- Email: [email protected]
Effect of tribochemistry on friction behavior of fluorinated amorphous carbon films against aluminum
Surface & Coatings Technology 304 (2016) 150–159
Contents lists available at ScienceDirect
Surface & Coatings Technology journal homepage: www.elsevier.com/locate/surfcoat
Effect of tribochemistry on friction behavior of fluorinated amorphous carbon films against aluminum Fu Wang a,c, Zhibin Lu a, Liping Wang a,b,⁎, Guangan Zhang a,⁎⁎, Qunji Xue a a b c
State Key Laboratory of Solid Lubrication, Lanzhou Institute of Chemical Physics, Chinese Academy of Sciences, Lanzhou 730000, PR China Key Laboratory of Marine Materials and Related Technologies, Ningbo Institute of Materials Technology and Engineering, Chinese Academy of Sciences, Ningbo 315201, PR China University of Chinese Academy of Sciences, Beijing 100039, PR China
a r t i c l e
i n f o
Article history: Received 21 February 2016 Revised 29 June 2016 Accepted in revised form 30 June 2016 Available online 1 July 2016 Keywords: a-C:H:F films Aluminum F content Friction Tribo-layer Tribochemistry
a b s t r a c t Fluorinated amorphous carbon (a-C:H:F) films synthesized from C2H2 and CF4 were examined by Raman spectra and X-ray photoelectron spectroscopy (XPS). Their tribological properties were tested against aluminum balls in dry sliding. The a-C:H:F films with low F content showed lower friction coefficient than hydrogenated amorphous carbon (a-C:H) film, whereas high F content in films resulted in a significant friction increase. Contact surfaces were analyzed in detail to elucidate the possible sliding mechanism. Results indicated that the friction behavior was closely related to the nature of a composite-like tribo-layer consisting of Al compounds and carbon components formed on Al ball, relying on the tribochemical processes of contact interface. The accumulated F atoms on tribo-layer reduced the adhesion across sliding interface because of strong repulsion between F atoms, and thereby lowered the friction of a-C:H:F films. However, with increasing F contents in films, the enhanced tribochemical reaction between Al and F caused crack, delamination and fragmentation of the tribolayer, and then a marked abrasive process at the sliding interface. Consequently, the increased shearing and abrasive actions strongly opposed the contribution of the reduced adhesion to friction, and result in a rather high friction of highly fluorinated a-C:H:F films. © 2016 Elsevier B.V. All rights reserved.
1. Introduction The use of lubricating coatings greatly alleviates aluminum adhesion and transfer to tools, a very pressing problem in Al alloys dry machining [1]. Diamond-like carbon (DLC) films, with excellent tribological properties, have shown better anti-adhesive effect and lower friction coefficient than other conventional TiN, ZrN, TiAlN and TiCN coatings when tested against Al [2–5]. This has greatly stimulated interest in using them as promising tool coatings for Al dry machining. However, to improve the performance of DLC films, lower friction in DLC/Al contact is a continual drive for optimizing and designing DLC films. Thus, it is quite necessary to probe the detailed friction mechanism of various DLC films against Al. Many studies have investigated the friction behavior and mechanism of DLC films and their modifications against Al under different conditions. Non-hydrogenated DLC films exhibit high friction coefficient (μ = 0.5–0.8) under high vacuum and inert (He, Ar and N2) environments, where the adhesive transfer of Al to DLC surface results in an undesirable Al\\Al contact [6,7]. On the contrary, low friction (μ b 0.2) is obtained as the introduction of H2 and H2O that passivate the ⁎ Correspondence to: L. Wang, State Key Laboratory of Solid Lubrication, Lanzhou Institute of Chemical Physics, Chinese Academy of Sciences, Lanzhou 730000, PR China. ⁎⁎ Corresponding author. E-mail addresses: [email protected] (L. Wang), [email protected] (G. Zhang).
http://dx.doi.org/10.1016/j.surfcoat.2016.06.087 0257-8972/© 2016 Elsevier B.V. All rights reserved.
unoccupied C\\σ bonds on carbon surface [6–8]. Oxygen results in a high friction (μ = 0.28–0.44) in DLC/Al contact due to the oxidation of Al ball and carbon surface, but no Al adhesion transfer was observed. [6,9,10]. Hydrogenated DLC films exhibited lower friction coefficient than H-free DLC at evaluated temperature and showed lower cutting forces and less adhesion, attributed to higher chemical inertness of Hcontaining DLC network. [11,12]. The a-C:H/a-Si:O coating showed low steady-state friction coefficient up to 400 °C, where a Si-containing carbon-based transfer layer terminated by H, O and OH was identified on Al surface [13]. Similarly, a nano-structural carbon tribo-layer terminated by H, HO and F on Al surface accounted for the low friction behavior of a-C:H:Si:O:F coating in both vacuum and ambient atmosphere [14]. In addition, the formation of WO3 on Al surface dominated the low friction behavior of W-DLC/Al at 400 °C [15]. Briefly, the passivation of free C\\σ bonds and the formation of tribo-layer on Al surface are fundamental to the achievement of low friction in DLC/Al contact. Fluorinated amorphous carbon (a-C:H:F) films have been investigated as anti-sticking coatings in nanoimprint lithography and biomedical applications because of their low surface energy property [16–18]. Theoretical calculations reveal that fluorinated carbon surfaces display lower adhesion tendency when facing each other [19–21]. Meanwhile, Donnet et al. [22] found that a-C:H:F films showed comparable tribological properties to non-fluorinated ones when F content is b20 at.%. Interestingly, when tested against Al counterparts, Sen et al. [23] recently found that FDLC film showed lower friction coefficient than HDLC film.
F. Wang et al. / Surface & Coatings Technology 304 (2016) 150–159
They pointed out that a fluorinated transfer layer on Al surface may reduce the friction level of FDLC/Al contact. This suggests a potential use of a-C:H:F films in Al anti-adhesion. However, this meaningful result was only observed and discussed for a certain FDLC film (3 at.% F, 25 at.% H) in ambient air, many issues are still open. Most importantly, considering the large electro-negativity of F and fluorinated tribo-layer, aC:H:F films perhaps signify more complicated tribochemical processes, especially with increasing F contents, which possibly markedly influence the friction behavior of a-C:H:F against Al. So, it is of great significance to know the relevant details, which can offer more complete information to further consider possible use of a-C:H:F films, e.g. in Al dry machining. The present work, accordingly, focuses on the influence of F content and tribochemistry on the friction behavior of a-C:H:F against Al. The aC:H:F films were produced from the glow plasma decomposition of C2H2 and CF4 mixture. The fundamental nature of films was characterized and understood. The friction behavior of films was assessed under dry contact conditions. The contact surfaces were examined to identify the factors that influenced the friction behavior of a-C:H:F films against Al.
151
friction tests were also carried out under dry N2 (RH: 1.0%), dry air (RH: 5.0%) and humid air (RH: 60.0%) to consider environmental effect. 2.3. Characterization methods Micro-Raman spectra were recorded by a LabRam HR800 JobinYvon spectrometer with an excitation wavelength of 532 nm. The hardness (H) and elastic modulus (E) of films were determined by a nanoindenter Hysitron TI 950 USA. The residual stress of films was determined through the deformation of Si wafer caused by the deposition of film (SuPro Instruments, FST-1000). Surface compositions and chemical states of samples were analyzed by an ESCALAB 250Xi (Thermo Fisher, USA) X-ray photoelectron spectroscopy (XPS) system with an Al Kα (1486.7 eV) X-ray source. A DSA 100 goniometer was employed to measured water contact angles. Surface morphology of wear tracks was acquired by 3D Laser Scanning Microscope (Keyence VK-X100/ X200, Japan). The transfer layers were observed using a JSM-5600LV scanning electron microscope (SEM) equipped with an energy-dispersive X-ray spectroscopy (EDS). 3. Results and discussion
2. Experimental procedure 3.1. Characterization of films 2.1. Materials The a-C:H:F films were deposited on silicon (110) and mirrorpolished stainless steel plats (roughness: Ra b 20 nm) by plasma enhanced chemical vapor deposition (PECVD) system, using acetylene (C2H2) and carbon tetrafluoride (CF4) as precursor gases. The base pressure of deposition chamber was better than 2 × 10− 3 Pa. Solvent cleaning and Ar+ etching procedures were performed to remove contaminants on substrate surfaces before deposition process. A thin aSi:H interlayer (~80 nm) and a-C:Si:H gradient layer (~500 nm) were produced to improve the adhesion of film to substrates. The a-C:H:F layer was synthesized from C2H2 and CF4 at a pressure of 6.5 Pa. A pulse power source was used (frequency, 1.5 kHz; duty ratio, 30%). The gas flow ratio (C2H2/CF4) and the corresponding sample are listed in Table. 1. The a-C:H (about 10 at.% H) and a-C:H:F-6 films were deposited at substrate voltage − 0.8 kV and − 1.0 kV, respectively, and the other films were produced at − 0.9 kV. The total thickness of films was about 2.0 μm (measured by SEM).
The micro-structure of films was investigated by Raman spectroscopy. As shown in Fig. 1, the broad, asymmetrical Raman peaks between 1000 and 1700 cm−1 suggested disordered carbon structures. According to recent publications by Neuville [24,25], the spectra can be resolved into two sub-bands: the so-called D band at 1340/1398 cm−1 and C5/C7 odd ring band at 1534/1552 cm−1. The so-called D band in fact corresponds to either disordered diamond (1330 cm−1) or to the G Aedge vibration modes of sp2 clusters (1350 cm−1). For a-C:H film, the D band at 1340 cm−1 mainly corresponded to disordered Csp3-Csp3 clusters. The C5/C7 band at 1534 cm−1 indicated that there were no significant hexagonal cyclic sp2 rings within film, which further signified that the D band at 1340 cm−1 could only encapsulate weak G Aedge
2.2. Friction and wear tests The sliding experiments were performed using a CSM tribometer and a ball-on-disk contact geometry with a reciprocating motion under a load of 3.0 N, a speed of 0.05 m/s (frequency: 5.0 Hz, stroke length: 2.5 mm). The counterpart was commercial purity (N 99%) aluminum ball with a diameter of 4.0 mm, Vickers hardness of about 270 MPa and the surface roughness, Ra, of ~60 nm. Aluminum balls were tested against films deposited on steel substrates in ambient air with a relative humidity (RH) of 30 ± 2.5% and halted after 20,000 laps, corresponding to a sliding distance of 200 m. The cross-sectional area of the track was determined by a profile-meter (KLA-Tencor, D-100). Furthermore, Table 1 Source gas ratio, F content, hardness (H), elastic modulus (E), internal stress (σ) and water contact angle (θ) of as-deposited films. Sample
C2H2/CF4/sccm
F/at.%
H/GPa
E/GPa
σ/GPa
θ/°
a-C:H a-C:H:F-1 a-C:H:F-2 a-C:H:F-3 a-C:H:F-4 a-C:H:F-5 a-C:H:F-6
120/0 60/20 60/40 60/60 50/70 40/80 30/90
0 1.7 3.4 6.5 9.6 16.7 18.6
14.8 15.4 14.6 13.4 12.9 11.8 10.3
127.8 131.5 134.4 122.7 116.3 99.1 89.4
−0.85 −0.70 −0.61 −0.43 −0.58 −0.41 −0.41
71.8 ± 0.7 78.3 ± 0.7 77.2 ± 1.5 78.0 ± 0.7 74.7 ± 0.8 81.7 ± 1.3 81.6 ± 2.4
Fig. 1. Raman spectra of a-C:H and a-C:H:F films.
