XPS analyses is most likely nanocrystalline or amorphous W2C. This was further confirmed by HRTEM investigation of the coating microstructure before and after annealing, as shown in Fig. 3. HRTEM images together with their associated selected area electron diffraction (SAED) patterns indicate that both of two samples are nanocomposite of nanocrystallines embedded in an amorphous matrix. In the asdeposited coating, there are small crystals about 2–5 nm in diameter with the interplanar distances of 2.4, 2.1 and ˚ (see Fig. 3a). These values can be correlated with the 1.3 A (111), (200) and (311) plane of WC1-x [9, 12], respectively. Focusing on the sample annealed at 500 °C, the ˚ can be measured interplanar distances of 2.1, 2.3 and 2.4 A corresponding to (200) WC1-x, (101) W2C and (111) WC1-x, crystal planes, respectively, corroborating the presence of precipitated W2C (see Fig. 3b). Moreover, the indexation of the SAED rings is in agreement with a mixture of WC1-x and W2C phases. Raman spectroscopy of the coatings was also analyzed to investigate the effect of annealing on the amorphous carbon matrix within the coatings. Figure 4 shows that no obvious difference among the Raman spectra of the asdeposited coating and the annealed coatings through the entire annealing sequence can be observed. All the spectra consist of two features: a G peak at 1554 cm-1 which is identified to reflect C sp2 vibrations, and a D disorder peak at 1367 cm-1 which indicates the presence of nano-size graphitic domains. According to Ferrari’s [33] theory about the Raman spectra of disordered and amorphous carbon, the peak positions and the intensity ratio of D and G peaks in the Raman spectrum are the most important parameters in understanding the structural information of the film. In general, the G peak position and ID/IG ratio of annealed coatings increase with the annealing temperature for

Tribol Lett (2016)63:14

Page 5 of 14

14

Fig. 4 Raman spectra of the as-deposited WC/a-C coating and the coatings annealed at different temperatures

visible for the as-deposited coating, and similar results are found in the coatings annealed at 100, 200 and 300 °C (see Fig. 5a–d). Such typical microstructure is most likely due to the simple geometric shadowing of the vapor beam by the growing film, and sensitive both to the nucleation and growth process and to the diffusion process. It is unstable and may be eliminated by high temperature annealing [35]. For the coatings annealed at 400 and 500 °C, it can be observed that the coarse columnar structure transforms to a more restrained columnar structure or fine-grained structure (see Fig. 5e, f). These results imply that the re-growth of the coatings like the ‘‘re-crystallization’’ occurs at this temperature, and the precipitated new carbide phase (W2C) may inhibit the growth of the columnar structure by separating the WC1-x grains. In addition, this lower re-growth temperature of the WC/a-C coatings may be due to its higher density of lattice defects, particularly the boundaries between nanocrystalline WC1-x and amorphous carbon. 3.2 Mechanical Properties Fig. 3 HRTEM micrographs and SAED patterns of the as-deposited WC/a-C coating (a) and the coating annealed at 500 °C (b)

amorphous carbon film due to coating graphitization [34]. However, the present study does not follow the same trend, indicates no graphitization up to 500 °C. The cross-sectional morphologies of the as-deposited WC/a-C coating and the coatings annealed at different temperatures are presented in Fig. 5. In order to acquire a comprehensive information, four different zones on the fracture surface were observed for each sample. Finally, almost the same morphological characteristics were collected from these four different zones indicating a good uniformity of the coatings. A clear columnar structure is

Figure 6 shows variations of coating hardness, elastic modulus and internal stress as a function of annealing temperature. Almost no change of the hardness and modulus value in annealed coatings is observed from 100 to 300 °C, and then, the hardness and modulus slightly increase with further increase of the annealing temperature, as shown in Fig. 6a. Generally, carbon-based coatings start to graphitize from 300 °C, which results in softening of the annealed coating. But refraining from the graphitization and the phase transformation during the heat treatments from 300 to 500 °C should be responsible for the increasing hardness and modulus. It can be supposed that the precipitated new carbide phase (W2C) around nanocrystalline WC1-x grains result in the formation of a

