- Email: [email protected]
Diallyl disulphide as natural organosulphur friction modifier via the in-situ tribo-chemical formation of tungsten disulphide
Applied Surface Science 428 (2018) 659–668
Contents lists available at ScienceDirect
Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc
Full length article
Diallyl disulphide as natural organosulphur friction modifier via the in-situ tribo-chemical formation of tungsten disulphide Manel Rodríguez Ripoll a,∗ , Vladimir Totolin a,∗ , Christoph Gabler a , Johannes Bernardi b , Ichiro Minami c a
AC2T research GmbH, Wiener Neustadt, Austria USTEM, Technische Universität Wien, Vienna, Austria c Division of Machine Elements, Luleå University of Technology, Luleå, Sweden b
a r t i c l e
i n f o
Article history: Received 17 May 2017 Received in revised form 1 September 2017 Accepted 13 September 2017 Available online 14 September 2017 Keywords: Diallyl disulphide Tungsten carbide Organosulphur compound Friction modifier Tungsten disulphide
a b s t r a c t The present work shows a novel method for generating in-situ low friction tribofilms containing tungsten disulphide in lubricated contacts using diallyl disulphide as sulphur precursor. The approach relies on the tribo-chemical interaction between the diallyl disulphide and a surface containing embedded sub-micrometer tungsten carbide particles. The results show that upon sliding contact between diallyl disulphide and the tungsten-containing surface, the coefficient of friction drops to values below 0.05 after an induction period. The reason for the reduction in friction is due to tribo-chemical reactions that leads to the in-situ formation of a complex tribofilm that contains iron and tungsten components. X-ray photoelectron spectroscopy analyses indicate the presence of tungsten disulphide at the contact interface, thus justifying the low coefficient of friction achieved during the sliding experiments. It was proven that the low friction tribofilms can only be formed by the coexistence of tungsten and sulphur species, thus highlighting the synergy between diallyl disulphide and the tungsten-containing surface. The concept of functionalizing surfaces to react with specific additives opens up a wide range of possibilities, which allows tuning on-site surfaces to target additive interactions. © 2017 Elsevier B.V. All rights reserved.
1. Introduction The reduction of friction in lubricated sliding components is a target pursued in order to reduce energy consumption and, as a consequence, our CO2 footprint. In order to achieve this goal, modern lubricants rely on functional additives in order to improve the properties of base oils. However, most of the commonly used additives pose environmental concerns more or less, in particular those composed of organic phosphorous, zinc and sulphur [1]. In recent years, nano-particles composed of transition metal dichalcogenides (TMDs) – such as MoS2 and WS2 – form emerged as potential substitutes of friction modifiers and anti-wear additives. TMDs have a lamellar structure with a strong bonding between the metal and chalcogenide atoms which contrasts with the weak chalcogenide–chalcogenide interaction between the layers, allowing them to easily slide over each other. This mechanism makes TMDs suitable to be used as solid lubricants. TMDs can be
∗ Corresponding authors at: AC2T research GmbH, Viktor−Kaplan−Strasse 2/C, 2700 Wiener Neustadt, Austria. E-mail addresses: [email protected] (M. Rodríguez Ripoll), [email protected] (V. Totolin). http://dx.doi.org/10.1016/j.apsusc.2017.09.100 0169-4332/© 2017 Elsevier B.V. All rights reserved.
obtained as nanoparticles in several morphologies, such as inorganic fullerenes or nanotubes [2–5]. In both cases, when mixed with additive-free oils such as poly-alpha-olefin (PAO), TMD nanoparticles are able to substantially reduce friction [6,7]. Among the main advantage of TMDs nanoparticles are their inertness, non-toxicity [8] and high thermal stability [9]. However, the interaction between TMD nanoparticles and conventional additives has been seldom investigated. While TMD nanoparticles seem to have a synergistic effect with ZDDP antiwear additive [10,11], their interaction with dispersants is antagonistic [12]. The latter point is crucial since dispersants are required for maintaining the lubrication performances. TMDs can also be applied directly to tribological components by various techniques. Magnetron sputtering is one of the most commonly used methods that is particularly employed in space applications, due to the excellent coating quality and good bonding with the substrate. Recently, W-S-C coatings doped with a large variety of metals have been proposed for enhancing the mechanical properties of TMDs coatings, while maintaining their low friction properties [13]. The major drawback of TMDs coatings in general is their poor performance in humid air, which limits their use mostly for vacuum and/or space applications.