152
F. Wang et al. / Surface & Coatings Technology 304 (2016) 150–159
band at ~1350 cm−1. This suggests that a-C:H film is the disordered sp3/ sp2 carbon matrix which can contain amorphous Csp3-Csp3 clusters. However, for a-C:H:F films, the D bands were significantly upshifted (above 1360 cm−1) and broadened. First, the higher deposition voltage of a-C:H:F films might upshift the observed D band because of possibly increased graphenic sp2 rings or sp2 clusters, which increased the G Aedge mode of sp2 clusters (~1350 cm−1). Second, the residual compressive stress in film (Table 1) upshifted the Raman frequencies of films to some extent. Third, owing to large electron-negativity of F atom, the formed C\\F bonds (stronger than C\\H bonds) resulted in the non-uniform distribution of bond angles, bond lengths and bond energies of C\\C bonds, and made them randomly distributed over a larger range around some nominal values. Thus, the corresponding Raman frequencies were also distributed over a broader Raman band. This explains why the D band frequencies of a-C:H:F films increased with increasing F contents, and also gives account for the D band broadening of a-C:H:F films as the F content. The broadening of D band in return reflected the increased atomic disorder of a-C:H:F films. Accordingly, the a-C:H:F films showed slightly increased sp2 clusters and increased atomic disorder compared to a-C:H film. By and large, agreeing with the micro-structural changes, both hardness and residual compressive stress of a-C:H:F films showed a decreasing tendency (Table 1). The surface F content of a-C:H:F films was determined by semi-quantitative XPS analysis, in the range of 1.7–18.6 at.% (Table 1). The measured water contact angles (Table 1) increased as the incorporation of F, indicating low surface energy property of a-C:H:F films. Fig. 2 shows high resolution XPS spectra of a-C:H:F films. All spectra were calibrated by fixing the C\\C/C\\H peaks at 284.8 eV. The C 1s spectra (Fig. 2a) appeared two shoulder peaks at ~ 287.5 and ~ 289.5 eV (C\\F group), which gradually enhanced with increasing F contents, while the ratio of C-C/C-H peaks dwindled. Together with the F 1s peaks (Fig. 2b), this well demonstrated the different fluorination degrees in individual films. With the increase of F content, both the C 1s shoulder peak and F 1s peak were slightly shifted towards higher binding energy. This could be attributed to the carbon matrix with a more fluorinated bonding environment, which increased their bonding energy because of the high electron-negativity of F atom [26]. In short, despite different fluorination degrees, F atoms were mainly bonded to carbon matrix by mono-fluorinated carbon atom (C\\F group). 3.2. Tribological behavior of a-C:H:F films against Al balls Fig. 3 shows the evolution of friction coefficient of films with sliding laps in ambient air (RH = 30%). For comparison, a-C:H film was also tested under the same conditions. From the plot, friction coefficient of films showed a slightly decreasing trend with the incorporation of a small amount of F (a-C:H:F-1, -2, -3), and the a-C:H:F-2 film with
3.4 at.% F exhibited the lowest friction value, in good agreement with the report by Sen et al. [23]. However, friction coefficient of films significantly increased with further increase in F content (a-C:H:F-4, -5, -6). The average steady-state friction coefficients of films calculated from 5000 to 20,000 laps were 0.20, 0.18, 0.21, 0.28, 0.32, and 0.47 for aC:H:F-1 to a-C:H:F-6, respectively, and 0.23 for a-C:H film. Therefore, it seemed that the reduction of friction was only achieved for lowly fluorinated films (≤6 at.% F). Note that all the friction curves (Fig. 3) were characterized by a rapid friction drop at the beginning of sliding, followed by a gradual increase and high steady-state value. This differed from the frequently observed cases, where a gradual decrease in friction level represents the running-in process [27], attributed to surface finish of contact bodies and the formation of transfer layer on the counterpart. Actually, similar behavior can also be seen in report by Sen et al. [23]. It was evident that the difference in hardness, hydrophobicity (Table 1) and bonding configuration of films (Figs. 1 and 2) cannot explain the observed variations in friction coefficient because of their inconsistent trends. For hard, amorphous carbon films deposited on smooth substrates, the adhesive, tribochemical and third-body interactions are the main factors influencing their friction behavior [28]. On account of high reactivity of Al towards O2 in air and F in film, it is considered that tribochemistry at the sliding interface could dominate the observed friction behavior. The wear rates of films were calculated by the wear volume of films divided by the product of normal load and sliding distance. Previous work by Konca et al. showed that 319 Al can cause more wear of the DLC films than WC and sapphire because of the formation of Al oxide abrasives at the sliding interface [29]. In current investigation, the wear rates of a-C:H:F films were of the same order of magnitude as that of a-C:H film (b 7 × 10−7 mm3/Nm). However, highly fluorinated a-C:H:F-6 film showed the highest friction coefficient but the lowest wear rate (1.8 × 10−7 mm3/Nm). This might be related to its different deposition process or other unclear reasons. Furthermore, the addition of F to films increased the wear of Al balls, as shown in Fig. 4. 3.3. Tribo-layer and wear track The tribological behavior of DLC films under dry sliding contact is normally governed by a carbon-rich transfer layer formed on counterpart surfaces, which prevents the direct contact between DLC films and counterpart materials and establishes a new sliding interface [30– 32]. To understand the friction behavior of a-C:H:F films, the contact surfaces of Al balls were firstly observed after sliding against a-C:H (Ffree), a-C:H:F-2 (low F) and a-C:H:F-5 (high F) films. Fig. 4 shows their SEM images after sliding 20,000 laps in ambient air. Obviously, a tribo-layer or transfer layer was established on Al ball surfaces for three films. When tested against a-C:H film, an integral and uniform tribo-layer was formed (Fig. 4a and b). Some fine cracks distributed
Fig. 2. High resolution XPS spectra of a-C:H:F films (a) C 1s and (b) F 1s.
F. Wang et al. / Surface & Coatings Technology 304 (2016) 150–159
Fig. 3. Friction curves of a-C:H and a-C:H:F films against Al balls in ambient air.
on the edge of the tribo-layer as sliding against a-C:H:F-2 film (Fig. 4c and d). However, when sliding contact with a-C:H:F-5 film, the tribolayer showed an apparent cracked appearance (Fig. 4e and f). This suggests that the incorporation of F can dramatically change the nature of tribo-layer. The easy shear and chemical states of the tribo-layer are essential to achieve low friction of DLC films. In the light of the fragmentation behavior in Fig. 4e and f, tribo-layer against a-C:H:F-5 film seemed to
153
have higher brittleness than those against a-C:H and a-C:H:F-2 films, which might be mainly associated with their chemical compositions. Although without a quantitative description of the mechanical properties of these layers, the high brittleness possibly suggested the increased shear resistance and hence increased friction. The cohesion of the tribo-layers and its adhesion to Al surface could affect their changes in SEM images. The formation of AlF3 at Al/tribo-layer interface had been confirmed by Sen et al. [14,23]. In terms of the formation of the tribolayer, the formed AlF3 on Al surface may prevent the chemical binding of carbon to Al surface as a result of surface termination, while the increased proportion of Al fluorides in tribo-layer would decrease the cross-link degree of the layer. With the increase of F content in film, more AlF3 was grown on Al surface due to the increased reaction rate between F and Al, which might markedly affect the adhesion strength of tribo-layer to Al surface, and also decrease the internal cross-linking density of the layer. Therefore, the tribo-layer formed from highly fluorinated a-C:H:F films would easily crack, delaminate and fragment into debris under continuous mechanical actions, some of which acted as abrasive particles and led to high friction. Fig. 5 presents the optical observation of the wear tracks formed on different films. Compared to a-C:H film, an obvious difference was that the sliding tracks on a-C:H:F films became wider. This can be associated with the tribochemical reactions between Al and F, which accelerated the wear of Al balls and hence increased the contact area. As shown in Fig. 4, despite decreasing hardness of films, the size of the wear scar increased with increasing F contents. Meanwhile, the low friction
Fig. 4. SEM images of the wear scars on Al ball surfaces after sliding against (a) a-C:H, (b) a-C:H:F-2 and (c) a-C:H:F-5 films in ambient air, and their local images (d), (e) and (f).
154
F. Wang et al. / Surface & Coatings Technology 304 (2016) 150–159
coefficient for a-C:H and a-C:H:F-2 films corresponded to (see in Fig. 3) a relatively smooth surface of wear tracks (Fig. 5a and b), whereas the high friction coefficient for a-C:H:F-5 and a-C:H:F-6 films corresponded to the rough wear track surfaces (Fig. 5c and d). In particular, the severer plowings grooves were clearly observed for a-C:H:F-6 film with the highest F content (Fig. 5d). This suggested that there were more hard wear particles introduced to the sliding interface when sliding against highly fluorinated films. Owing to the susceptibility of Al to O2, it is believed that the oxidation of Al surface and/or Al2O3 abrasives significantly affect the friction and wear behavior of non-hydrogenated DLC film [29]. While in a-C:H:F films the large electron-negativity difference between F (3.98) and Al (1.61) also facilitates the formation of Al fluorides. It was thereby thought that stable aluminum oxides (Al2O3) and fluorides at sliding interface could act as hard abrasives that contribute to high friction, and more abrasive particles were introduced into the sliding interface as the F content increased. 3.4. Chemical analyses of sliding interfaces 3.4.1. XPS analysis of tribo-layers and wear tracks The tribo-layers against a-C:H, a-C:H:F-2, and a-C:H:F-5 films were examined by XPS to acquire their surface chemistry. Fig. 6a shows their full range XPS spectra and chemical compositions. The tribolayer from a-C:H film consisted of C, O and Al, while F also became one of the major components of the tribo-layer against a-C:H:F films. Fluorine was clearly accumulated on the tribo-layer surfaces for two a-C:H:F films (Fig. 6a). In particular, when sliding against a-C:H:F-2 film, the tribo-layer had four-fold higher F concentration (16.0 at.%) compared to 3.4 at.% F in film. Moreover, it was clear that Al was an important component of all the tribo-layers, and its content increased as the addition of F to film. This might be associated with the increased wear of Al balls when tested against a-C:H:F films. To further address the chemical difference of the tribo-layers, high resolution XPS spectra of C 1s, F 1s and Al 2p are presented in Fig. 6b– d. The C 1s spectra in Fig. 6b mainly consisted of three peaks at 284.8, 286.4 and 288.7 eV, which are assigned to the C\\C/C\\H, C\\OH/ C\\O\\C and O\\C_O groups, respectively. But C atoms became less oxidized with the increase of F contents within films. The
hydrophobicity of a-C:H:F films could reduce tribo-oxidation of carbon atoms caused by absorbed water molecules. In the case of a-C:H:F-2 and a-C:H:F-5 films, the peaks of 287.5 and 289.3 eV were also identified and assigned to the C\\F groups, but their low ratios signified that there was only a small portion (~5%) of C atoms fluorinated on tribo-layers. This well certified the break of C\\F bonds during rubbing, and was further demonstrated by the lower F 1s binding energy of tribo-layer than that of film (Fig. 6d). An unidentified peak at ~ 293.3 eV was also observed for all three tribo-layers. As a consequence, for a-C:H:F films, the majority of the accumulated F atoms on tribo-layer were not chemically bound to carbon skeleton to create a fluorinated carbon surface, which is very beneficial for achieving the low friction of a-C:H:F films because of repulsive interaction between two fluorinated carbon surfaces [19,20]. High resolution Al 2p spectra (Fig. 6c) appeared at 74.5, 74.9 and 75.4 eV, corresponding to a-C:H, a-C:H:F-2 and a-C:H:F-5 respectively. This indicated that the Al 2p binding energy of tribo-layer increased as increasing F contents in film. For a-C:H film, the Al 2p peak of 74.5 eV suggested the presence of Al2O3 phase on the surface of the tribolayer. However, for a-C:H:F films, the Al 2p binding energy of tribolayers was higher than that of Al2O3 and slightly increased with the F content, but lower than that of AlF3 (76.3 eV). This could represent a mixture of Al oxides, Al fluorides and Al oxyfluorides (AlOxFy). The binding energies of Al 2p and F 1s can be shifted to a lower level as the substitution of F atom in AlF3 by O atom [33,34]. Consequently, the tribolayer on Al ball surfaces seemed to be a composite layer including carbon and complex Al compound components, which dominated the sliding behavior of films against Al. Note that the composition and nature of the layer varied as film chemistry. As shown in Fig. 6a, the increase of F content in film induced more Al compounds in tribo-layer. This could result in the increased brittleness and the decreased cohesion of tribolayer, and hence increase the possibility of formation of the Al-containing wear particles. The wear track regions on a-C:H, a-C:H:F-2 and a-C:H:F-5 were also examined using XPS after sliding tests. As seen in Fig. 7a, the chemical composition inside and outside the wear track on a-C:H:F-2 film was quite similar, whereas the wear track on a-C:H:F-5 film contained lower F concentration (10.7 at.%) compared to the unworn region
Fig. 5. Optical images of the wear tracks formed on (a) a-C:H, (b) a-C:H:F-2, (c) a-C:H:F-5 and (d) a-C:H:F-6 films after 20,000 laps in ambient air.
F. Wang et al. / Surface & Coatings Technology 304 (2016) 150–159
155
Fig. 6. (a) Full range XPS spectra and composition of tribo-layers against a-C:H, a-C:H:F-2 and a-C:H:F-5 films, their high resolution XPS spectra of (b) C 1s, (c) Al 2p and (d) F 1s. Film in (d) represents as-deposited a-C:H:F-5 film.