123

14

Page 6 of 14

Tribol Lett (2016)63:14

Fig. 5 Cross-sectional micrographs of the as-deposited WC/a-C coating (a) and the coatings annealed at different temperatures: (b) 100 °C; (c) 200 °C; (d) 300 °C; (e) 400 °C; (f) 500 °C

Fig. 6 Hardness, Young’s modulus (a) and the internal stress (b) of the as-deposited WC/a-C coating and the coatings annealed at different temperatures

123

composite structure, consisting of well-segregated nanocrystalline WC1-x, W2C and amorphous carbon phase. In such composites, the nanocrystalline phase serves as a dislocation barrier, adding to grain boundary barriers, which together cause the hardness and modulus increase. Meanwhile, all the WC/a-C coatings exhibit compressive intrinsic stress (see Fig. 6b) and the intrinsic stress continuously decreases with the increase of annealing temperature. The release of intrinsic stress of the WC/a-C coatings associated with annealing involves two stages. The first stage is the release of the ion bombardment induced stress when annealing temperature below 200 °C. According to the Davis’s model [36] for the formation of compressive stress in thin films by ion bombardment, compressive stress arises when a growing film is bombarded by atoms or ions with energies of tens or hundreds of electronvolts by a process of ‘‘atomic peening’’. This kind of stress can be reduced by providing the additional energy required to release implanted atoms from their metastable positions within the film, such as low temperature annealing [22, 23]. The second stage is the stress release owing to the conversion of sp3 sites to sp2 during high temperature annealing. For the latter, Sullivan et al. [37, 38] proposed a structural model for stress relief, in which a small fraction of sp3 sites convert to sp2 to relieve stress. It is noted that while the atomic volume of sp2 site exceeds that of a sp3 site, its in-plane size is less, due to its shorter bond length. Thus, the formation of sp2 sites with their r plane aligned in the plane of compression will

Tribol Lett (2016)63:14

relieve a biaxial compressive stress. Through the derivation [39], only a small increase (*1 %) in sp2 fraction is needed to account for the stress relief, such a slight structural variation may be ignored in the visible Raman spectra. More than reducing the intrinsic stress, vacuum annealing also improved the toughness of the coatings. This is illustrated in Fig. 7, which shows the micrographs of Rockwell C test indents (see Fig. 7a), the crack density and average crack length around the indents (see Fig. 7b) of the as-deposited coating and the coatings annealed at different temperatures. Although there are many methods for thin film toughness assessment [40], however, until

Page 7 of 14

14

now, there is neither standard procedure nor commonly accepted methodology to follow. Observing morphology of the indents or measuring the crack density and average crack length in Rockwell C test may be an intuitive and effective way to evaluate the coating toughness according to the original definition of itself [40]. Almost the same hardness (RHB 88 ± 1) of the substrates was measured before and after each annealing process. The same pressed depth (*104 lm) and width (*510 lm) indicates the same substrate deformations during the Rockwell C indentation tests. Thus, it is reasonable to relate crack analysis results to the coating toughness at a qualitative level. It can be seen that all of the coatings exhibit radial cracks but no delamination at the indent corners under the same load of 600 N, with the increase of the annealing temperature the crack density and average crack length around the indents present a trend of decrease, particularly above 200 °C. These results mean that an improved coating toughness can be obtained after a high temperature annealing. This may be attributed to the release of the intrinsic stress and the composite coating structure caused by the phase transition as discussed above. Such composite structure not only provides effective barriers against the initiation and propagation of cracks, but also allow grain boundary sliding [20], thus generate a remarkable coating toughness.

3.3 Tribological Properties

Fig. 7 Micrographs of Rockwell C test indents (a), the crack density and average crack length around the indents (b) of the as-deposited WC/a-C coating and the coatings annealed at different temperatures