660
M. Rodríguez Ripoll et al. / Applied Surface Science 428 (2018) 659–668
TMD tribofilms, in particular WS2 , can also be formed by many different routes involving coatings, bulk materials and fluids containing W and S in different states but not necessarily including crystalline WS2 . A summary of these routes was recently reviewed by Gustavsson and Jacobson [14]. One of these routes relies on the synergy between W-doped DLC coatings lubricated with Scontaining additives [15,16]. In this case the reduction in friction achieved under reciprocating sliding is rather modest (0.24 down to 0.15) and a major drawback is the consumption of the coating during the tribochemical reaction. A recent alternative to overcome this problem relies in the in-situ generation of WS2 using surfaces containing tungsten carbide embedded submicron particles. By these means, WS2 can be formed at the contact interface through tribo-chemical reactions, thus leading to the same low friction values (∼0.05) as those observed in TMD coatings but that are independent of humidity [17]. Furthermore, this method proved to be effective in overcoming the challenge of bringing the nanoparticles into the contact area in the presence of dispersants, since the W-containing particles are already initially embedded on the surface thus making dispersants superfluous. While this approach is promising and can be readily implemented in various engineering applications, the extreme pressure (EP) additives used as sulphur carrier in this study were sulphurised olefins. It has been reported that these type of EP additives may pose environmental concerns, in particular during their manufacturing process that requires the use of sulphur monochloride which leads to the presence of chlorine residues in the final product [18]. For this reason, the current study reports new fundamental insights regarding the in-situ generation of W-S containing tribofilms using lubricant additives that are conformed to the principles of green chemistry [19]. The aim of the present work is the in-situ generation of low friction W-S containing tribofilms using natural organosulphur compounds. The selected natural organosulphur compound, diallyl disulphide, is inexpensive and typically used in the food industry. Further, since food additives are foreseen to be used for human consumption, the toxicity and bioaccumulative properties of such compounds are thoroughly investigated, continuously monitored [20] and strictly regulated [21]. Diallyl disulphide is a non-polar organic compound typically found in garlic that is responsible for its characteristic odour. Diallyl disulphide is soluble in fats, oils and non-polar solvents such as hexane. Diallyl disulphide has been linked with potential health benefits, in particular with a reduction of cardiovascular diseases and some types of cancer [22]. The compound is green to the environment and it can be found on environmentally friendly nematicides [23]. It is also used in the food industry to improve the smell and taste of products. So far, only a single reference could be found in literature, where diallyl-disulphide was used in tribological applications, in particular as one of the components of natural garlic oil proposed as extreme pressure additive [24]. In this work, we exploit the presence of diallyl disulphide for generating in-situ WS2 under the presence of WC functionalised surfaces. The use of natural organosulphur compounds as lubricant additives in combination with WC functionalized surfaces offers a novel potential application as friction modifiers for this environmentally safe compound.
Fig. 1. Molecular structure of diallyl disulphide. Table 1 Summary of the test parameters. Normal load Stroke Frequency Test duration Counterbody Ball diameter
10 N 2 mm 25 Hz 14400 s 100Cr6 Ball 10 mm
mon synthetic base oil that is widely used in industrial components as well as automotive industry. The PAO base oil had a viscosity of 45.5 and 7.9 mm2 /s at 40 and 100 ◦ C, respectively. Its viscosity index (VI) according to ASTM D2270-04 was 146.2 and the density at 15 ◦ C was 0.83 g/cm3 . All the physical properties of the PAO base oil were obtained with a Stabinger viscometer SVM 3000 (Anton Paar GmbH, Austria). The prepared mixtures had a concentration of 1.5 and 10 wt% diallyl disulphide. Since diallyl disulphide is soluble in hydrocarbon oils, the mixture of PAO and diallyl disulphide was stable for al concentrations and no phase separation was observed even after 3 weeks after its initial formulation. Stainless steel AISI 304 surfaces were functionalized by embedding tungsten carbide (WC) particles with a diameter of 0.8 m using a machine hammer peening technique [25]. The WC particles were mixed with non-additivated oil until forming a homogeneous suspension, which was deposited on the surface of the steel samples. Afterwards, the sample was peened at a 90◦ angle with a hammer equipped with a semi-spherical WC ball (8 mm diameter) along a predefined path using an impact frequency of 200 Hz (Fig. 2). The impact energy of the hammer on the oil suspension resulted in the WC particles being mechanically embedded into the surface of the steel. The sample was peened line by line until the complete steel surface was hit by the hammer. The process was repeated twice in order to achieve a final surface coverage by the WC particles of 20%. A thorough description of the embedding process along with a detailed characterisation of the resulting functionalized surfaces can be found elsewhere [26]. 2.2. Tribological evaluation The tribological performance of the lubricant mixtures was ® evaluated under reciprocating sliding conditions using an SRV tribometer (Optimol, Germany). The tests were performed at an oscillating frequency of 25 Hz with a stroke of 2 mm. As counterbody, a 10 mm diameter 100Cr6 bearing steel ball was used. The normal load was set to 10 N. The tribological tests were performed at room temperature, following the testing parameters of our preliminary work (Table 1) [17]. Prior to the tests, tungsten carbide containing surfaces and the bearing steel balls were cleaned for 10 min in ultrasonic bath using toluene and petroleum ether. Afterwards the WC functionalized surfaces were tested for four hours under fully immersed contact conditions using 0.2 ml of lubricant mixture. Throughout the test, the coefficient of friction was recorded.
2. Experimental 2.3. Morphological and chemical analyses of the tribofilms 2.1. Lubricant mixtures and WC functionalized surfaces The lubricant mixtures were prepared using commercially available diallyl disulphide (4,5-dithia-1,7-octadiene, Sigma Aldrich, USA) (Fig. 1). The diallyl disulphide was mixed in poly-alpha-olefin 8 base oil (PAO), which was used as the carrier fluid. PAO is a com-
Scanning electron microscopy (SEM) with energy dispersive xray analyses (EDX) was used for evaluating the morphology and elemental composition of the generated tribofilms. A JEOL JSM 6500 F (Jeol, Japan) was operated using an acceleration voltage of 20 kV. The SEM was used in secondary electron and in backscattered
M. Rodríguez Ripoll et al. / Applied Surface Science 428 (2018) 659–668
661
Fig. 2. Schematic representation of the surface functionalization process along with a top view microscopy image of the resulting surface containing embedded WC particles.