(16.7 at.% F). Such variation in F concentration for a-C:H:F-5 film may largely be associated with the inherent nature of a-C:H:F films, and not a result of friction. When the growth process of film ends, more passivation atoms are required to form a stable and low energy surface. Thus, for high CF4 ratio in source gas, F atom controls this process, whereas for high C2H2 ratio, H atom dominates the same process. This suggests that higher F concentration can be observed on the surface of highly fluorinated a-C:H:F films. Sliding contact might not induce a significant change in chemical states of carbon-based network. With regard to Al, a weak Al 2p signal was only detected on the surface of the
wear track on a-C:H:F-5 film (Fig. 7b). This indicated that more Al-containing particles possibly detached from the tribo-layer and acted as abrasive particles, which actively supported the observation in Figs. 3–5. 3.4.2. EDS mapping of wear particles In DLC/steel contact, the iron-rich debris is closely related to their relatively high friction value (μ ≈ 0.15) in ambient air [35,36]. The wear debris aside the wear tracks on a-C:H and a-C:H:F-5 were also examined by SEM. Fig. 8 shows their SEM images and EDS elemental mappings. The more scattered and larger particles were observed for a-
Fig. 7. (a) Full range XPS spectra of wear tracks and films for a-C:H:F-2 and a-C:H:F-5 films, (b) high resolution Al 2p spectra of wear tracks on a-C:H, a-C:H:F-2 and a-C:H:F-5 films.
156
F. Wang et al. / Surface & Coatings Technology 304 (2016) 150–159
C:H:F-5 (Fig. 8d). The mappings of O and Al suggested Al-containing granular or powdery debris, possible Al oxides debris. Although the F and C mappings did not give available information on wear debris (data not shown), it was deduced that F was also contained in wear particles for a-C:H:F films because of their fluorinated tribo-layer with Al fluorides. Together with the grooves in Fig. 5c and d, it seemed that these Al-containing particles, as the third-body abrasive, caused the high friction and high F content in film aggravated this process. 3.4.3. Raman spectra of tribo-layers Normally, a graphitized carbon layer formed on the counterpart is in favor of the low friction and wear behavior of DLC films [31,37]. XPS results in Fig. 6a and b indicated that carbon was still an important component of the tribo-layers and formed various organic groups, including alcohol, ether, and carboxyl. Their microstructure was further identified at a larger scale by Raman spectra. As shown in Fig. 9, however, all tribo-layers did not show a frequently observed ‘graphite-like’ feature (separated D peak and G peak), only very weak amorphous carbon signal between 1000 and 1800 cm−1 for a-C:H and a-C:H:F-5 films. Thereby, carbon in tribo-layer might be organized in either non-continuous tiny structures or polymer-like chain structures terminated by\\H, \\OH, \\O, and\\F. The formation of Al compounds and the catalytic degradation of carbon network by formed Al compounds (Al2O3 or AlF3) may partly explain such structures [38,39]. Although there was no continuous carbon-based transfer layer formed, these carbon species in a composite-like tribo-layer might increase the crosslink of the layer, and also effectively prevented the adhesive transfer of Al to film surface. In contrast, complex Al compounds likely increased the shear resistance of tribo-layer and induced the generation of hard abrasive particles. Therefore, when sliding contact with Al counterpart, DLC films exhibit higher friction coefficient compared with other mating materials, e.g. Al2O3, ZrO2, SiC, and Si3N4 [40].
Fig. 9. Raman spectra of tribo-layers against a-C:H, a-C:H:F-2 and a-C:H:F-5 films.
testing under ambient conditions did not induce an effective carbon-based transfer layer on Al surface [40], in accord with the Raman observation in Fig. 9. This may mainly be because that: a) the H and OH passivation of carbon surface prevents the generation of free C\\σ bonds; b) the growth of oxides on Al surface reduces the exposure of active Al sites (ΔGo(Al2O3) = − 1690 kJ/mol); c) soft Al tends to be tore and transferred to the surface of hard DLC [7,10,41]; d) the formed C\\Al bonds are easily decomposed into C\\H and Al\\O by water molecules, as listed in formulas (2)–(4). C\H=C\C=C\O þ Al→C\Al=C\O⋯Al
ð1Þ
3.5. Tribochemistry at sliding interfaces
C\C=C\O þ H2 O→C\H=C\OH
ð2Þ
The foregoing observations indicated that complex chemical processes occurred during rubbing, and significantly changed the nature of tribo-layer with increasing F contents. In terms of surface carbon, the generation of free C\\σ bonds can promote the formation of C\\Al bonds, which is highly conducive to the adhesion of carbon layer to Al surface. The C\\O bonds can also increase the adhesion of carbon layer as the formation of C\\O\\Al bonds (in formula (1)). Nevertheless, even if a relatively strong bonding between Al and free C\\σ bonds (the standard Gibbs free energy of formation: ΔGo(Al4C3) = −218 kJ/mol),
Al þ O2 =C\O→Al oxides
ð3Þ
C\Al þ H2 O→C\H þ Al oxides
ð4Þ
Apart from these possible processes, for a-C:H:F films, the large electron-negativity difference between F (3.98) and Al (1.61) facilitates the production of Al fluorides (ΔGo(AlF3) = −1530 kJ/mol). According to TEM evidence by Sen et al. [23], an AlF3 layer, rather Al2O3 layer, is formed at Al/tribo-layer interface in FDLC/Al sliding contact. This
Fig. 8. SEM images of wear debris aside the wear tracks on (a) a-C:H and (d) a-C:H:F-5, and element mappings: (b), (e) for O and (c), (f) for Al.
F. Wang et al. / Surface & Coatings Technology 304 (2016) 150–159
suggests that although Al readily oxidizes under ambient conditions, the formed oxide film on its surface is removed during rubbing. It is also indicated that the reaction of Al with F is kinetically supported because of the direct contact between Al and F (formula (5)). As depicted in Section 3.3, the formed AlF3 affected the bonding strength of tribo-layer to Al surface. Low F content, to some extent, might also promote the formation of C\\Al bonds because of the cleavage of C\\F bonds and nascent Al surface (formula (5)), whereas high F content markedly weakened the adhesion of tribo-layer and lead to a cracked tribo-layer with more Al fluorides formed (Fig. 4e and f). In addition, AlF3, a strong Lewis acid, is readily coordinated by alkoxy groups (C\\O) and H2O molecules to form C\\O…AlF3 and AlF3·xH2O. The former might promote the attachment of carbon component, but the latter might cause the detachment of AlF3 layer and could be transformed into AlOxFy under mechanical actions and friction heat (formula (7)). Therefore, oxygen-rich AlOxFy could be formed for lowly fluorinated a-C:H:F films, while fluorine-rich AlOxFy was formed for highly fluorinated a-C:H:F films, as with the XPS spectra in Fig. 6c and d. C\F þ Al→C\Al þ AlF3
ð5Þ
AlF3 þ C\O→C\O…AlF3
ð6Þ
AlF3 þ xH2 O→AlF3 xH2 O→Al ðOHÞx Fy →AlOx Fy
ð7Þ
3.6. Proposed sliding mechanism for a-C:H:F/Al contact Given that the a-C:H:F films were deposited from a carbonhydrogen-fluorine system, it is though that diamond nano-crystals can be facilitated during deposition because of thermal activation involving fluorine [24]. Thus, it should be considered that formed diamond nanocrystals might cause the increase of friction with increasing F contents. However, on the basis of the Raman spectra and hardness data of films, either the increased sp2 clusters or the decreased hardness of aC:H:F films disapproved of the increase of possible diamond nanocrystals with the F content. This suggested that the sliding behavior of a-C:H:F films in contact with Al was mainly associated with the
157
passivation of surface C\\σ bonds and the interfacial tribochemistry, rather than diamond nano-structure. The saturation of free C\\σ bonds can eliminate strong covalent adhesion between contact carbon surfaces [42]. The water molecules in ambient air efficiently passivated the surface free C\\σ bonds and formed a low-adhesive sliding surface. And hence the adhesion transfer of Al was not observed for all tests. As shown in Fig. 3, the initially low friction of films might be attributed to the H and OH passivation of dangling C\\σ bonds, but the subsequent increase in friction suggested that some detrimental factors appeared with continuous sliding. This may correspond to the formation of a specific tribo-layer on Al surface. On the basis of the Raman observation (Fig. 9) and the formation of Al compounds (Fig. 6), a composite-like tribo-layer including carbon structures and Al compounds may largely cause such increase in friction. The addition of F clearly induced different tribochemical processes and altered the nature of tribo-layer on Al surface. In comparison with a-C:H film, the friction behavior of a-C:H:F films exhibited two opposite trends. One is the friction reduction for films with low F content (b6 at.%) and another is the significant friction increment for films with high F content (N 9 at.%). For the first case, it is proposed that the repulsive force arising from fluorinated transfer layer including C\\F groups and AlF3 is responsible for the observed friction reduction [23]. However, based on current observations, different friction mechanisms should be considered for a-C:H:F/Al contact. According to the discussion of Fontaine [43], there are three main components contributing to the friction coefficient of DLC films, μ = μabrasion + μshearing + μadhesion, where μabrasion, μshearing and μadhesion represent the abrasive, shearing and adhesive component of friction coefficient, respectively. For most DLC sliding systems, a carbon-based transfer layer well avoids the generation of third-body abrasives and offers low shearing resistance [43]. Nevertheless, for the a-C:H:F/Al contact, for one thing, the accumulated F atoms on a composite-like tribo-layer reduced interface adhesion because of repulsion between F atoms, and thereby reduced the adhesive component of friction, μadhesion [19,20,44]. For another, the presence of Al and F in tribo-layer decreased the crosslink density of the layer and increased the shear resistance of the layer (Fig. 4), and their contents in tribo-layer increased with the increase of F content in film (Fig. 4). In the meantime, the increased Al fluorides at Al/tribo-layer interface weakened the adhesion of tribo-layer to Al
Fig. 10. Scheme of sliding interfaces for a-C:H and a-C:H:F films against Al balls in ambient air, (a) lowly fluorinated a-C:H:F films, (b) a-C:H films and (c) highly fluorinated a-C:H:F films.
158
F. Wang et al. / Surface & Coatings Technology 304 (2016) 150–159
Fig. 11. Friction curves of (a) a-C:H and (b) a-C:H:F-2 films against Al balls under dry N2 (RH: 1.0%), dry air (RH: 5.0%) and humid air (RH: 60.0%).
surface. Therefore, the tribo-layer derived from highly fluorinated films easily cracks, delaminates, and fragments into third-body abrasives due to the enhanced tribochemical processes. As a result, lowly fluorinated a-C:H:F films did not markedly change the μabrasion and μshearing components when compared to the a-C:H film, conversely, the reduced μadhesion lowered friction coefficient. However, highly fluorinated aC:H:F films enhanced tribochemical reaction of Al with F, which generated more Al compounds, especially complex Al fluorides. This promoted the production of third-body abrasives, and so the increased μabrasion and μshearing greatly raised the friction coefficient. As shown in Fig. 10, a proposed scheme depicts the interfacial behavior of a-C:H and a-C:H:F films sliding against Al. Relying on the F content of films (i.e. tribochemical process), the reduced μshearing component and the increased μabrasion and μshearing components dominate competitively the friction response of a-C:H:F/Al contact. Considering the susceptibility of tribochemical process to environments, additional sliding experiments were carried out under different atmosphere. As seen in Fig. 11, when tests were performed under dry N2 (RH: 1.0%), dry air (RH: 5.0%) and humid air (RH: 60.0%), the aC:H:F-2 film did not exhibit a lower friction value than a-C:H film. Accordingly, although the lowly a-C:H:F films can achieve a lower friction coefficient under ambient conditions when tested against Al, the tribochemical process involving F may restrict such friction merit in varied environments, e.g. Al alloys dry machining. Nevertheless, notably, the build-up of the fluorinated tribo-layer, suggesting a low adhesive interface, may be of great interest for tribological applications of a-C:H:F films. At present, the use of Al mating material may not quite fulfil the prerequisites to form a fluorinated carbon-based transfer because of its high softness, high affinity towards F and relatively weak affinity towards C. But when mating materials are those strong carbide formers (e.g. WC, Ti, SiC, Si3N4), a fluorinated carbon transfer layer can be expected to provide excellent tribological properties due to strong repulsive forces between two fluorinated carbon surfaces. Then the optimization of composition and structure of fluorinated carbon-based network would become crucial for their tribological applications [45, 46]. 4. Conclusions The a-C:H:F films synthesized from C2H2 and CF4 plasma were slid against Al balls to investigated the effect of F content on their friction behavior. Contact surfaces were examined to understand possible sliding mechanism. The main conclusions are as follows: (1) In ambient air, the friction coefficient of a-C:H:F films decreased first and then markedly increased with increasing F contents in films. An optimum F content for a-C:H:F films, about 3 at.%, can achieve the lowest friction coefficient in a-C:H:F/Al contact. (2) A fluorinated composite-like tribo-layer, including Al compounds
and carbon components, formed on Al surface dominates the sliding behavior of a-C:H:F/Al contact, and its nature changes with tribochemical process involving F. (3) For lowly fluorinated a-C:H:F films, the reduced interfacial adhesion resulting from fluorinated tribo-layer lowers friction coefficient. For highly fluorinated a-C:H:F films, due to the enhanced tribochemical reaction between Al and F, the increased shearing and abrasive components significantly raise the friction coefficient. (4) A fluorinated carbon-based transfer layer can be expected to provide excellent tribological properties when those strong carbide former slide against a-C:H:F films with appropriate F content and optimized structure.