The variation of coefficients of friction (COF) as a function of reciprocating sliding time of as-deposited WC/a-C coating and the coatings annealed at different temperatures under dry contact at the temperatures of 25 °C are presented in Fig. 8a. For the as-deposited coating, friction coefficient stabilizes above 0.1 after about 600 s. After this period, friction coefficient begins to decrease and finally stabilize to a value of 0.05. For the coatings annealed at 100 and 200 °C, they exhibit a similar trend of an increasing friction after a short running-in period and start to decline at about 0.15 and 0.11, respectively, and both reduce to 0.05 after 3000 s. This reduction in friction occurs earlier for the annealed coatings than the asdeposited one. A more stable and lower friction coefficient of about 0.07 is observed over the entire test range on the coating annealed at 300 °C. By further rising the annealing temperature to 400 and 500 °C, corresponding annealed coatings start stepping into the stage of low friction after a very short running-in stage (after 1200 s). Thus, it can be claimed that the higher temperature annealing is conducted for the coatings, the earlier it can enter into the stage of low friction. To further investigate the friction behaviors of the

123

14

Page 8 of 14

Fig. 8 Friction coefficients (a) and wear rates (b) of the as-deposited WC/a-C coating and the coatings annealed at different temperatures

WC/a-C coatings, SEM micrographs of wear scars for steel balls sliding against the WC/a-C coatings are shown in Fig. 9. Transfer film formation is observed for all coatings and is especially pronounced for the coatings annealed at high temperatures (see Fig. 9e, f). Such an intact transfer film could lead to a low friction. This observation is in good agreement with the friction behaviors as shown in Fig. 8a. Therefore, the low friction coefficient, usually attributed to the coating, is really due to interfacial sliding between the transfer film and the worn coating surface [41]. This will be further discussed latter. The specific wear rates of the WC/a-C coatings are shown in Fig. 8b. It clearly indicates that the coatings exhibit apparently different wear properties for different annealing temperatures. Overall, variation in wear rates of the WC/a-C coatings along the series appears to correlate with the friction characteristics. With the increase of the annealing temperature, the wear rate gradually decreases to the lowest value at 400 °C and then recovery to a high value at 500 °C. It can be further confirmed by the 3D images and 2D cross-sectional profiles of the wear tracks as

123

Tribol Lett (2016)63:14

shown in Fig. 10. At the initiation of sliding, because the coating is much harder than the steel ball, the contact area of the steel ball is worn away rapidly and reacts with the environment resulting in a rough worn surface with amount of protrusions (Fe and iron oxides) [16, 42]. These protrusions are sufficiently hard that they penetrate the coating on the mating surface. Plastic grooves caused by the penetration and plowing of hard protrusions are observed in wear tracks of all coatings, and some small isolated spalls are apparent that follow the trajectory of the grooves in the coatings annealed at low temperatures (see Fig. 10a–c). It is worth noting that an anomalous smooth but deep wear track is presented in the coating annealed at 500 °C (see Fig. 10f) that obviously differ from the shallow grooves as observed in the coatings annealed at 300 and 400 °C (see Fig. 10d, e). The improved wear resistance of the annealed coatings can be attributed to the following two factors. First, it is clear that steady low friction coefficients occur when well adhered and intact transfer films cover the contact area of the steel ball as discussed in Figs. 8 and 9. Meanwhile, the partition effect of the transfer films can restrict direct contact of the steel ball with the coatings and thus prevent intense wear [41]. Second, the hardening and toughening caused by phase transition of high temperature annealed coatings may be the primary cause of the improved wear resistance. Generally, once the contact conditions are severe enough to exceed the cracking threshold for the WC/a-C coatings, because of its low toughness and moderate hardness, both median and lateral cracks form. The lateral cracks extension creates spalls during reciprocating sliding resulting in a high wear rate, just like the as-deposited coating as shown in Figs. 8b and 10a. For the coatings annealed at 300 and 400 °C, the relative high hardness provides good bearing capacity and against the penetration of hard protrusions on the mating steel surface. Moreover, the improved toughness can dissipate strain energy under high levels of normal and tangential loads in sliding contact, thereby preventing the crack initiation and propagation. Thus, an optimized wear resistance can be obtained, as shown in Fig. 8b, while the deteriorated wear resistance of the coating annealed at 500 °C may be mainly attributed to the severe oxidation of the coating surface due to the presence of a sufficient W2C. It can be confirmed by the following component analysis of the worn coating surface and the corresponding transfer films. Another ancillary friction test was also carried out under the same condition but only run 30 min for the purposes of investigating how the high temperature annealing rapidly resulting in a low friction. After this test, the steel ball contact surfaces are examined by SEM to reveal the morphological features of the worn area, while the elemental composition of this area is determined using EDS, as