electron mode, as required, in order to visualise the most relevant features of the tribofilms in comparison with the untested surface. The generated tribofilms were analysed at higher resolution using transmission electron microscopy (TEM). To this end, cross section lamellas were cut perpendicular to the sliding direction. The tribofilms were protected with organometallic Pt precursor and subsequently milled using focused ion beam. The TEM analysis were performed with a TECNAI F20 field emission TEM (FEI, Hilsboro, OR, USA) using an acceleration voltage of 200 kV. Scanning transmission images (STEM) were taken using a high angle annular dark field detector operated in dark field mode. EDX analysis were performed using an EDAX Apollo XLTW Silicon Drift Detector. The chemical composition of the tribofilms was investigated using X-ray photoelectron spectroscopy (XPS) with a Thermo Fisher Scientific Theta Probe (East Grinstead, UK) equipped with a monochromatic Al K␣ X-ray source (h = 1486.6 eV) and Ar+ ion gun. During the measurements, the base pressure inside the XPS chamber was kept at values in the range of 10−7 Pa. Prior to the XPS analysis, the tested samples were ultrasonically cleaned during 10 min in toluene followed by 10 additional minutes in petroleum ether (both HPLC grade). Afterwards, the samples were sputtered inside the XPS chamber using soft Ar+ for 20 s, with 3 kV and 1 A sputter current in order to remove the remaining contaminants. The sputtered area was approximately 3 × 3 mm. The elemental composition of the surface was obtained using survey scans acquired at a spot resolution of 400 m and a pass energy of 200 eV. The identified elements were then acquired using high resolution scans (at a pass energy of 50 eV) and the resulting binding energies were referenced to the adventitious carbon at a binding energy of 284.6 eV. XPS chemical mapping was performed both inside and outside the wear scar at a resolution of 100 m step size and 100 m spot diameter. All the acquired spectra were processed with the Average Data System software 5.945 (Thermo Fisher Scientific, East Grinstead, UK), using Gaussian/Lorentzian peak fitting for the high resolution scans. Raman spectroscopy analyses were performed on selected samples using a Micro-Raman spectrometer LabRam ARAMIS-UV Vis (Horiba, Japan). The samples were excited with a 532 nm wavelength laser. The spectra were acquire during 60 s over a frequency range from 200 to 2000 cm−1 .
3. Results 3.1. Friction and wear behaviour The frictional behaviour of the lubricant mixtures containing diallyl disulphide was investigated against WC functionalized surfaces under reciprocating sliding conditions (Fig. 3). The base PAO oil led to a steady coefficient of friction (COF) value of around 0.13, after an initial running in period. When adding a concentration of 1.5 wt% of diallyl disulphide to the PAO base oil, the coefficient of friction steadily decreased after an initial running in period at a friction level of 0.22. The final COF measured (∼0.07) was significantly lower than the values achieved with pure PAO, but no steady-state conditions were achieved throughout the four hours duration of the test. Additional experiments revealed that about 5 h are required to reach a steady-state COF under these conditions. Therefore, in order to enhance the performance of the lubricant mixture, the diallyl disulphide concentration was raised up to 10 wt%. In this case, the lubricant mixture led to an initial coefficient of friction of 0.19, but shortly afterwards, the COF started rapidly to decrease until a COF of 0.04 was reached in less than two hours after starting the rubbing process. Afterwards, the COF remained steady throughout the entire test. These results suggest that after an initial running-in period, the observed drop in friction might be due to the progressive formation of a tribofilm at the contact interface between the WC functionalized surface and the bearing steel ball. The performance of the tribofilm depends on the concentration of diallyl disulphide but in the best case is able to dramatically reduce friction to values below 0.05. This friction value lies within the value range reported in literature for WS2 in boundary lubricated contacts [27] and for W-S containing coatings in dry air contact conditions [13]. The role of tungsten at the contact interface was addressed by testing the best performing lubricant mixture (10 wt% diallyl disulphide) against conventional AISI 304 steel samples, using the same experimental conditions as described in Table 1. The results show that throughout the test, the coefficient of friction remains fairly stable at a value of about 0.15 (see supplementary material). In this case, the relatively high friction together with the lack of friction reduction as the experiment progresses is a hint that indicates the absence of low friction tribofilm formation during the test.
662
M. Rodríguez Ripoll et al. / Applied Surface Science 428 (2018) 659–668
Fig. 3. Representative friction curves of all tested lubricant mixtures against WC functionalized surfaces.
Fig. 4. Surface topography using secondary electrons of WC functionalized surfaces after being tested using a lubricant mixture of 10 wt% diallyl disulphide in PAO (a and b). Detail of the tribofilm formed using backscattered electron images (c and d).
The morphologies of the wear scars formed on the WC functionalized surfaces revealed the presence of a dark tribofilm under light microscopy (see supplementary material). No visible wear could be observed on the investigated functionalized surfaces. It is known that tungsten carbide particles have a superior wear resistance when compared to conventional steel substrates, even when tested under PAO (supplementary material) [26]. The low coefficient of friction values around 0.04 reported within the present section accompanied by lack of significant mechanical wear can only be achieved by the simultaneous combination of a sulphur car-
rier compound and the tungsten-containing functionalized surface. Under the absence of tungsten at the contact interface, the results showed a clearly deficient friction and wear performance and the values of the coefficient of friction obtained are by a factor of three higher compared to the values achieved when tungsten is present on the surface (supplementary material). These results show that diallyl disulphide per se is not able to provide any benefits in terms of friction under the selected testing conditions and emphasises its synergistic effect with the WC functionalized surfaces.