Acknowledgments This work was supported by the National Natural Science Foundation of China (Grant Nos. 51322508 and 21373249). References [1] H. Coldwell, R. Dewes, D. Aspinwall, N. Renevier, D. Teer, The use of soft/lubricating coatings when dry drilling BS L168 aluminium alloy, Surf. Coat. Technol. 177 (2004) 716–726. [2] M. Lahres, P. Müller-Hummel, O. Doerfel, Applicability of different hard coatings in dry milling aluminium alloys, Surf. Coat. Technol. 91 (1) (1997) 116–121. [3] H. Fukui, J. Okida, N. Omori, H. Moriguchi, K. Tsuda, Cutting performance of DLC coated tools in dry machining aluminum alloys, Surf. Coat. Technol. 187 (1) (2004) 70–76. [4] A. Riahi, A. Alpas, Adhesion of AA5182 aluminum sheet to DLC and TiN coatings at 25 °C and 420 °C, Surf. Coat. Technol. 202 (4) (2007) 1055–1061. [5] N. Wain, N. Thomas, S. Hickman, J. Wallbank, D. Teer, Performance of low-friction coatings in the dry drilling of automotive Al–Si alloys, Surf. Coat. Technol. 200 (5) (2005) 1885–1892. [6] Y. Qi, E. Konca, A.T. Alpas, Atmospheric effects on the adhesion and friction between non-hydrogenated diamond-like carbon (DLC) coating and aluminum–a first principles investigation, Surf. Sci. 600 (15) (2006) 2955–2965. [7] E. Konca, Y.-T. Cheng, A. Weiner, J. Dasch, A.T. Alpas, The role of hydrogen atmosphere on the tribological behavior of non-hydrogenated DLC coatings against aluminum, Tribol. Trans. 50 (2) (2007) 178–186. [8] Y. Qi, L.G. Hector Jr., Hydrogen effect on adhesion and adhesive transfer at aluminum/diamond interfaces, Phys. Rev. B 68 (20) (2003) 201403. [9] A.A. Gharam, M. Lukitsch, Y. Qi, A.T. Alpas, Role of oxygen and humidity on the tribochemical behaviour of non-hydrogenated diamond-like carbon coatings, Wear 271 (9) (2011) 2157–2163. [10] E. Konca, Y.-T. Cheng, A. Weiner, J. Dasch, A.T. Alpas, Effect of test atmosphere on the tribological behaviour of the non-hydrogenated diamond-like carbon coatings against 319 aluminum alloy and tungsten carbide, Surf. Coat. Technol. 200 (5) (2005) 1783–1791. [11] W. Ni, Y.-T. Cheng, A.M. Weiner, T.A. Perry, Tribological behavior of diamond-likecarbon (DLC) coatings against aluminum alloys at elevated temperatures, Surf. Coat. Technol. 201 (6) (2006) 3229–3234. [12] S. Bhowmick, A.T. Alpas, The performance of hydrogenated and non-hydrogenated diamond-like carbon tool coatings during the dry drilling of 319 Al, Int. J. Mach. Tools Manuf. 48 (7) (2008) 802–814.
F. Wang et al. / Surface & Coatings Technology 304 (2016) 150–159 [13] S. Bhowmick, A. Banerji, M. Lukitsch, A.T. Alpas, The high temperature tribological behavior of Si, O containing hydrogenated diamond-like carbon (a-C:H/a-Si:O) coating against an aluminum alloy, Wear 330 (2015) 261–271. [14] F. Sen, X. Meng-Burany, M. Lukitsch, Y. Qi, A.T. Alpas, Low friction and environmentally stable diamond-like carbon (DLC) coatings incorporating silicon, oxygen and fluorine sliding against aluminum, Surf. Coat. Technol. 215 (2013) 340–349. [15] A. Abou Gharam, M. Lukitsch, M. Balogh, N. Irish, A.T. Alpas, High temperature tribological behavior of W-DLC against aluminum, Surf. Coat. Technol. 206 (7) (2011) 1905–1912. [16] M. Ishihara, T. Kosaka, T. Nakamura, K. Tsugawa, M. Hasegawa, F. Kokai, Y. Koga, Antibacterial activity of fluorine incorporated DLC films, Diam. Relat. Mater. 15 (4) (2006) 1011–1014. [17] K.-I. Nakamatsu, N. Yamada, K. Kanda, Y. Haruyama, S. Matsui, Fluorinated diamondlike carbon coating as antisticking layer on nanoimprint mold, Jpn. J. Appl. Phys. 45 (9 L) (2006) L954. [18] T. Hasebe, A. Shimada, T. Suzuki, Y. Matsuoka, T. Saito, S. Yohena, A. Kamijo, N. Shiraga, M. Higuchi, K. Kimura, Fluorinated diamond-like carbon as antithrombogenic coating for blood-contacting devices, J. Biomed. Mater. Res. Part A 76 (1) (2006) 86–94. [19] S. Bai, T. Onodera, R. Nagumo, R. Miura, A. Suzuki, H. Tsuboi, N. Hatakeyama, H. Takaba, M. Kubo, A. Miyamoto, Friction reduction mechanism of hydrogen-and fluorine-terminated diamond-like carbon films investigated by molecular dynamics and quantum chemical calculation, J. Phys. Chem. C 116 (23) (2012) 12559–12565. [20] J. Wang, F. Wang, J. Li, Q. Sun, P. Yuan, Y. Jia, Comparative study of friction properties for hydrogen-and fluorine-modified diamond surfaces: a first-principles investigation, Surf. Sci. 608 (2013) 74–79. [21] S. Bhowmick, F. Sen, A. Banerji, A.T. Alpas, Friction and adhesion of fluorine containing hydrophobic hydrogenated diamond-like carbon (FH-DLC) coating against magnesium alloy AZ91, Surf. Coat. Technol. 267 (2015) 21–31. [22] C. Donnet, J. Fontaine, A. Grill, V. Patel, C. Jahnes, M. Belin, Wear-resistant fluorinated diamond-like carbon films, Surf. Coat. Technol. 94 (1997) 531–536. [23] F.G. Sen, Y. Qi, A.T. Alpas, Material transfer mechanisms between aluminum and fluorinated carbon interfaces, Acta Mater. 59 (7) (2011) 2601–2614. [24] S. Neuville, Quantum electronic mechanisms of atomic rearrangements during growth of hard carbon films, Surf. Coat. Technol. 206 (4) (2011) 703–726. [25] S. Neuville, Carbon Structure Analysis with Differentiated Raman Spectroscopy, LAP Academic Publishing, Saarbrucken Germany, 2014 1–176. [26] M. Rubio-Roy, E. Bertran, E. Pascual, M. Polo, J. Andújar, Fluorinated DLC deposited by pulsed-DC plasma for antisticking surface applications, Diam. Relat. Mater. 17 (7) (2008) 1728–1732. [27] Y. Pei, D. Galvan, J.T.M. De Hosson, Nanostructure and properties of TiC/a-C:H composite coatings, Acta Mater. 53 (17) (2005) 4505–4521. [28] A. Erdemir, C. Donnet, Tribology of diamond-like carbon films: recent progress and future prospects, J. Phys. D. Appl. Phys. 39 (18) (2006) R311. [29] E. Konca, Y.T. Cheng, A.M. Weiner, J.M. Dasch, A.T. Alpas, Elevated temperature tribological behavior of non-hydrogenated diamond-like carbon coatings against 319 aluminum alloy, Surf. Coat. Technol. 200 (12) (2006) 3996–4005.
159
[30] J. Fontaine, T. Le Mogne, J. Loubet, M. Belin, Achieving superlow friction with hydrogenated amorphous carbon: some key requirements, Thin Solid Films 482 (1) (2005) 99–108. [31] Y. Liu, A. Erdemir, E. Meletis, An investigation of the relationship between graphitization and frictional behavior of DLC coatings, Surf. Coat. Technol. 86 (1996) 564–568. [32] J. Sanchez-Lopez, A. Erdemir, C. Donnet, T. Rojas, Friction-induced structural transformations of diamondlike carbon coatings under various atmospheres, Surf. Coat. Technol. 163 (2003) 444–450. [33] S.M. Coman, M. Verziu, A. Tirsoaga, B. Jurca, C. Teodorescu, V. Kuncser, V.I. Parvulescu, G. Scholz, E. Kemnitz, NbF5–AlF3 catalysts: design, synthesis, and application in lactic acid synthesis from cellulose, ACS Catal. 5 (5) (2015) 3013–3026. [34] D. Kim, Y. Yu, K. Char, Characterization of aluminum oxyfluoride barrier in magnetic tunnel junctions, J. Appl. Phys. 96 (4) (2004) 2278–2285. [35] H. Fukui, M. Irie, Y. Utsumi, K. Oda, H. Ohara, An investigation of the wear track on DLC (a-C:H) film by time-of-flight secondary ion mass spectroscopy, Surf. Coat. Technol. 146 (2001) 378–383. [36] S.J. Park, K.R. Lee, D.H. Ko, Tribochemical reaction of hydrogenated diamond-like carbon films: a clue to understand the environmental dependence, Tribol. Int. 37 (11) (2004) 913–921. [37] A. Erdemir, C. Bindal, J. Pagan, P. Wilbur, Characterization of transfer layers on steel surfaces sliding against diamond-like hydrocarbon films in dry nitrogen, Surf. Coat. Technol. 76 (1995) 559–563. [38] J. Wei, W. Fong, D.B. Bogy, C.S. Bhatia, The decomposition mechanisms of a perfluoropolyether at the head/disk interface of hard disk drives, Tribol. Lett. 5 (2–3) (1998) 203–209. [39] P.H. Kasai, W.T. Tang, P. Wheeler, Degradation of perfluoropolyethers catalyzed by aluminum oxide, Appl. Surf. Sci. 51 (3) (1991) 201–211. [40] H. Liu, A. Tanaka, T. Kumagai, Influence of sliding mating materials on the tribological behavior of diamond-like carbon films, Thin Solid Films 352 (1) (1999) 145–150. [41] X. Meng-Burany, A.T. Alpas, FIB and TEM studies of damage mechanisms in DLC coatings sliding against aluminum, Thin Solid Films 516 (2) (2007) 325–335. [42] A. Konicek, D. Grierson, A. Sumant, T. Friedmann, J. Sullivan, P. Gilbert, W. Sawyer, R.W. Carpick, Influence of surface passivation on the friction and wear behavior of ultrananocrystalline diamond and tetrahedral amorphous carbon thin films, Phys. Rev. B 85 (15) (2012) 155448. [43] J. Fontaine, C. Donnet, A. Erdemir, Fundamentals of the tribology of DLC coatings, in: C. Donnet, A. Erdemir (Eds.), Tribology of Diamond-like Carbon Films, Springer 2008, pp. 139–154. [44] C. Bailey, S. Mukhopadhyay, A. Wander, B. Searle, N. Harrison, Structure and stability of α-AlF3 surfaces, J. Phys. Chem. C 113 (12) (2009) 4976–4983. [45] S. Neuville, A. Matthews, A perspective on the optimisation of hard carbon and related coatings for engineering applications, Thin Solid Films 515 (17) (2007) 6619–6653. [46] S. Neuville, New application perspective for tetrahedral amorphous carbon coatings, QScience Connect (2014).