Tribol Lett (2016)63:14

Page 9 of 14

14

Fig. 9 SEM images of wear scars on the counterpart balls after sliding against the as-deposited WC/a-C coating (a) and the coatings annealed at different temperatures: (b) 100 °C; (c) 200 °C; (d) 300 °C; (e) 400 °C; (f) 500 °C

Fig. 10 3D images and 2D cross-sectional profiles of wear tracks on the as-deposited WC/a-C coating (a) and the coatings annealed at different temperatures: (b) 100 °C; (c) 200 °C; (d) 300 °C; (e) 400 °C; (f) 500 °C

123

14

Page 10 of 14

Tribol Lett (2016)63:14

Fig. 11 SEM images and elemental EDS maps of wear scars on the counterpart balls after sliding against the as-deposited WC/a-C coating (a) and the coatings annealed at specific temperatures: (b) 100 °C; (c) 300 °C; (d) 500 °C

shown in Fig. 11. For steel balls sliding against the asdeposited coating and the coating annealed at 100 °C (see Fig. 11a, b), no clear transfer film can be observed and the corresponding elemental distributions indicate an enrichment of O on wear scars due to oxidation. The Raman spectra of the wear scars further confirm that neither a decent amount of carbonaceous layers nor tungsten oxides transfer to the steel ball after the friction test, as shown in Fig. 12. It only shows the presence of two bands at 720 and 945 cm-1 which can be assigned to ferritungstite [43] (see Fig. 12a). The peak at 945 cm-1 can also possibly be assigned to the stretching mode of the W=O bonds that appeared on the boundaries of the amorphous or nanostructured tungsten oxides [44]. That is why these two coatings still exhibit high friction after 30 min as shown in Fig. 8. However, some inconsecutive transfer film is observed on the steel ball sliding against the coating annealed at 300 °C and mainly composed of C, O and W (see Fig. 11b). Two noticeable changes occur in the Raman spectra of this transfer film (see Fig. 12a); one is the appearance of a broad peak in the range of 600–1000 cm-1 from WO3 [43, 45]. The strong broadening of this band suggests that the formed compounds present a high structural disorder, indicating low crystallinity. And the other is the presentation of a typical feature of polycrystalline

123

graphite at about 1345 and 1600 cm-1, revealing that the evolution of the sp3-hybridized to the more stable sp2hybridized carbon (graphitization) at the friction contact faces as well as the transformation from short-range to long-range ordered sp2-bonded structure (crystallization) inside the transfer film must occur [33, 46]. So, the relatively low friction of the coating annealed at 300 °C should be attributed to the interfacial sliding between a thin layer of graphitized carbon on the coating surface and a carbonrich tribofilm which contains poor crystalized WO3. In particular, the steel ball sliding against the coating annealed at 500 °C is entirely covered by a clear transfer layer, and amount of transferred W with O is the notable feature (see Fig. 11d). Also notice that the Raman spectra of this transfer film show a small peak at 275 cm-1 and a sharp peak at 880 cm-1 which are associated with the formation of crystalline WO3 [9, 45, 47], and almost no carbon signal is detected, as shown in Fig. 12a. It means the transfer film is dominated by crystalline WO3. Low friction caused by friction-induced WO3-rich tribofilm has been previously reported [6, 13, 48]; thus, the outstanding low friction of the high temperature annealed coating at the primary stage of friction test is ultimately plausible. The Raman spectra of the transfer film formed after 60 min friction test are shown in Fig. 12b. It is revealed