M. Rodríguez Ripoll et al. / Applied Surface Science 428 (2018) 659–668
663
Fig. 5. High resolution TEM images and their corresponding EDX maps of the low friction tribofilm formed using 10 wt% diallyl disulphide as friction modifier.
Table 2 Wear scar diameter as a function of diallyl disulphide concentration. Lubricant mixture
Wear scar diameter [m]
PAO 1.5 wt% diallyl disulphide in PAO 10 wt% diallyl disulphide in PAO
455 ± 39 523 ± 69 858 ± 14
Wear of the 100Cr6 bearing steel ball used as counterbody was addressed by measuring the wear scar diameters using light microscopy. The results are summarised in Table 2 and show that the wear scar increases for higher concentrations of diallyl disulphide. 3.2. Morphology and composition of the low friction tribofilms Hereinafter, the analyses of the tribofilms will be focused mainly on the solution of diallyl disulphide at the concentration of 10 wt%, since it showed the best friction performance and will serve to illustrate better the underlying friction mechanism. SEM micrographs using secondary electrons reveal the presence of a dark tribofilm covering the wear scar (Fig. 4a). Within the tribofilm, darker regions can be identified, which correspond to sulphur-rich areas (Fig. 4b). These dark dots homogeneously scattered throughout the surface are more clearly visible using backscattered electrons (Fig. 4c). By this means, it can be distinguish a tribofilm composed by the embedded tungsten carbide particles seen as white dots due to the large atomic mass, regions of light grey colour formed mainly by iron and oxygen and the dark spots found to be sulphur rich areas (Fig. 4d). Moreover, no signs of abrasion or material transfer could be observed in any of the SEM micrographs.
A cross sectional lamella obtained from the tribofilm was analysed using transmission electron microscopy (Fig. 5). The bright field images show the presence of discontinuous tribofilm with a thickness oscillating between 50 and 200 nm. Occasionally, tungsten carbide particles embedded in the surface can be observed in the vicinity of this tribofilm. EDX analyses reveal that the chemical composition of the tribofilm consists mainly of iron, oxygen, tungsten and sulphur. The sulphur and tungsten rich areas are overlapping, thus suggesting the formation of W-S compounds. The rest of the tribofilm is mostly iron rich. The homogenous oxygen distribution along both, the iron and the tungsten rich area indicates the possible presence of iron and tungsten oxides. The dark field images verify the complexity of the tribofilm. A top layer with a thickness of 5–10 nm can be identified, which is formed by lamellar structures. This thin top layer covers most of the surface and can be even found in regions where no tribofilm is present. The layer was too thin for making an accurate EDX analysis but its morphology and structure coincide with the tribofilms found by Polcar et al. on counterparts sliding against W-S-C/Cr coatings in humid air [28] and consequently, could be attributed to WS2 . Beneath the layer, a tribofilm composed of amorphous, crystalline and lamellar structures of different interspacing can be found indicating a complex mixture of iron and tungsten oxides and sulphides. Similar complex tribofilms containing iron and tungsten oxides and sulphides were observed in point contacts lubricated by dispersions containing PAO base oil with WS2 nanoparticles [29]. An additional TEM lamella obtained from the ball used as counteracting body reveals the occasional presence of a thin sulphur-rich area close to the surface (Fig. 6). The presence of a thin top layer covering most of the functionalized surface along with the
664
M. Rodríguez Ripoll et al. / Applied Surface Science 428 (2018) 659–668 Table 3 Relative surface atomic concentrations inside the wear scars as detected by the XPS analysis. Atomic orbital
Atomic concentration (%) 10 wt% diallyl disulphide in PAO
W4f S2p C1s O1s Fe2p
2.3 6.9 59.9 27.3 2.9
3.3. Chemical composition of the low friction tribofilms Fig. 6. High resolution TEM image of the tribofilm formed on the steel ball used as the counterbody.
presence of sulphur rich inclusions in the counteracting body suggest that the W-S tribofilm is formed on scattered spots across the functionalized surface and during sliding the counterbody smears it, enabling low friction.
The chemical composition of the tribofilms formed on the WCfunctionalized surface lubricated with 10 wt% diallyl disulphide in PAO is shown in Table 3. The high resolution W4f spectrum showed that the main component was WC, as outlined by the welldefined doublet W4f7/2 and W4f5/2 at binding energies of 31.4 eV and 33.6 eV, respectively (Fig. 7, top). Additionally, the spectrum reveals the presence of WS2 binding states (31.8 eV and 34.2 eV). The other binding states include WO2 and WO3 [17,28]. The presence of WS2 inside the wear scar generated on the WCfunctionalized surface lubricated with 10 wt% diallyl disulphide in
Fig. 7. High resolution XPS spectra of W4f (top) and S2p (bottom) for 10 wt% diallyl disulphide in PAO. The XPS analyses were performed inside the wear scars generated on the WC functionalized disc surfaces.