Contents lists available at ScienceDirect
Surface & Coatings Technology journal homepage: www.elsevier.com/locate/surfcoat
Effect of tribochemistry on friction behavior of fluorinated amorphous carbon films against aluminum Fu Wang a,c, Zhibin Lu a, Liping Wang a,b,⁎, Guangan Zhang a,⁎⁎, Qunji Xue a a b c
State Key Laboratory of Solid Lubrication, Lanzhou Institute of Chemical Physics, Chinese Academy of Sciences, Lanzhou 730000, PR China Key Laboratory of Marine Materials and Related Technologies, Ningbo Institute of Materials Technology and Engineering, Chinese Academy of Sciences, Ningbo 315201, PR China University of Chinese Academy of Sciences, Beijing 100039, PR China
a r t i c l e
i n f o
Article history: Received 21 February 2016 Revised 29 June 2016 Accepted in revised form 30 June 2016 Available online 1 July 2016 Keywords: a-C:H:F films Aluminum F content Friction Tribo-layer Tribochemistry
a b s t r a c t Fluorinated amorphous carbon (a-C:H:F) films synthesized from C2H2 and CF4 were examined by Raman spectra and X-ray photoelectron spectroscopy (XPS). Their tribological properties were tested against aluminum balls in dry sliding. The a-C:H:F films with low F content showed lower friction coefficient than hydrogenated amorphous carbon (a-C:H) film, whereas high F content in films resulted in a significant friction increase. Contact surfaces were analyzed in detail to elucidate the possible sliding mechanism. Results indicated that the friction behavior was closely related to the nature of a composite-like tribo-layer consisting of Al compounds and carbon components formed on Al ball, relying on the tribochemical processes of contact interface. The accumulated F atoms on tribo-layer reduced the adhesion across sliding interface because of strong repulsion between F atoms, and thereby lowered the friction of a-C:H:F films. However, with increasing F contents in films, the enhanced tribochemical reaction between Al and F caused crack, delamination and fragmentation of the tribolayer, and then a marked abrasive process at the sliding interface. Consequently, the increased shearing and abrasive actions strongly opposed the contribution of the reduced adhesion to friction, and result in a rather high friction of highly fluorinated a-C:H:F films. © 2016 Elsevier B.V. All rights reserved.
1. Introduction The use of lubricating coatings greatly alleviates aluminum adhesion and transfer to tools, a very pressing problem in Al alloys dry machining [1]. Diamond-like carbon (DLC) films, with excellent tribological properties, have shown better anti-adhesive effect and lower friction coefficient than other conventional TiN, ZrN, TiAlN and TiCN coatings when tested against Al [2–5]. This has greatly stimulated interest in using them as promising tool coatings for Al dry machining. However, to improve the performance of DLC films, lower friction in DLC/Al contact is a continual drive for optimizing and designing DLC films. Thus, it is quite necessary to probe the detailed friction mechanism of various DLC films against Al. Many studies have investigated the friction behavior and mechanism of DLC films and their modifications against Al under different conditions. Non-hydrogenated DLC films exhibit high friction coefficient (μ = 0.5–0.8) under high vacuum and inert (He, Ar and N2) environments, where the adhesive transfer of Al to DLC surface results in an undesirable Al\\Al contact [6,7]. On the contrary, low friction (μ b 0.2) is obtained as the introduction of H2 and H2O that passivate the ⁎ Correspondence to: L. Wang, State Key Laboratory of Solid Lubrication, Lanzhou Institute of Chemical Physics, Chinese Academy of Sciences, Lanzhou 730000, PR China. ⁎⁎ Corresponding author. E-mail addresses: [email protected] (L. Wang), [email protected] (G. Zhang).
http://dx.doi.org/10.1016/j.surfcoat.2016.06.087 0257-8972/© 2016 Elsevier B.V. All rights reserved.
unoccupied C\\σ bonds on carbon surface [6–8]. Oxygen results in a high friction (μ = 0.28–0.44) in DLC/Al contact due to the oxidation of Al ball and carbon surface, but no Al adhesion transfer was observed. [6,9,10]. Hydrogenated DLC films exhibited lower friction coefficient than H-free DLC at evaluated temperature and showed lower cutting forces and less adhesion, attributed to higher chemical inertness of Hcontaining DLC network. [11,12]. The a-C:H/a-Si:O coating showed low steady-state friction coefficient up to 400 °C, where a Si-containing carbon-based transfer layer terminated by H, O and OH was identified on Al surface [13]. Similarly, a nano-structural carbon tribo-layer terminated by H, HO and F on Al surface accounted for the low friction behavior of a-C:H:Si:O:F coating in both vacuum and ambient atmosphere [14]. In addition, the formation of WO3 on Al surface dominated the low friction behavior of W-DLC/Al at 400 °C [15]. Briefly, the passivation of free C\\σ bonds and the formation of tribo-layer on Al surface are fundamental to the achievement of low friction in DLC/Al contact. Fluorinated amorphous carbon (a-C:H:F) films have been investigated as anti-sticking coatings in nanoimprint lithography and biomedical applications because of their low surface energy property [16–18]. Theoretical calculations reveal that fluorinated carbon surfaces display lower adhesion tendency when facing each other [19–21]. Meanwhile, Donnet et al. [22] found that a-C:H:F films showed comparable tribological properties to non-fluorinated ones when F content is b20 at.%. Interestingly, when tested against Al counterparts, Sen et al. [23] recently found that FDLC film showed lower friction coefficient than HDLC film.
F. Wang et al. / Surface & Coatings Technology 304 (2016) 150–159
They pointed out that a fluorinated transfer layer on Al surface may reduce the friction level of FDLC/Al contact. This suggests a potential use of a-C:H:F films in Al anti-adhesion. However, this meaningful result was only observed and discussed for a certain FDLC film (3 at.% F, 25 at.% H) in ambient air, many issues are still open. Most importantly, considering the large electro-negativity of F and fluorinated tribo-layer, aC:H:F films perhaps signify more complicated tribochemical processes, especially with increasing F contents, which possibly markedly influence the friction behavior of a-C:H:F against Al. So, it is of great significance to know the relevant details, which can offer more complete information to further consider possible use of a-C:H:F films, e.g. in Al dry machining. The present work, accordingly, focuses on the influence of F content and tribochemistry on the friction behavior of a-C:H:F against Al. The aC:H:F films were produced from the glow plasma decomposition of C2H2 and CF4 mixture. The fundamental nature of films was characterized and understood. The friction behavior of films was assessed under dry contact conditions. The contact surfaces were examined to identify the factors that influenced the friction behavior of a-C:H:F films against Al.
151
friction tests were also carried out under dry N2 (RH: 1.0%), dry air (RH: 5.0%) and humid air (RH: 60.0%) to consider environmental effect. 2.3. Characterization methods Micro-Raman spectra were recorded by a LabRam HR800 JobinYvon spectrometer with an excitation wavelength of 532 nm. The hardness (H) and elastic modulus (E) of films were determined by a nanoindenter Hysitron TI 950 USA. The residual stress of films was determined through the deformation of Si wafer caused by the deposition of film (SuPro Instruments, FST-1000). Surface compositions and chemical states of samples were analyzed by an ESCALAB 250Xi (Thermo Fisher, USA) X-ray photoelectron spectroscopy (XPS) system with an Al Kα (1486.7 eV) X-ray source. A DSA 100 goniometer was employed to measured water contact angles. Surface morphology of wear tracks was acquired by 3D Laser Scanning Microscope (Keyence VK-X100/ X200, Japan). The transfer layers were observed using a JSM-5600LV scanning electron microscope (SEM) equipped with an energy-dispersive X-ray spectroscopy (EDS). 3. Results and discussion
2. Experimental procedure 3.1. Characterization of films 2.1. Materials The a-C:H:F films were deposited on silicon (110) and mirrorpolished stainless steel plats (roughness: Ra b 20 nm) by plasma enhanced chemical vapor deposition (PECVD) system, using acetylene (C2H2) and carbon tetrafluoride (CF4) as precursor gases. The base pressure of deposition chamber was better than 2 × 10− 3 Pa. Solvent cleaning and Ar+ etching procedures were performed to remove contaminants on substrate surfaces before deposition process. A thin aSi:H interlayer (~80 nm) and a-C:Si:H gradient layer (~500 nm) were produced to improve the adhesion of film to substrates. The a-C:H:F layer was synthesized from C2H2 and CF4 at a pressure of 6.5 Pa. A pulse power source was used (frequency, 1.5 kHz; duty ratio, 30%). The gas flow ratio (C2H2/CF4) and the corresponding sample are listed in Table. 1. The a-C:H (about 10 at.% H) and a-C:H:F-6 films were deposited at substrate voltage − 0.8 kV and − 1.0 kV, respectively, and the other films were produced at − 0.9 kV. The total thickness of films was about 2.0 μm (measured by SEM).
The micro-structure of films was investigated by Raman spectroscopy. As shown in Fig. 1, the broad, asymmetrical Raman peaks between 1000 and 1700 cm−1 suggested disordered carbon structures. According to recent publications by Neuville [24,25], the spectra can be resolved into two sub-bands: the so-called D band at 1340/1398 cm−1 and C5/C7 odd ring band at 1534/1552 cm−1. The so-called D band in fact corresponds to either disordered diamond (1330 cm−1) or to the G Aedge vibration modes of sp2 clusters (1350 cm−1). For a-C:H film, the D band at 1340 cm−1 mainly corresponded to disordered Csp3-Csp3 clusters. The C5/C7 band at 1534 cm−1 indicated that there were no significant hexagonal cyclic sp2 rings within film, which further signified that the D band at 1340 cm−1 could only encapsulate weak G Aedge
2.2. Friction and wear tests The sliding experiments were performed using a CSM tribometer and a ball-on-disk contact geometry with a reciprocating motion under a load of 3.0 N, a speed of 0.05 m/s (frequency: 5.0 Hz, stroke length: 2.5 mm). The counterpart was commercial purity (N 99%) aluminum ball with a diameter of 4.0 mm, Vickers hardness of about 270 MPa and the surface roughness, Ra, of ~60 nm. Aluminum balls were tested against films deposited on steel substrates in ambient air with a relative humidity (RH) of 30 ± 2.5% and halted after 20,000 laps, corresponding to a sliding distance of 200 m. The cross-sectional area of the track was determined by a profile-meter (KLA-Tencor, D-100). Furthermore, Table 1 Source gas ratio, F content, hardness (H), elastic modulus (E), internal stress (σ) and water contact angle (θ) of as-deposited films. Sample
C2H2/CF4/sccm
F/at.%
H/GPa
E/GPa
σ/GPa
θ/°
a-C:H a-C:H:F-1 a-C:H:F-2 a-C:H:F-3 a-C:H:F-4 a-C:H:F-5 a-C:H:F-6
120/0 60/20 60/40 60/60 50/70 40/80 30/90
0 1.7 3.4 6.5 9.6 16.7 18.6
14.8 15.4 14.6 13.4 12.9 11.8 10.3
127.8 131.5 134.4 122.7 116.3 99.1 89.4
−0.85 −0.70 −0.61 −0.43 −0.58 −0.41 −0.41
71.8 ± 0.7 78.3 ± 0.7 77.2 ± 1.5 78.0 ± 0.7 74.7 ± 0.8 81.7 ± 1.3 81.6 ± 2.4
Fig. 1. Raman spectra of a-C:H and a-C:H:F films.