Tribol Lett (2016)63:14

Page 11 of 14

14

Fig. 12 Raman spectra of transfer film formed on the steel balls sliding against the as-deposited WC/a-C coating and the coatings annealed at 100, 300 and 500 °C after 30 min (a) and 60 min (b) friction test

that the transfer film originated from the as-deposited coating and the coatings annealed at 100 and 300 °C consist of crystalline WO3 according to the Raman peaks at 275, 880 and 945 cm-1, and graphitized carbon according to the two well-defined peaks at 1355 and 1600 cm-1. It is well consistent with the friction coefficients as shown in Fig. 8a that both of them achieve low friction after a relatively long time due to the formation of an effective tribofilm. An interesting observation for the transfer film originated from the coating annealed at 500 °C is the appearance of a stronger signal of crystalline WO3 with little carbon, which is very similar with the one observed after 30 min friction test. It is indicated that friction-induced coating oxidation takes place in the whole friction process with a WO3-rich tribofilm continually transferring to the steel ball, thus resulting in a steady low friction but sever oxidation wear as shown in Figs. 8 and 10f. Friction-induced formation of WO3 on sliding contact surface was further confirmed by XPS analysis in wear tracks of the coatings annealed at 100, 300 and 500 °C, as shown in Fig. 13. Two XPS peaks from WO3 can be observed at *35.5 (W 4f 7/2) and *37.5 eV (W 4f 5/2) for the annealed coatings, especially pronounced for the

coating annealed at 500 °C. In addition to the peaks from WC1-x at *32.0 (W 4f 7/2) and at *34.1 eV (W 4f 5/2), two clear peaks from W2C at *31.6 (W 4f 7/2) and *33.7 eV (W 4f 5/2) are also detected for the coatings annealed at 300 and 500 °C. It further confirmed that the phase transition as discussed above: WC1-x ? W2C, takes place through the entire coating instead of just on the coating surface. While the precipitated W2C phase is not stable, causing immediate oxidation to form tungsten oxide when exposed to air, let alone under dry sliding condition. That is, friction-induced formation of WO3 will be much easier for the coatings annealed at a higher temperature, and a steady flow of precipitated W2C phase attributes to the formation of a continuously compacted WO3-rich tribofilm. This in combination with the carbon graphitization in the worn coating surface indicates that the low friction property is the result of interfacial sliding between the WO3-rich tribofilm and a thin self-lubricating carbon (graphite) layer. To further confirm the primary low-friction mechanism, the effects of humidity on the friction coefficient of the coating annealed at 500 °C were investigated, as shown in Fig. 14. The friction coefficient, in the steady state,

123

14

Page 12 of 14

Tribol Lett (2016)63:14

stabilizes at about 0.05–0.06 both at 0 and 30 % RH, and increases to 1.0 at 50 % RH and 0.17 at 70 % RH. It goes against the concept of a higher humidity leading to a lower friction coefficient due to passivation mechanism [49], which is well known in the hydrogen-free amorphous carbon coatings. We therefore claim that it is the WO3-rich tribofilm that dominates the friction behavior rather than graphitized carbon on the coating surface. So, a low friction can be still achieved even in dry air condition due to absence of a sufficient source for adhesion (less dangling bonds of carbon atoms on the tribofilm and weak interactions between WO3 and carbon) at the sliding interface. The agglomeration of wear debris was enhanced in a high humid environment, which impedes the sliding then resulting in a relatively high friction at 50 and 70 % RH [50]. Thus, the low-friction mechanism of the high temperature annealed coatings can be summarized in Fig. 15. The friction-induced WO3 upon a thin graphitized carbon layer gradually adheres strongly to steel ball forming a stationary WO3-rich tribofilm. In addition to the lubrication of graphitized carbon on the coating surface, the interfacial sliding between the WO3-rich tribofilm and the graphitized carbon layer minimizes adhesion and interaction of the tribopair thus resulting in a steady low friction.

4 Summary and Conclusions

Fig. 13 XPS spectra of W 4f energy region for wear tracks of the coatings annealed at 100, 300 and 500 °C

Fig. 14 Friction coefficients of the coating annealed at 500 °C under different humidity conditions

123

The influence of post-deposition annealing on microstructure, mechanical and tribological properties of WC/a-C nanocomposite coatings was investigated. It has been demonstrated that the nanocrystalline WC1-x partially decays to metastable W2C by a vacuum annealing at 300–500 °C and no graphitization of amorphous carbon matrix starts up to 500 °C. These annealed coatings exhibit a slightly increased hardness and an improved toughness as well as a gradually decreased internal stress with the increase of annealing temperature, especially pronounced above 200 °C. It is also found that the higher temperature annealing is conducted for the coatings, the earlier it can enter into the stage of low friction. Optimized tribological properties with low friction coefficient and enhanced wear resistance of the WC/a-C coating are obtained by annealing at 400 °C. The annealing effect on the friction and wear behaviors is associated with changes in the structure and mechanical properties of the coatings. The interfacial sliding between the friction-induced WO3-rich tribofilm and the graphitized carbon layer should be responsible for the low friction. The improved hardness and toughness, combining with the formation of continuously compacted WO3-rich tribofilm, is a critical factor in achieving a low wear rate.