M. Rodríguez Ripoll et al. / Applied Surface Science 428 (2018) 659–668
Fig. 8. XPS chemical mapping of sulphur at a binding energy of 162.5 eV. Table 4 Relative surface atomic concentrations inside the wear scars as detected by the XPS analysis. Atomic orbital
Compound
Binding energy [eV]
Atomic%
1.5 wt% diallyl disulphide in PAO
10 wt% diallyl disulphide in PAO
C 1s
Graphite WC C O C O
284.6 283.0 286.2 289.2
25.8 1.7 4.2 2.4
14.0 1.2 – –
Fe 2p
Fe2 O3 Fe metallic Fe metallic FeSx Fe3 O4
710.8 706.8 707.9 713.9 709.2
8.7 – – 2.1 5.0
9.8 3.8 5.7 3.1 9.2
O 1s
Fe2 O3 WOx FeSO4
530.1 531.4 532.3
18.6 9.1 2.9
29.8 14.4 –
S 2p
FeS FeS2 SOx
161.4 162.7 167.5
8.1 6.1 1.6
5.7 2.7 –
W 4f
WC WS2 WO3 WO2
31.5 31.8 35.1 36.5
1.2 0.2 0.4 0.1
0.3 – 0.2 –
PAO is further outlined in the high resolution S2p peak from Fig. 7 (bottom). A doublet S2p3/2 and S2p1/2 could be detected at 162.5 eV and 163.7 eV, respectively and was ascribed to W-S bonds, according to prior findings [17,28,30]. The intensity of the WS2 peak is given by its relatively high surface atomic concentration (4.5 at.%). To further evidence the existence of WS2 type tribofilms on the WC-functionalized surfaces lubricated with 10 wt% diallyl disulphide in PAO, a XPS mapping was performed at one of the turning points of the shallow worn area as shown in Fig. 8. The distribution of sulphur inside and outside the worn area clearly showed a high intensity of S2p3/2 at a binding energy of 162.5 eV which is related to WS2 and was exclusively detected inside the worn area. This clearly implies that WS2 can only be formed by the frictional energy in the tribocontact via an in-situ tribo-chemical reaction. Moreover, it should be noted that the intensity is fairly homogeneous inside the worn area, indicating the presence of a continuous tribofilm. XPS analyses were also performed on the balls used as counterbody and are summarised in Table 4 for the two concentrations of diallyl disulphide used. The tribofilms formed on the surface of the balls mainly consisted of a combination of iron oxides and iron
665
sulphides mixed with tungsten oxides and carbides, probably due to the transfer of debris originated at the WC embedded particles. In this case, the presence of WS2 on the ball could not be detected by XPS. Raman spectroscopy measurements were performed inside the disc wear scars in order to gain a better insight into the formation of WS2 tribofilms and their role on the friction reduction mechanism. In this case, the selected samples were the surfaces lubricated with 1.5 wt% diallyl disulphide and tested for 4 and 8 h. As reported in Section 3.1 and shown in Fig. 3, the samples tested during 4 h using 1.5 wt% diallyl disulphide were able to substantially reduce friction but a steady coefficient of friction was not reached. The second sample tested during 8 h, reached a steady-state friction value of 0.075, slightly higher than the value obtained for 10 wt%. The Raman spectrum after 4 h test shows the presence of two peaks at 337 and 370 cm−1 and a small peak at 421 cm−1 (Fig. 9). The characteristic peaks for WS2 are at 352 and 421 cm−1 [31,32] even though some authors reported the first peak at 340 cm−1 when investigating W-S-N coatings [33]. The peaks corresponding to FeS2 are located at 340 and 375 cm−1 [34]. Regardless of the surface investigated, it can be clearly seen that the peak at 352 cm−1 is overlapping with the peaks at 337 and 370 cm−1 . After 4 h test, due to the low intensity of the peak at 421 cm−1 , the spectrum indicates that the tribofilm is mainly composed of FeS2 . After 8 h test, and coincident with a steady low friction, the peaks found in the Raman spectrum are identical, and the main highlight is the higher intensity of the peak at 421 cm−1 , which indicates a higher presence of WS2 inside the wear track. The peak found at 1550 cm−1 could not be fully identified, but it may be related to the presence of WC, whose peaks are located at 1340 and 1580 cm−1 [35]. 3.4. Reaction mechanism between diallyl disulphide and tungsten carbide It has been reported that dialkyl disulphides react with iron at rubbing surface and yield iron sulphide [36]. The detailed mechanism of iron sulphide formation is discussed in S. Plaza [37]. Although the previous papers focus on Fe alloys, we expect analogous reaction for WC at rubbing contact, because both Fe and W are transition elements whose reactivity is determined by the availability of vacant d-atomic orbitals. In particular, the electron configuration of W is [Xe]4f14 5d4 6s2 , so that W is ready to accept maximum 6 electrons on the 5d-atomic orbital. Diallyl disulphide can donate non-bonding electron pair (2 electrons) on S and - electrons (2 electrons) on double C C bond. W and the half moiety of diallyl disulphide can form a bidentate type intermediate through coordination of electrons, depicted as dotted lines (Fig. 10). A coordination of both moieties of diallyl disulphide to W C W is possible. It should be pointed out that disulphides with allyl, benzyl, or phenyl substituent exhibit better anti-wear properties compared to saturated hydrocarbons [38]. The reactions of transition metals with allylic compounds are well-known for allylation of nucleophiles [39]. A rearrangement of chemical bonds through radical reaction is likely. The reaction yields “allylated” WC (WC CH2 CH CH2 ) and thio-radical species accompanied with W S bond formation. The intra-molecular rearrangement of the radical species forms another W S bond (Fig. 10). The resultant radical species can further react with diallyl disulphide to form other W S bonds sequentially. The following friction reduction mechanism can be thus inferred from the results obtained (Fig. 11). During the rubbing process, the sulphur present in the lubricant mixtures reacts with the embedded WC particles as described. This tribo-chemical reaction, results in the localised formation of S rich spots containing WS2 , which are found scattered over the contact area. Thanks to the characteristic hexagonal crystal structure with chemical bonds between W
666
M. Rodríguez Ripoll et al. / Applied Surface Science 428 (2018) 659–668
Fig. 9. Raman spectra measured on functionalized surfaces tested using a lubricant mixture containing 1.5 wt% of diallyl disulphide during 4 and 8 h.