152
F. Wang et al. / Surface & Coatings Technology 304 (2016) 150–159
band at ~1350 cm−1. This suggests that a-C:H film is the disordered sp3/ sp2 carbon matrix which can contain amorphous Csp3-Csp3 clusters. However, for a-C:H:F films, the D bands were significantly upshifted (above 1360 cm−1) and broadened. First, the higher deposition voltage of a-C:H:F films might upshift the observed D band because of possibly increased graphenic sp2 rings or sp2 clusters, which increased the G Aedge mode of sp2 clusters (~1350 cm−1). Second, the residual compressive stress in film (Table 1) upshifted the Raman frequencies of films to some extent. Third, owing to large electron-negativity of F atom, the formed C\\F bonds (stronger than C\\H bonds) resulted in the non-uniform distribution of bond angles, bond lengths and bond energies of C\\C bonds, and made them randomly distributed over a larger range around some nominal values. Thus, the corresponding Raman frequencies were also distributed over a broader Raman band. This explains why the D band frequencies of a-C:H:F films increased with increasing F contents, and also gives account for the D band broadening of a-C:H:F films as the F content. The broadening of D band in return reflected the increased atomic disorder of a-C:H:F films. Accordingly, the a-C:H:F films showed slightly increased sp2 clusters and increased atomic disorder compared to a-C:H film. By and large, agreeing with the micro-structural changes, both hardness and residual compressive stress of a-C:H:F films showed a decreasing tendency (Table 1). The surface F content of a-C:H:F films was determined by semi-quantitative XPS analysis, in the range of 1.7–18.6 at.% (Table 1). The measured water contact angles (Table 1) increased as the incorporation of F, indicating low surface energy property of a-C:H:F films. Fig. 2 shows high resolution XPS spectra of a-C:H:F films. All spectra were calibrated by fixing the C\\C/C\\H peaks at 284.8 eV. The C 1s spectra (Fig. 2a) appeared two shoulder peaks at ~ 287.5 and ~ 289.5 eV (C\\F group), which gradually enhanced with increasing F contents, while the ratio of C-C/C-H peaks dwindled. Together with the F 1s peaks (Fig. 2b), this well demonstrated the different fluorination degrees in individual films. With the increase of F content, both the C 1s shoulder peak and F 1s peak were slightly shifted towards higher binding energy. This could be attributed to the carbon matrix with a more fluorinated bonding environment, which increased their bonding energy because of the high electron-negativity of F atom [26]. In short, despite different fluorination degrees, F atoms were mainly bonded to carbon matrix by mono-fluorinated carbon atom (C\\F group). 3.2. Tribological behavior of a-C:H:F films against Al balls Fig. 3 shows the evolution of friction coefficient of films with sliding laps in ambient air (RH = 30%). For comparison, a-C:H film was also tested under the same conditions. From the plot, friction coefficient of films showed a slightly decreasing trend with the incorporation of a small amount of F (a-C:H:F-1, -2, -3), and the a-C:H:F-2 film with
3.4 at.% F exhibited the lowest friction value, in good agreement with the report by Sen et al. [23]. However, friction coefficient of films significantly increased with further increase in F content (a-C:H:F-4, -5, -6). The average steady-state friction coefficients of films calculated from 5000 to 20,000 laps were 0.20, 0.18, 0.21, 0.28, 0.32, and 0.47 for aC:H:F-1 to a-C:H:F-6, respectively, and 0.23 for a-C:H film. Therefore, it seemed that the reduction of friction was only achieved for lowly fluorinated films (≤6 at.% F). Note that all the friction curves (Fig. 3) were characterized by a rapid friction drop at the beginning of sliding, followed by a gradual increase and high steady-state value. This differed from the frequently observed cases, where a gradual decrease in friction level represents the running-in process [27], attributed to surface finish of contact bodies and the formation of transfer layer on the counterpart. Actually, similar behavior can also be seen in report by Sen et al. [23]. It was evident that the difference in hardness, hydrophobicity (Table 1) and bonding configuration of films (Figs. 1 and 2) cannot explain the observed variations in friction coefficient because of their inconsistent trends. For hard, amorphous carbon films deposited on smooth substrates, the adhesive, tribochemical and third-body interactions are the main factors influencing their friction behavior [28]. On account of high reactivity of Al towards O2 in air and F in film, it is considered that tribochemistry at the sliding interface could dominate the observed friction behavior. The wear rates of films were calculated by the wear volume of films divided by the product of normal load and sliding distance. Previous work by Konca et al. showed that 319 Al can cause more wear of the DLC films than WC and sapphire because of the formation of Al oxide abrasives at the sliding interface [29]. In current investigation, the wear rates of a-C:H:F films were of the same order of magnitude as that of a-C:H film (b 7 × 10−7 mm3/Nm). However, highly fluorinated a-C:H:F-6 film showed the highest friction coefficient but the lowest wear rate (1.8 × 10−7 mm3/Nm). This might be related to its different deposition process or other unclear reasons. Furthermore, the addition of F to films increased the wear of Al balls, as shown in Fig. 4. 3.3. Tribo-layer and wear track The tribological behavior of DLC films under dry sliding contact is normally governed by a carbon-rich transfer layer formed on counterpart surfaces, which prevents the direct contact between DLC films and counterpart materials and establishes a new sliding interface [30– 32]. To understand the friction behavior of a-C:H:F films, the contact surfaces of Al balls were firstly observed after sliding against a-C:H (Ffree), a-C:H:F-2 (low F) and a-C:H:F-5 (high F) films. Fig. 4 shows their SEM images after sliding 20,000 laps in ambient air. Obviously, a tribo-layer or transfer layer was established on Al ball surfaces for three films. When tested against a-C:H film, an integral and uniform tribo-layer was formed (Fig. 4a and b). Some fine cracks distributed
Fig. 2. High resolution XPS spectra of a-C:H:F films (a) C 1s and (b) F 1s.
F. Wang et al. / Surface & Coatings Technology 304 (2016) 150–159
Fig. 3. Friction curves of a-C:H and a-C:H:F films against Al balls in ambient air.
on the edge of the tribo-layer as sliding against a-C:H:F-2 film (Fig. 4c and d). However, when sliding contact with a-C:H:F-5 film, the tribolayer showed an apparent cracked appearance (Fig. 4e and f). This suggests that the incorporation of F can dramatically change the nature of tribo-layer. The easy shear and chemical states of the tribo-layer are essential to achieve low friction of DLC films. In the light of the fragmentation behavior in Fig. 4e and f, tribo-layer against a-C:H:F-5 film seemed to
153
have higher brittleness than those against a-C:H and a-C:H:F-2 films, which might be mainly associated with their chemical compositions. Although without a quantitative description of the mechanical properties of these layers, the high brittleness possibly suggested the increased shear resistance and hence increased friction. The cohesion of the tribo-layers and its adhesion to Al surface could affect their changes in SEM images. The formation of AlF3 at Al/tribo-layer interface had been confirmed by Sen et al. [14,23]. In terms of the formation of the tribolayer, the formed AlF3 on Al surface may prevent the chemical binding of carbon to Al surface as a result of surface termination, while the increased proportion of Al fluorides in tribo-layer would decrease the cross-link degree of the layer. With the increase of F content in film, more AlF3 was grown on Al surface due to the increased reaction rate between F and Al, which might markedly affect the adhesion strength of tribo-layer to Al surface, and also decrease the internal cross-linking density of the layer. Therefore, the tribo-layer formed from highly fluorinated a-C:H:F films would easily crack, delaminate and fragment into debris under continuous mechanical actions, some of which acted as abrasive particles and led to high friction. Fig. 5 presents the optical observation of the wear tracks formed on different films. Compared to a-C:H film, an obvious difference was that the sliding tracks on a-C:H:F films became wider. This can be associated with the tribochemical reactions between Al and F, which accelerated the wear of Al balls and hence increased the contact area. As shown in Fig. 4, despite decreasing hardness of films, the size of the wear scar increased with increasing F contents. Meanwhile, the low friction
Fig. 4. SEM images of the wear scars on Al ball surfaces after sliding against (a) a-C:H, (b) a-C:H:F-2 and (c) a-C:H:F-5 films in ambient air, and their local images (d), (e) and (f).
154
F. Wang et al. / Surface & Coatings Technology 304 (2016) 150–159
coefficient for a-C:H and a-C:H:F-2 films corresponded to (see in Fig. 3) a relatively smooth surface of wear tracks (Fig. 5a and b), whereas the high friction coefficient for a-C:H:F-5 and a-C:H:F-6 films corresponded to the rough wear track surfaces (Fig. 5c and d). In particular, the severer plowings grooves were clearly observed for a-C:H:F-6 film with the highest F content (Fig. 5d). This suggested that there were more hard wear particles introduced to the sliding interface when sliding against highly fluorinated films. Owing to the susceptibility of Al to O2, it is believed that the oxidation of Al surface and/or Al2O3 abrasives significantly affect the friction and wear behavior of non-hydrogenated DLC film [29]. While in a-C:H:F films the large electron-negativity difference between F (3.98) and Al (1.61) also facilitates the formation of Al fluorides. It was thereby thought that stable aluminum oxides (Al2O3) and fluorides at sliding interface could act as hard abrasives that contribute to high friction, and more abrasive particles were introduced into the sliding interface as the F content increased. 3.4. Chemical analyses of sliding interfaces 3.4.1. XPS analysis of tribo-layers and wear tracks The tribo-layers against a-C:H, a-C:H:F-2, and a-C:H:F-5 films were examined by XPS to acquire their surface chemistry. Fig. 6a shows their full range XPS spectra and chemical compositions. The tribolayer from a-C:H film consisted of C, O and Al, while F also became one of the major components of the tribo-layer against a-C:H:F films. Fluorine was clearly accumulated on the tribo-layer surfaces for two a-C:H:F films (Fig. 6a). In particular, when sliding against a-C:H:F-2 film, the tribo-layer had four-fold higher F concentration (16.0 at.%) compared to 3.4 at.% F in film. Moreover, it was clear that Al was an important component of all the tribo-layers, and its content increased as the addition of F to film. This might be associated with the increased wear of Al balls when tested against a-C:H:F films. To further address the chemical difference of the tribo-layers, high resolution XPS spectra of C 1s, F 1s and Al 2p are presented in Fig. 6b– d. The C 1s spectra in Fig. 6b mainly consisted of three peaks at 284.8, 286.4 and 288.7 eV, which are assigned to the C\\C/C\\H, C\\OH/ C\\O\\C and O\\C_O groups, respectively. But C atoms became less oxidized with the increase of F contents within films. The
hydrophobicity of a-C:H:F films could reduce tribo-oxidation of carbon atoms caused by absorbed water molecules. In the case of a-C:H:F-2 and a-C:H:F-5 films, the peaks of 287.5 and 289.3 eV were also identified and assigned to the C\\F groups, but their low ratios signified that there was only a small portion (~5%) of C atoms fluorinated on tribo-layers. This well certified the break of C\\F bonds during rubbing, and was further demonstrated by the lower F 1s binding energy of tribo-layer than that of film (Fig. 6d). An unidentified peak at ~ 293.3 eV was also observed for all three tribo-layers. As a consequence, for a-C:H:F films, the majority of the accumulated F atoms on tribo-layer were not chemically bound to carbon skeleton to create a fluorinated carbon surface, which is very beneficial for achieving the low friction of a-C:H:F films because of repulsive interaction between two fluorinated carbon surfaces [19,20]. High resolution Al 2p spectra (Fig. 6c) appeared at 74.5, 74.9 and 75.4 eV, corresponding to a-C:H, a-C:H:F-2 and a-C:H:F-5 respectively. This indicated that the Al 2p binding energy of tribo-layer increased as increasing F contents in film. For a-C:H film, the Al 2p peak of 74.5 eV suggested the presence of Al2O3 phase on the surface of the tribolayer. However, for a-C:H:F films, the Al 2p binding energy of tribolayers was higher than that of Al2O3 and slightly increased with the F content, but lower than that of AlF3 (76.3 eV). This could represent a mixture of Al oxides, Al fluorides and Al oxyfluorides (AlOxFy). The binding energies of Al 2p and F 1s can be shifted to a lower level as the substitution of F atom in AlF3 by O atom [33,34]. Consequently, the tribolayer on Al ball surfaces seemed to be a composite layer including carbon and complex Al compound components, which dominated the sliding behavior of films against Al. Note that the composition and nature of the layer varied as film chemistry. As shown in Fig. 6a, the increase of F content in film induced more Al compounds in tribo-layer. This could result in the increased brittleness and the decreased cohesion of tribolayer, and hence increase the possibility of formation of the Al-containing wear particles. The wear track regions on a-C:H, a-C:H:F-2 and a-C:H:F-5 were also examined using XPS after sliding tests. As seen in Fig. 7a, the chemical composition inside and outside the wear track on a-C:H:F-2 film was quite similar, whereas the wear track on a-C:H:F-5 film contained lower F concentration (10.7 at.%) compared to the unworn region
Fig. 5. Optical images of the wear tracks formed on (a) a-C:H, (b) a-C:H:F-2, (c) a-C:H:F-5 and (d) a-C:H:F-6 films after 20,000 laps in ambient air.
F. Wang et al. / Surface & Coatings Technology 304 (2016) 150–159
155
Fig. 6. (a) Full range XPS spectra and composition of tribo-layers against a-C:H, a-C:H:F-2 and a-C:H:F-5 films, their high resolution XPS spectra of (b) C 1s, (c) Al 2p and (d) F 1s. Film in (d) represents as-deposited a-C:H:F-5 film.