Tribol Lett (2016)63:14

Page 13 of 14

14

Fig. 15 Schematic representation of the steadystate contact condition for the high temperature annealed WC/ a-C coating

Acknowledgments The authors are grateful for financial support from the National Basic Research Program of China (973 Program, Grant No. 2013CB632302), National Natural Science Foundation of China (Grant No. 51322508) and the Program of the Light of the Chinese Academy of Sciences in China’s Western Region (2013).

References 1. Trindade, B., Vieira, M.T., Bauer-Grosse, E.: Amorphous phase forming ability in (W–C)-based sputtered films. Acta Mater. 46, 1731–1739 (1998) 2. Voevodin, A.A., O’Neill, J.P., Prasad, S.V., Zabinski, J.S.: Nanocrystalline WC and WC/a-C composite coatings produced from intersected plasma fluxes at low deposition temperatures. J. Vac. Sci. Technol. A 17, 986–992 (1999) 3. Park, S.J., Lee, K.R., Ko, D.H., Eun, K.Y.: Microstructure and mechanical properties of WC-C nanocomposite films. Diam. Relat. Mater. 11, 1747–1752 (2002) 4. Aubert, P., Perrie`re, J.: Microstructural and mechanical investigations of tungsten carbide films deposited by reactive RF sputtering. Surf. Coat. Technol. 200, 6469–6473 (2006) 5. Mateos, J.J.M.E., Cuetos, J.M., Fernandez, E., Vijande, R.: Tribological behaviour of plasma-sprayed WC coatings with and without laser remelting. Wear 239, 274–281 (2000) 6. Banerji, A., Bhowmick, S., Alpas, A.T.: High temperature tribological behavior of W containing diamond-like carbon (DLC) coating against titanium alloys. Surf. Coat. Technol. 241, 93–104 (2014) 7. Liu, Y., Gubisch, M., Hild, W., Scherge, M., Spiess, L., Knedlik, Ch., Schaefer, J.A.: Nanoscale multilayer WC/C coatings developed for nanopositioning, part II: friction and wear. Thin Solid Films 488, 140–148 (2005) 8. Voevodin, A.A., O’Neill, J.P., Zabinski, J.S.: Tribological performance and tribochemistry of nanocrystalline WC/amorphous diamond-like carbon composites. Thin Solid Films 342, 194–200 (1999) 9. Abad, M.D., Mun˜oz-Ma´rquez, M.A., El Mrabet, S., Justo, A., Sa´nchez-Lo´pez, J.C.: Tailored synthesis of nanostructured WC/aC coatings by dual magnetron sputtering. Surf. Coat. Technol. 204, 3490–3500 (2010) 10. Roth, D., Rau, B., Roth, S., Mai, J., Dittrich, K.-H.: Large area and three-dimensional deposition of diamond-like carbon films for industrial applications. Surf. Coat. Technol. 74(75), 637–641 (1995)