Fig. 10. Proposed reaction mechanism. Step 1: Interaction of W with Diallyl disulphide moiety and rearrangement of chemical bonds through radical reaction Step 2: Intra-molecular rearrangement and W S W S bond formation
and S atoms, WS2 has a 2D flat sheet structure [40]. The repulsion between layers results in easy slip between the lamellas that at the macroscopic level is evidenced by a low friction coefficient of about 0.04 [27]. Due to the high polarisation of the lamellas, WS2 has a good adhesion to metallic surfaces so that during rubbing, there is transfer of material to the counterbody, even this could not be measured by XPS in the present work. By this means the WS2 lamellas generated in individual spots are smeared over the contacting surface, as evidenced in the XPS mapping. Similar kinds of tribofilms have been reported in literature for contacts lubricated by base oil containing MoS2 nanoparticles [41]. In this case, the authors reported the presence of Mo in places of the contact interface, where no tribofilm was apparently visible, probably as a consequence of local exfoliation of MoS2 nanoparticles and smearing over the surface. In their case, this mechanism analogous to the one observed in the present work could not be proved, since the Mo signal may also indicate the presence of nanoparticles due to poor
cleaning of the surface. In our case, since no TMD were explicitly added to the base lubricants, this local generation of lamellas and smearing over the surface seems to be more plausible. Regarding the counterbody, the formation of a tribofilm that consisted mainly of iron oxides and sulphides as well as tungsten oxides and carbides, leads to a higher ball wear for higher concentrations of diallyl disulphide. This could be attributed to the higher relative atomic concentration of iron sulphides as well as to the presence of iron sulphates detected inside the wear scars generated on the surfaces lubricated with 1.5 wt% diallyl disulphide. Conversely, in case of 10 wt% diallyl disulphide, no iron sulphates could be detected inside the wears scars and the relative surface atomic concentration of iron sulphide was lower when compared to the 1.5 wt% diallyl disulphide. According to previous studies, both iron sulphide and iron sulphate are known to improve tribological properties of counteracting metal surfaces [42].
M. Rodríguez Ripoll et al. / Applied Surface Science 428 (2018) 659–668
667
Fig. 11. Schematic representation of the proposed lubrication mechanism.
4. Discussion The main potential advantages of the in-situ formation of low friction tribofilms when compared to previous approaches is that in contrast to TMD nanoparticles there is no need for producing stable suspensions, since in our case, the WC particles are embedded and homogeneously distributed across the surface. This concept requires the use of sulphur carrier additives in the base oil, but compared to nanoparticles, avoids the need of using dispersants for preparing stable emulsions. The latter point is crucial since dispersants have been reported to have an antagonistic effect with TMD nanoparticles by preventing adhesion of TMD lamellas at the contact interface, thus resulting in the lack of low friction tribofilm [12]. A further advantage of the in-situ generation of low friction tribofilms is that the tribofilm is formed on site and it allows achieving very low friction levels under humid air conditions. As highlighted in the present work, low friction tribofilms can be formed from natural products that are not able to provide friction modifier properties per se, but can reach this goal via a tribo-chemical reaction in synergy with a functionalized substrate. Conventionally, in many engineering applications where low friction is desired, this goal is achieved either by using suitable oil additives, not a few of them have the risk of environmental concern and that often rely on the presence of ferrous surfaces for being functional, or by using low friction TMD coatings that are not able to perform under humid environments. The performance of diallyl disulphide as friction modifier via the in-situ formation of WS2 is promising despite that the concentrations required (up to 10 wt%) are considerably higher than the concentrations of sulfurized olefin polysulfide EP additive used in our previous work. Further in the latter case, the low friction state was typically achieved between 1.5 and 2 h after rubbing, whereas for 10 wt% diallyl disulphide, the time required was slightly longer (between 2 and 2.5 h). The concept of functionalizing surfaces to react with specific additives opens up a wide range of possibilities, which allows tuning surfaces to target additive interactions. Our new approach deliberately exploits the synergy between lubricant and surface treatment and opens the door for the design and construction of tribologically effective systems. 5. Conclusions The present work has shown the feasibility of using diallyl disulphide as natural organosulphur friction modifier via the
in-situ formation of low friction W-S tribofilms, which are tribochemically formed at the contact interface on demand. The low friction tribofilms formed under the presence of diallyl disulphide and tungsten carbide have a chemical composition formed by iron and tungsten oxides and sulphides with a thickness of 50–200 m. On top of this tribofilm there is a thin layer of 5–10 nm formed by WS2 . The formation of these tribofilms requires the simultaneous presence of both, a sulphur carrier additive and a surface functionalized with tungsten carbide particles. Under the absence of either element, friction remains at much higher values. The in-situ generation mechanism of WS2 suggests that the tribo-chemical reaction occurs at individual spots scattered around the surface where WS2 lamellas are transferred to the counterbody and smeared over the surface, which results in a coefficient of friction of 0.04. The presented approached relies in the synergy between the lubricant additives and the material surface and enable the use of environmentally friendly compounds as efficient friction modifiers. Acknowledgments This work was funded by the Austrian COMET Programme (Project K2 XTribology. No. 849109) and carried out at the “Excellence Centre of Tribology”. The authors would like to thank Mr. Lukas Spiller for running the tribological tests. The functionalized surfaces were manufactured by Dr. Christoph Lechner (Technische Universität Wien). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.apsusc.2017.09. 100. References [1] W.J. Bartz, Tribol. Int. 31 (1998) 35–47. [2] R. Tenne, L. Margulis, M. Genut, G. Hodes, Nature 360 (1992) 444–446. [3] Y. Feldman, E. Wasserman, D.J. Srolovitz, R. Tenne, Science 267 (1995) 222–225. [4] L. Rapoport, Y. Bilik, Y. Feldman, M. Homyonfer, S.R. Cohen, R. Tenne, Nature 387 (1997) 791–793. ˇ [5] M. Remˇskar, Z. Skraba, M. Regula, C. Ballif, R. Sanjinés, F. Lévy, Adv. Mater. 10 (1998) 246–249. [6] J. Kogovˇsek, M. Remˇskar, A. Mrzel, M. Kalin, Tribol. Int. 61 (2013) 40–47. [7] M. Kalin, J. Kogovˇsek, M. Remˇskar, Wear 303 (2013) 480–485. [8] M. Pardo, T. Shuster-Meiseles, S. Levin-Zaidman, A. Rudich, Y. Rudich, Environ. Sci. Technol. 48 (2014) 3457–3466.