(16.7 at.% F). Such variation in F concentration for a-C:H:F-5 film may largely be associated with the inherent nature of a-C:H:F films, and not a result of friction. When the growth process of film ends, more passivation atoms are required to form a stable and low energy surface. Thus, for high CF4 ratio in source gas, F atom controls this process, whereas for high C2H2 ratio, H atom dominates the same process. This suggests that higher F concentration can be observed on the surface of highly fluorinated a-C:H:F films. Sliding contact might not induce a significant change in chemical states of carbon-based network. With regard to Al, a weak Al 2p signal was only detected on the surface of the
wear track on a-C:H:F-5 film (Fig. 7b). This indicated that more Al-containing particles possibly detached from the tribo-layer and acted as abrasive particles, which actively supported the observation in Figs. 3–5. 3.4.2. EDS mapping of wear particles In DLC/steel contact, the iron-rich debris is closely related to their relatively high friction value (μ ≈ 0.15) in ambient air [35,36]. The wear debris aside the wear tracks on a-C:H and a-C:H:F-5 were also examined by SEM. Fig. 8 shows their SEM images and EDS elemental mappings. The more scattered and larger particles were observed for a-
Fig. 7. (a) Full range XPS spectra of wear tracks and films for a-C:H:F-2 and a-C:H:F-5 films, (b) high resolution Al 2p spectra of wear tracks on a-C:H, a-C:H:F-2 and a-C:H:F-5 films.
156
F. Wang et al. / Surface & Coatings Technology 304 (2016) 150–159
C:H:F-5 (Fig. 8d). The mappings of O and Al suggested Al-containing granular or powdery debris, possible Al oxides debris. Although the F and C mappings did not give available information on wear debris (data not shown), it was deduced that F was also contained in wear particles for a-C:H:F films because of their fluorinated tribo-layer with Al fluorides. Together with the grooves in Fig. 5c and d, it seemed that these Al-containing particles, as the third-body abrasive, caused the high friction and high F content in film aggravated this process. 3.4.3. Raman spectra of tribo-layers Normally, a graphitized carbon layer formed on the counterpart is in favor of the low friction and wear behavior of DLC films [31,37]. XPS results in Fig. 6a and b indicated that carbon was still an important component of the tribo-layers and formed various organic groups, including alcohol, ether, and carboxyl. Their microstructure was further identified at a larger scale by Raman spectra. As shown in Fig. 9, however, all tribo-layers did not show a frequently observed ‘graphite-like’ feature (separated D peak and G peak), only very weak amorphous carbon signal between 1000 and 1800 cm−1 for a-C:H and a-C:H:F-5 films. Thereby, carbon in tribo-layer might be organized in either non-continuous tiny structures or polymer-like chain structures terminated by\\H, \\OH, \\O, and\\F. The formation of Al compounds and the catalytic degradation of carbon network by formed Al compounds (Al2O3 or AlF3) may partly explain such structures [38,39]. Although there was no continuous carbon-based transfer layer formed, these carbon species in a composite-like tribo-layer might increase the crosslink of the layer, and also effectively prevented the adhesive transfer of Al to film surface. In contrast, complex Al compounds likely increased the shear resistance of tribo-layer and induced the generation of hard abrasive particles. Therefore, when sliding contact with Al counterpart, DLC films exhibit higher friction coefficient compared with other mating materials, e.g. Al2O3, ZrO2, SiC, and Si3N4 [40].
Fig. 9. Raman spectra of tribo-layers against a-C:H, a-C:H:F-2 and a-C:H:F-5 films.
testing under ambient conditions did not induce an effective carbon-based transfer layer on Al surface [40], in accord with the Raman observation in Fig. 9. This may mainly be because that: a) the H and OH passivation of carbon surface prevents the generation of free C\\σ bonds; b) the growth of oxides on Al surface reduces the exposure of active Al sites (ΔGo(Al2O3) = − 1690 kJ/mol); c) soft Al tends to be tore and transferred to the surface of hard DLC [7,10,41]; d) the formed C\\Al bonds are easily decomposed into C\\H and Al\\O by water molecules, as listed in formulas (2)–(4). C\H=C\C=C\O þ Al→C\Al=C\O⋯Al
ð1Þ
3.5. Tribochemistry at sliding interfaces
C\C=C\O þ H2 O→C\H=C\OH
ð2Þ
The foregoing observations indicated that complex chemical processes occurred during rubbing, and significantly changed the nature of tribo-layer with increasing F contents. In terms of surface carbon, the generation of free C\\σ bonds can promote the formation of C\\Al bonds, which is highly conducive to the adhesion of carbon layer to Al surface. The C\\O bonds can also increase the adhesion of carbon layer as the formation of C\\O\\Al bonds (in formula (1)). Nevertheless, even if a relatively strong bonding between Al and free C\\σ bonds (the standard Gibbs free energy of formation: ΔGo(Al4C3) = −218 kJ/mol),
Al þ O2 =C\O→Al oxides
ð3Þ
C\Al þ H2 O→C\H þ Al oxides
ð4Þ
Apart from these possible processes, for a-C:H:F films, the large electron-negativity difference between F (3.98) and Al (1.61) facilitates the production of Al fluorides (ΔGo(AlF3) = −1530 kJ/mol). According to TEM evidence by Sen et al. [23], an AlF3 layer, rather Al2O3 layer, is formed at Al/tribo-layer interface in FDLC/Al sliding contact. This
Fig. 8. SEM images of wear debris aside the wear tracks on (a) a-C:H and (d) a-C:H:F-5, and element mappings: (b), (e) for O and (c), (f) for Al.
F. Wang et al. / Surface & Coatings Technology 304 (2016) 150–159
suggests that although Al readily oxidizes under ambient conditions, the formed oxide film on its surface is removed during rubbing. It is also indicated that the reaction of Al with F is kinetically supported because of the direct contact between Al and F (formula (5)). As depicted in Section 3.3, the formed AlF3 affected the bonding strength of tribo-layer to Al surface. Low F content, to some extent, might also promote the formation of C\\Al bonds because of the cleavage of C\\F bonds and nascent Al surface (formula (5)), whereas high F content markedly weakened the adhesion of tribo-layer and lead to a cracked tribo-layer with more Al fluorides formed (Fig. 4e and f). In addition, AlF3, a strong Lewis acid, is readily coordinated by alkoxy groups (C\\O) and H2O molecules to form C\\O…AlF3 and AlF3·xH2O. The former might promote the attachment of carbon component, but the latter might cause the detachment of AlF3 layer and could be transformed into AlOxFy under mechanical actions and friction heat (formula (7)). Therefore, oxygen-rich AlOxFy could be formed for lowly fluorinated a-C:H:F films, while fluorine-rich AlOxFy was formed for highly fluorinated a-C:H:F films, as with the XPS spectra in Fig. 6c and d. C\F þ Al→C\Al þ AlF3
ð5Þ
AlF3 þ C\O→C\O…AlF3
ð6Þ
AlF3 þ xH2 O→AlF3 xH2 O→Al ðOHÞx Fy →AlOx Fy
ð7Þ
3.6. Proposed sliding mechanism for a-C:H:F/Al contact Given that the a-C:H:F films were deposited from a carbonhydrogen-fluorine system, it is though that diamond nano-crystals can be facilitated during deposition because of thermal activation involving fluorine [24]. Thus, it should be considered that formed diamond nanocrystals might cause the increase of friction with increasing F contents. However, on the basis of the Raman spectra and hardness data of films, either the increased sp2 clusters or the decreased hardness of aC:H:F films disapproved of the increase of possible diamond nanocrystals with the F content. This suggested that the sliding behavior of a-C:H:F films in contact with Al was mainly associated with the
157
passivation of surface C\\σ bonds and the interfacial tribochemistry, rather than diamond nano-structure. The saturation of free C\\σ bonds can eliminate strong covalent adhesion between contact carbon surfaces [42]. The water molecules in ambient air efficiently passivated the surface free C\\σ bonds and formed a low-adhesive sliding surface. And hence the adhesion transfer of Al was not observed for all tests. As shown in Fig. 3, the initially low friction of films might be attributed to the H and OH passivation of dangling C\\σ bonds, but the subsequent increase in friction suggested that some detrimental factors appeared with continuous sliding. This may correspond to the formation of a specific tribo-layer on Al surface. On the basis of the Raman observation (Fig. 9) and the formation of Al compounds (Fig. 6), a composite-like tribo-layer including carbon structures and Al compounds may largely cause such increase in friction. The addition of F clearly induced different tribochemical processes and altered the nature of tribo-layer on Al surface. In comparison with a-C:H film, the friction behavior of a-C:H:F films exhibited two opposite trends. One is the friction reduction for films with low F content (b6 at.%) and another is the significant friction increment for films with high F content (N 9 at.%). For the first case, it is proposed that the repulsive force arising from fluorinated transfer layer including C\\F groups and AlF3 is responsible for the observed friction reduction [23]. However, based on current observations, different friction mechanisms should be considered for a-C:H:F/Al contact. According to the discussion of Fontaine [43], there are three main components contributing to the friction coefficient of DLC films, μ = μabrasion + μshearing + μadhesion, where μabrasion, μshearing and μadhesion represent the abrasive, shearing and adhesive component of friction coefficient, respectively. For most DLC sliding systems, a carbon-based transfer layer well avoids the generation of third-body abrasives and offers low shearing resistance [43]. Nevertheless, for the a-C:H:F/Al contact, for one thing, the accumulated F atoms on a composite-like tribo-layer reduced interface adhesion because of repulsion between F atoms, and thereby reduced the adhesive component of friction, μadhesion [19,20,44]. For another, the presence of Al and F in tribo-layer decreased the crosslink density of the layer and increased the shear resistance of the layer (Fig. 4), and their contents in tribo-layer increased with the increase of F content in film (Fig. 4). In the meantime, the increased Al fluorides at Al/tribo-layer interface weakened the adhesion of tribo-layer to Al
Fig. 10. Scheme of sliding interfaces for a-C:H and a-C:H:F films against Al balls in ambient air, (a) lowly fluorinated a-C:H:F films, (b) a-C:H films and (c) highly fluorinated a-C:H:F films.
158
F. Wang et al. / Surface & Coatings Technology 304 (2016) 150–159
Fig. 11. Friction curves of (a) a-C:H and (b) a-C:H:F-2 films against Al balls under dry N2 (RH: 1.0%), dry air (RH: 5.0%) and humid air (RH: 60.0%).
surface. Therefore, the tribo-layer derived from highly fluorinated films easily cracks, delaminates, and fragments into third-body abrasives due to the enhanced tribochemical processes. As a result, lowly fluorinated a-C:H:F films did not markedly change the μabrasion and μshearing components when compared to the a-C:H film, conversely, the reduced μadhesion lowered friction coefficient. However, highly fluorinated aC:H:F films enhanced tribochemical reaction of Al with F, which generated more Al compounds, especially complex Al fluorides. This promoted the production of third-body abrasives, and so the increased μabrasion and μshearing greatly raised the friction coefficient. As shown in Fig. 10, a proposed scheme depicts the interfacial behavior of a-C:H and a-C:H:F films sliding against Al. Relying on the F content of films (i.e. tribochemical process), the reduced μshearing component and the increased μabrasion and μshearing components dominate competitively the friction response of a-C:H:F/Al contact. Considering the susceptibility of tribochemical process to environments, additional sliding experiments were carried out under different atmosphere. As seen in Fig. 11, when tests were performed under dry N2 (RH: 1.0%), dry air (RH: 5.0%) and humid air (RH: 60.0%), the aC:H:F-2 film did not exhibit a lower friction value than a-C:H film. Accordingly, although the lowly a-C:H:F films can achieve a lower friction coefficient under ambient conditions when tested against Al, the tribochemical process involving F may restrict such friction merit in varied environments, e.g. Al alloys dry machining. Nevertheless, notably, the build-up of the fluorinated tribo-layer, suggesting a low adhesive interface, may be of great interest for tribological applications of a-C:H:F films. At present, the use of Al mating material may not quite fulfil the prerequisites to form a fluorinated carbon-based transfer because of its high softness, high affinity towards F and relatively weak affinity towards C. But when mating materials are those strong carbide formers (e.g. WC, Ti, SiC, Si3N4), a fluorinated carbon transfer layer can be expected to provide excellent tribological properties due to strong repulsive forces between two fluorinated carbon surfaces. Then the optimization of composition and structure of fluorinated carbon-based network would become crucial for their tribological applications [45, 46]. 4. Conclusions The a-C:H:F films synthesized from C2H2 and CF4 plasma were slid against Al balls to investigated the effect of F content on their friction behavior. Contact surfaces were examined to understand possible sliding mechanism. The main conclusions are as follows: (1) In ambient air, the friction coefficient of a-C:H:F films decreased first and then markedly increased with increasing F contents in films. An optimum F content for a-C:H:F films, about 3 at.%, can achieve the lowest friction coefficient in a-C:H:F/Al contact. (2) A fluorinated composite-like tribo-layer, including Al compounds
and carbon components, formed on Al surface dominates the sliding behavior of a-C:H:F/Al contact, and its nature changes with tribochemical process involving F. (3) For lowly fluorinated a-C:H:F films, the reduced interfacial adhesion resulting from fluorinated tribo-layer lowers friction coefficient. For highly fluorinated a-C:H:F films, due to the enhanced tribochemical reaction between Al and F, the increased shearing and abrasive components significantly raise the friction coefficient. (4) A fluorinated carbon-based transfer layer can be expected to provide excellent tribological properties when those strong carbide former slide against a-C:H:F films with appropriate F content and optimized structure.