11. Joachim, F., Kurz, N., Glatthaar, B.: Influence of coatings and surface improvements on the lifetime of gears. Gear Technol. 21, 50–56 (2004) 12. Pu, J., He, D., Wang, L.: Effects of WC phase contents on the microstructure, mechanical properties and tribological behaviors of WC/a-C superlattice coatings. Appl. Surf. Sci. 357, 2039–2047 (2015) 13. Wa¨nstrand, O., Larsson, M., Hedenqvist, P.: Mechanical and tribological evaluation of PVD WC/C coatings. Surf. Coat. Technol. 111, 247–254 (1999) 14. Wang, A.Y., Lee, K.R., Ahn, J.P., Han, J.H.: Structure and mechanical properties of W incorporated diamond-like carbon films prepared by a hybrid ion beam deposition technique. Carbon 44, 1826–1832 (2006) 15. Yao, N., Evans, A.G., Cooper, C.V.: Wear mechanism operating in W-DLC coatings in contact with machined steel surfaces. Surf. Coat. Technol. 179, 306–313 (2004) 16. Voevodin, A.A., Zabinski, J.S.: Supertough wear-resistant coatings with ‘chameleon’ surface adaptation. Thin Solid Films 370, 223–231 (2000) 17. Gorishnyy, T.Z., Olson, L.G., Oden, M., Aouadi, S.M., Rohde, S.L.: Optimization of wear-resistant coating architectures using finite element analysis. J. Vac. Sci. Technol. A 21, 332–339 (2003) 18. Qi, Z.Q., Meletis, E.I.: Mechanical and tribological behavior of nanocomposite multilayered Cr/a-C thin films. Thin Solid Films 479, 174–181 (2005) 19. Zhang, S., Sun, D., Fu, Y., Du, H.: Toughening of hard nanostructural thin films: a critical review. Surf. Coat. Technol. 198, 2–8 (2005) 20. Friedmann, T.A., Sullivan, J.P., Knapp, J.A., Tallant, D.R., Follstaedt, D.M., Medlin, D.L., Mirkarimi, P.B.: Thick stress-free amorphous-tetrahedral carbon films with hardness near that of diamond. Appl. Phys. Lett. 71, 3820–3822 (1997) 21. Li, H., Xu, T., Wang, C., Chen, J., Zhou, H., Liu, H.: Annealing effect on the structure, mechanical and tribological properties of hydrogenated diamond-like carbon films. Thin Solid Films 515, 2153–2160 (2006) 22. Zhang, W., Tanaka, A., Wazumi, K., Koga, Y., Xu, B.S.: The effect of annealing on mechanical and tribological properties of diamond-like carbon multilayer films. Diam. Relat. Mater. 13, 2166–2169 (2004) 23. Beck, P.A.: Annealing of cold worked metals. Adv. Phys. 3, 245–324 (1954) 24. Doerner, M.F., Nix, W.D.: Stresses and deformation processes in thin films on substrates. Crit. Rev. Solid State Mater. Sci. 14, 225–268 (1988)

123

14

Page 14 of 14

25. Mrabet, S.E., Abad, M.D., Lo´pez-Cartes, C., Martı´nez-Martı´nez, D., Sa´nchez-Lo´pez, J.C.: Thermal evolution of WC/C nanostructured coatings by Raman and in situ XRD analysis. Plasma Process. Polym. 6, S444–S449 (2009) 26. Radic, N., Grzeta, B., Milat, O., Ivkov, J., Stubicar, M.: Tungsten–carbon films prepared by reactive sputtering from argon– benzene discharges. Thin Solid Films 320, 192–197 (1998) 27. Kong, P.C., Suzuki, M., Young, R., Pfender, E.: Synthesis of bWC1-x in an atmospheric-pressure, thermal plasma jet reactor. Plasma Chem. Plasma Process. 3, 115–133 (1983) 28. Morimitsu, L.C.A., Ospina, R.O., Carmona, J.M.G., Parra, E.R., Arango, P.A.: Deposition and computational analysis of WC thin films grown by PAPVD. Revista Mexicana de Fı´sica 59, 106–111 (2013) 29. Kojima, I., Miyazaki, E., Inoue, Y., Yasumori, I.: Catalysis by transition metal carbides. VI, Hydrogenation of carbon monoxide over WC, W2C, and W powder catalysts. Bull. Chem. Soc. Jpn. 58, 611–617 (1985) 30. Ospina, R., Castillo, H.A., Benavides, V., Restrepo, E., Arango, Y.C., Arias, D.F., Devia, A.: Influence of the annealing temperature on a crystal phase of W/WC bilayers grown by pulsed arc discharge. Vacuum 81, 373–377 (2006) 31. Kurlov, A.S., Gusev, A.I.: Tungsten carbides and WC phase diagram. Inorg. Mater. 42, 121–127 (2006) 32. Lin, M.-H.: Synthesis of nanophase tungsten carbide by electrical discharge machining. Ceram. Int. 31, 1109–1115 (2005) 33. Ferrari, A.C., Robertson, J.: Interpretation of Raman spectra of disordered and amorphous carbon. Phys. Rev. B 61, 14095– 14107 (2000) 34. Shroder, R.E., Nemanich, R.J., Glass, J.T.: Analysis of the composite structures in diamond thin films by Raman spectroscopy. Phys. Rev. B 41, 3738–3745 (1990) 35. Dirks, A.G., Leamy, H.J.: Columnar microstructure in vapordeposited thin films. Thin Solid Films 47, 219–233 (1977) 36. Davis, C.A.: A simple model for the formation of compressive stress in thin films by ion bombardment. Thin Solid Films 226, 30–34 (1993) 37. Sullivan, J.P., Friedmann, T.A., Baca, A.G.: Stress relaxation and thermal evolution of film properties in amorphous carbon. J. Electron. Mater. 26, 1021–1029 (1997) 38. Sullivan, J.P., Friedmann, T.A., Dunn, R.G., Stechel, E.B., Schultz, P.A., Siegal, M.P., Missert, N.: The electronic transport