668
M. Rodríguez Ripoll et al. / Applied Surface Science 428 (2018) 659–668
[9] L. Chang, H. Yang, W. Fu, N. Yang, J. Chen, M. Li, G. Zou, J. Li, Mater. Res. Bull. 41 (2006) 1242–1248. [10] P.U. Aldana, B. Vacher, T. Le Mogne, M. Belin, B. Thiebaut, F. Dassenoy, Tribol. Lett. 56 (2014) 249–258. [11] A. Tomala, B. Vengudusamy, M. Rodríguez Ripoll, A. Naveira Suarez, M. Remˇskar, R. Rosentsveig, Tribol. Lett. 59 (2015) 1–18. [12] P. Rabaso, F. Dassenoy, F. Ville, M. Diaby, B. Vacher, T. Le Mogne, M. Belin, J. Cavoret, Tribol. Lett. 55 (2014) 503–516. [13] T. Polcar, A. Cavaleiro, Surf. Coatings Technol. 206 (2011) 686–695. [14] F. Gustavsson, S. Jacobson, Tribol. Int. 101 (2016) 340–347. [15] B. Podgornik, D. Hren, J. Viˇzintin, Thin Solid Films 476 (2005) 92–100. [16] B. Podgornik, D. Hren, J. Viˇzintin, S. Jacobson, N. Stavlid, S. Hogmark, Wear 261 (2006) 32–40. [17] V. Totolin, M. Rodríguez Ripoll, M. Jech, B. Podgornik, Tribol. Int. 94 (2016) 269–278. [18] L.R. Rudnick, Lubricant Additives: Chemistry and Applications, 2nd ed., CRC Press, 2009. [19] P. Anastas, N. Eghbali, Green Chemistry: Principles and Practice, 2010. [20] Commission Regulation (EU), No 257/2010 of 25 March 2010 Setting up a Programme for the Re-evaluation of Approved Food Additives in Accordance with Regulation (EC) No 1333/2008 of the European Parliament and of the Council on Food Additives, 2010. [21] Regulation (EC), No 1331/2008 of the European Parliament and of the Council of 16 December 2008 Establishing a Common Authorisation Procedure for Food Additives, Food Enzymes and Food Flavourings, 2008. [22] L. Yi, Q. Su, Food Chem. Toxicol. 57 (2013) 362–370. [23] I.-K. Park, J.-Y. Park, K.-H. Kim, K.-S. Choi, I.-H. Choi, C.-S. Kim, S.-C. Shin, Nematology 7 (2005) 767–774.
[24] W. Li, C. Jiang, M. Chao, X. Wang, ACS Sustain. Chem. Eng. 2 (2014) 798–803. [25] F. Bleicher, C. Lechner, C. Habersohn, M. Obermair, F. Heindl, M. Rodriguez Ripoll, CIRP Ann. Manuf. Technol. 62 (2013) 239–242. [26] M.R. Ripoll, F. Heindl, C. Lechner, V. Totolin, M. Jech, F. Bleicher, Tribol. Trans. 60 (2017) 479–489. [27] V.B. Niste, M. Ratoi, Mater. Today Commun. 8 (2016) 1–11. [28] T. Polcar, F. Gustavsson, T. Thersleff, S. Jacobson, A. Cavaleiro, Faraday Discuss. 156 (2012) 383–401. [29] M. Ratoi, V.B. Niste, J. Walker, J. Zekonyte, Tribol. Lett. 52 (2013) 81–91. [30] NIST X-Ray Photoelectron Spectroscopy Database, 2016 http://srdata.nist. gov/xps/. [31] L. Joly-Pottuz, F. Dassenoy, M. Belin, B. Vacher, J.M. Martin, N. Fleischer, Tribol. Lett. 18 (2005) 477–485. [32] T.W. Scharf, A. Rajendran, R. Banerjee, F. Sequeda, Thin Solid Films 517 (2009) 5666–5675. [33] J. Sundberg, H. Nyberg, E. Särhammar, T. Nyberg, S. Jacobson, U. Jansson, Surf. Coatings Technol. 258 (2014) 86–94. [34] M. Miyajima, K. Kitamura, K. Matsumoto, Tribol. Online 11 (2016) 382–388. [35] X. Zhou, Y. Qiu, J. Yu, J. Yin, S. Gao, Int. J. Hydrogen Energy 36 (2011) 7398–7404. [36] J. Tannous, B.M.I. de Bouchet, T. Le-Mogne, P. Charles, J.M. Martin, Tribol. Mater. Surfaces Interfaces 1 (2007) 98–104. [37] S. Plaza, A S L E Trans. 30 (1986) 493–500. [38] K.G. Allum, E.S. Forbes, J. Inst. Pet. 53 (1967) 173–185. [39] J. Tsuji, I. Minami, Acc. Chem. Res. 20 (1987) 140–145. [40] E. Gibney, Nature 522 (2015) 274–276. [41] J. Kogovˇsek, M. Kalin, Tribol. Lett. 53 (2014) 585–597. [42] V. Totolin, N. Ranetcaia, V. Hamciuc, N. Shore, N. Dörr, C. Ibanescu, B.C. Simionescu, V. Harabagiu, Tribol. Int. 67 (2013) 1–10.