Acknowledgments This work was supported by the National Natural Science Foundation of China (Grant Nos. 51322508 and 21373249). References [1] H. Coldwell, R. Dewes, D. Aspinwall, N. Renevier, D. Teer, The use of soft/lubricating coatings when dry drilling BS L168 aluminium alloy, Surf. Coat. Technol. 177 (2004) 716–726. [2] M. Lahres, P. Müller-Hummel, O. Doerfel, Applicability of different hard coatings in dry milling aluminium alloys, Surf. Coat. Technol. 91 (1) (1997) 116–121. [3] H. Fukui, J. Okida, N. Omori, H. Moriguchi, K. Tsuda, Cutting performance of DLC coated tools in dry machining aluminum alloys, Surf. Coat. Technol. 187 (1) (2004) 70–76. [4] A. Riahi, A. Alpas, Adhesion of AA5182 aluminum sheet to DLC and TiN coatings at 25 °C and 420 °C, Surf. Coat. Technol. 202 (4) (2007) 1055–1061. [5] N. Wain, N. Thomas, S. Hickman, J. Wallbank, D. Teer, Performance of low-friction coatings in the dry drilling of automotive Al–Si alloys, Surf. Coat. Technol. 200 (5) (2005) 1885–1892. [6] Y. Qi, E. Konca, A.T. Alpas, Atmospheric effects on the adhesion and friction between non-hydrogenated diamond-like carbon (DLC) coating and aluminum–a first principles investigation, Surf. Sci. 600 (15) (2006) 2955–2965. [7] E. Konca, Y.-T. Cheng, A. Weiner, J. Dasch, A.T. Alpas, The role of hydrogen atmosphere on the tribological behavior of non-hydrogenated DLC coatings against aluminum, Tribol. Trans. 50 (2) (2007) 178–186. [8] Y. Qi, L.G. Hector Jr., Hydrogen effect on adhesion and adhesive transfer at aluminum/diamond interfaces, Phys. Rev. B 68 (20) (2003) 201403. [9] A.A. Gharam, M. Lukitsch, Y. Qi, A.T. Alpas, Role of oxygen and humidity on the tribochemical behaviour of non-hydrogenated diamond-like carbon coatings, Wear 271 (9) (2011) 2157–2163. [10] E. Konca, Y.-T. Cheng, A. Weiner, J. Dasch, A.T. Alpas, Effect of test atmosphere on the tribological behaviour of the non-hydrogenated diamond-like carbon coatings against 319 aluminum alloy and tungsten carbide, Surf. Coat. Technol. 200 (5) (2005) 1783–1791. [11] W. Ni, Y.-T. Cheng, A.M. Weiner, T.A. Perry, Tribological behavior of diamond-likecarbon (DLC) coatings against aluminum alloys at elevated temperatures, Surf. Coat. Technol. 201 (6) (2006) 3229–3234. [12] S. Bhowmick, A.T. Alpas, The performance of hydrogenated and non-hydrogenated diamond-like carbon tool coatings during the dry drilling of 319 Al, Int. J. Mach. Tools Manuf. 48 (7) (2008) 802–814.
F. Wang et al. / Surface & Coatings Technology 304 (2016) 150–159 [13] S. Bhowmick, A. Banerji, M. Lukitsch, A.T. Alpas, The high temperature tribological behavior of Si, O containing hydrogenated diamond-like carbon (a-C:H/a-Si:O) coating against an aluminum alloy, Wear 330 (2015) 261–271. [14] F. Sen, X. Meng-Burany, M. Lukitsch, Y. Qi, A.T. Alpas, Low friction and environmentally stable diamond-like carbon (DLC) coatings incorporating silicon, oxygen and fluorine sliding against aluminum, Surf. Coat. Technol. 215 (2013) 340–349. [15] A. Abou Gharam, M. Lukitsch, M. Balogh, N. Irish, A.T. Alpas, High temperature tribological behavior of W-DLC against aluminum, Surf. Coat. Technol. 206 (7) (2011) 1905–1912. [16] M. Ishihara, T. Kosaka, T. Nakamura, K. Tsugawa, M. Hasegawa, F. Kokai, Y. Koga, Antibacterial activity of fluorine incorporated DLC films, Diam. Relat. Mater. 15 (4) (2006) 1011–1014. [17] K.-I. Nakamatsu, N. Yamada, K. Kanda, Y. Haruyama, S. Matsui, Fluorinated diamondlike carbon coating as antisticking layer on nanoimprint mold, Jpn. J. Appl. Phys. 45 (9 L) (2006) L954. [18] T. Hasebe, A. Shimada, T. Suzuki, Y. Matsuoka, T. Saito, S. Yohena, A. Kamijo, N. Shiraga, M. Higuchi, K. Kimura, Fluorinated diamond-like carbon as antithrombogenic coating for blood-contacting devices, J. Biomed. Mater. Res. Part A 76 (1) (2006) 86–94. [19] S. Bai, T. Onodera, R. Nagumo, R. Miura, A. Suzuki, H. Tsuboi, N. Hatakeyama, H. Takaba, M. Kubo, A. Miyamoto, Friction reduction mechanism of hydrogen-and fluorine-terminated diamond-like carbon films investigated by molecular dynamics and quantum chemical calculation, J. Phys. Chem. C 116 (23) (2012) 12559–12565. [20] J. Wang, F. Wang, J. Li, Q. Sun, P. Yuan, Y. Jia, Comparative study of friction properties for hydrogen-and fluorine-modified diamond surfaces: a first-principles investigation, Surf. Sci. 608 (2013) 74–79. [21] S. Bhowmick, F. Sen, A. Banerji, A.T. Alpas, Friction and adhesion of fluorine containing hydrophobic hydrogenated diamond-like carbon (FH-DLC) coating against magnesium alloy AZ91, Surf. Coat. Technol. 267 (2015) 21–31. [22] C. Donnet, J. Fontaine, A. Grill, V. Patel, C. Jahnes, M. Belin, Wear-resistant fluorinated diamond-like carbon films, Surf. Coat. Technol. 94 (1997) 531–536. [23] F.G. Sen, Y. Qi, A.T. Alpas, Material transfer mechanisms between aluminum and fluorinated carbon interfaces, Acta Mater. 59 (7) (2011) 2601–2614. [24] S. Neuville, Quantum electronic mechanisms of atomic rearrangements during growth of hard carbon films, Surf. Coat. Technol. 206 (4) (2011) 703–726. [25] S. Neuville, Carbon Structure Analysis with Differentiated Raman Spectroscopy, LAP Academic Publishing, Saarbrucken Germany, 2014 1–176. [26] M. Rubio-Roy, E. Bertran, E. Pascual, M. Polo, J. Andújar, Fluorinated DLC deposited by pulsed-DC plasma for antisticking surface applications, Diam. Relat. Mater. 17 (7) (2008) 1728–1732. [27] Y. Pei, D. Galvan, J.T.M. De Hosson, Nanostructure and properties of TiC/a-C:H composite coatings, Acta Mater. 53 (17) (2005) 4505–4521. [28] A. Erdemir, C. Donnet, Tribology of diamond-like carbon films: recent progress and future prospects, J. Phys. D. Appl. Phys. 39 (18) (2006) R311. [29] E. Konca, Y.T. Cheng, A.M. Weiner, J.M. Dasch, A.T. Alpas, Elevated temperature tribological behavior of non-hydrogenated diamond-like carbon coatings against 319 aluminum alloy, Surf. Coat. Technol. 200 (12) (2006) 3996–4005.
159
[30] J. Fontaine, T. Le Mogne, J. Loubet, M. Belin, Achieving superlow friction with hydrogenated amorphous carbon: some key requirements, Thin Solid Films 482 (1) (2005) 99–108. [31] Y. Liu, A. Erdemir, E. Meletis, An investigation of the relationship between graphitization and frictional behavior of DLC coatings, Surf. Coat. Technol. 86 (1996) 564–568. [32] J. Sanchez-Lopez, A. Erdemir, C. Donnet, T. Rojas, Friction-induced structural transformations of diamondlike carbon coatings under various atmospheres, Surf. Coat. Technol. 163 (2003) 444–450. [33] S.M. Coman, M. Verziu, A. Tirsoaga, B. Jurca, C. Teodorescu, V. Kuncser, V.I. Parvulescu, G. Scholz, E. Kemnitz, NbF5–AlF3 catalysts: design, synthesis, and application in lactic acid synthesis from cellulose, ACS Catal. 5 (5) (2015) 3013–3026. [34] D. Kim, Y. Yu, K. Char, Characterization of aluminum oxyfluoride barrier in magnetic tunnel junctions, J. Appl. Phys. 96 (4) (2004) 2278–2285. [35] H. Fukui, M. Irie, Y. Utsumi, K. Oda, H. Ohara, An investigation of the wear track on DLC (a-C:H) film by time-of-flight secondary ion mass spectroscopy, Surf. Coat. Technol. 146 (2001) 378–383. [36] S.J. Park, K.R. Lee, D.H. Ko, Tribochemical reaction of hydrogenated diamond-like carbon films: a clue to understand the environmental dependence, Tribol. Int. 37 (11) (2004) 913–921. [37] A. Erdemir, C. Bindal, J. Pagan, P. Wilbur, Characterization of transfer layers on steel surfaces sliding against diamond-like hydrocarbon films in dry nitrogen, Surf. Coat. Technol. 76 (1995) 559–563. [38] J. Wei, W. Fong, D.B. Bogy, C.S. Bhatia, The decomposition mechanisms of a perfluoropolyether at the head/disk interface of hard disk drives, Tribol. Lett. 5 (2–3) (1998) 203–209. [39] P.H. Kasai, W.T. Tang, P. Wheeler, Degradation of perfluoropolyethers catalyzed by aluminum oxide, Appl. Surf. Sci. 51 (3) (1991) 201–211. [40] H. Liu, A. Tanaka, T. Kumagai, Influence of sliding mating materials on the tribological behavior of diamond-like carbon films, Thin Solid Films 352 (1) (1999) 145–150. [41] X. Meng-Burany, A.T. Alpas, FIB and TEM studies of damage mechanisms in DLC coatings sliding against aluminum, Thin Solid Films 516 (2) (2007) 325–335. [42] A. Konicek, D. Grierson, A. Sumant, T. Friedmann, J. Sullivan, P. Gilbert, W. Sawyer, R.W. Carpick, Influence of surface passivation on the friction and wear behavior of ultrananocrystalline diamond and tetrahedral amorphous carbon thin films, Phys. Rev. B 85 (15) (2012) 155448. [43] J. Fontaine, C. Donnet, A. Erdemir, Fundamentals of the tribology of DLC coatings, in: C. Donnet, A. Erdemir (Eds.), Tribology of Diamond-like Carbon Films, Springer 2008, pp. 139–154. [44] C. Bailey, S. Mukhopadhyay, A. Wander, B. Searle, N. Harrison, Structure and stability of α-AlF3 surfaces, J. Phys. Chem. C 113 (12) (2009) 4976–4983. [45] S. Neuville, A. Matthews, A perspective on the optimisation of hard carbon and related coatings for engineering applications, Thin Solid Films 515 (17) (2007) 6619–6653. [46] S. Neuville, New application perspective for tetrahedral amorphous carbon coatings, QScience Connect (2014).