123

Tribol Lett (2016)63:14

39.

40.

41.

42.

43.

44.

45.

46. 47.

48.

49.

50.

mechanism in amorphous tetrahedrally-coordinated carbon films. Mater. Res. Soc. Symp. Proc. 498, 97–102 (1998) Ferrari, A.C., Kleinsorge, B., Morrison, N.A., Hart, A., Stolojan, V., Robertson, J.: Stress reduction and bond stability during thermal annealing of tetrahedral amorphous carbon. J. Appl. Phys. 85, 7191–7197 (1999) Zhang, S., Sun, D., Fu, Y., Du, H.: Toughness measurement of thin films: a critical review. Surf. Coat. Technol. 198, 74–84 (2005) Scharf, T.W., Singer, I.L.: Role of the transfer film on the friction and wear of metal carbide reinforced amorphous carbon coatings during run-in. Tribol. Lett. 36, 43–53 (2009) Park, S.J., Lee, K.R., Ko, D.H.: Tribochemical reaction of hydrogenated diamond-like carbon films: a clue to understand the environmental dependence. Tribol. Int. 37, 913–921 (2004) Tarassov, M.P., Marinov, M.S., Konstantinov, L.L., Zotov, N.S.: Raman spectroscopy of ferritungstite: experimental and model spectra. Phys. Chem. Miner. 21, 63–66 (1994) Baserga, A., Russo, V., Fonzo, F.D., Bailini, A., Cattaneo, D., Casari, C.S., Bassi, A.L., Bottani, C.E.: Nanostructured tungsten oxide with controlled properties: synthesis and Raman characterization. Thin Solid Films 515, 6465–6469 (2007) Gharam, A.A., Lukitsch, M.J., Balogh, M.P., Irish, N., Alpas, A.T.: High temperature tribological behavior of W-DLC against aluminum. Surf. Coat. Technol. 206, 1905–1912 (2011) Chu, P.K., Li, L.: Characterization of amorphous and nanocrystalline carbon films. Mater. Chem. Phys. 96, 253–277 (2006) Mrabet, S.E., Abad, M.D., Sa´nchez-Lo´pez, J.C.: Identification of the wear mechanism on WC/C nanostructured coatings. Surf. Coat. Technol. 206, 1913–1920 (2011) Harlin, P., Carlsson, P., Bexell, U., Olsson, M.: Influence of surface roughness of PVD coatings on tribological performance in sliding contacts. Surf. Coat. Technol. 201, 4253–4259 (2006) Konicek, A.R., Grierson, D.S., Sumant, A.V., Friedmann, T.A., Sullivan, J.P., Gilbert, P.U.P.A., Sawyer, W.G., Carpick, W.: Influence of surface passivation on the friction and wear behavior of ultrananocrystalline diamond and tetrahedral amorphous carbon thin films. Phys. Rev. B 85, 155448 (2012) Yang, S.H., Kong, H., Lee, K.R., Park, S., Kim, D.E.: Effect of environment on the tribological behavior of Si-incorporated diamond-like carbon films. Wear 252, 70–79 (2002)