Contents lists available at ScienceDirect
Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc
Full length article
Diallyl disulphide as natural organosulphur friction modifier via the in-situ tribo-chemical formation of tungsten disulphide Manel Rodríguez Ripoll a,∗ , Vladimir Totolin a,∗ , Christoph Gabler a , Johannes Bernardi b , Ichiro Minami c a
AC2T research GmbH, Wiener Neustadt, Austria USTEM, Technische Universität Wien, Vienna, Austria c Division of Machine Elements, Luleå University of Technology, Luleå, Sweden b
a r t i c l e
i n f o
Article history: Received 17 May 2017 Received in revised form 1 September 2017 Accepted 13 September 2017 Available online 14 September 2017 Keywords: Diallyl disulphide Tungsten carbide Organosulphur compound Friction modifier Tungsten disulphide
a b s t r a c t The present work shows a novel method for generating in-situ low friction tribofilms containing tungsten disulphide in lubricated contacts using diallyl disulphide as sulphur precursor. The approach relies on the tribo-chemical interaction between the diallyl disulphide and a surface containing embedded sub-micrometer tungsten carbide particles. The results show that upon sliding contact between diallyl disulphide and the tungsten-containing surface, the coefficient of friction drops to values below 0.05 after an induction period. The reason for the reduction in friction is due to tribo-chemical reactions that leads to the in-situ formation of a complex tribofilm that contains iron and tungsten components. X-ray photoelectron spectroscopy analyses indicate the presence of tungsten disulphide at the contact interface, thus justifying the low coefficient of friction achieved during the sliding experiments. It was proven that the low friction tribofilms can only be formed by the coexistence of tungsten and sulphur species, thus highlighting the synergy between diallyl disulphide and the tungsten-containing surface. The concept of functionalizing surfaces to react with specific additives opens up a wide range of possibilities, which allows tuning on-site surfaces to target additive interactions. © 2017 Elsevier B.V. All rights reserved.
1. Introduction The reduction of friction in lubricated sliding components is a target pursued in order to reduce energy consumption and, as a consequence, our CO2 footprint. In order to achieve this goal, modern lubricants rely on functional additives in order to improve the properties of base oils. However, most of the commonly used additives pose environmental concerns more or less, in particular those composed of organic phosphorous, zinc and sulphur [1]. In recent years, nano-particles composed of transition metal dichalcogenides (TMDs) – such as MoS2 and WS2 – form emerged as potential substitutes of friction modifiers and anti-wear additives. TMDs have a lamellar structure with a strong bonding between the metal and chalcogenide atoms which contrasts with the weak chalcogenide–chalcogenide interaction between the layers, allowing them to easily slide over each other. This mechanism makes TMDs suitable to be used as solid lubricants. TMDs can be
∗ Corresponding authors at: AC2T research GmbH, Viktor−Kaplan−Strasse 2/C, 2700 Wiener Neustadt, Austria. E-mail addresses: [email protected] (M. Rodríguez Ripoll), [email protected] (V. Totolin). http://dx.doi.org/10.1016/j.apsusc.2017.09.100 0169-4332/© 2017 Elsevier B.V. All rights reserved.
obtained as nanoparticles in several morphologies, such as inorganic fullerenes or nanotubes [2–5]. In both cases, when mixed with additive-free oils such as poly-alpha-olefin (PAO), TMD nanoparticles are able to substantially reduce friction [6,7]. Among the main advantage of TMDs nanoparticles are their inertness, non-toxicity [8] and high thermal stability [9]. However, the interaction between TMD nanoparticles and conventional additives has been seldom investigated. While TMD nanoparticles seem to have a synergistic effect with ZDDP antiwear additive [10,11], their interaction with dispersants is antagonistic [12]. The latter point is crucial since dispersants are required for maintaining the lubrication performances. TMDs can also be applied directly to tribological components by various techniques. Magnetron sputtering is one of the most commonly used methods that is particularly employed in space applications, due to the excellent coating quality and good bonding with the substrate. Recently, W-S-C coatings doped with a large variety of metals have been proposed for enhancing the mechanical properties of TMDs coatings, while maintaining their low friction properties [13]. The major drawback of TMDs coatings in general is their poor performance in humid air, which limits their use mostly for vacuum and/or space applications.
660
M. Rodríguez Ripoll et al. / Applied Surface Science 428 (2018) 659–668
TMD tribofilms, in particular WS2 , can also be formed by many different routes involving coatings, bulk materials and fluids containing W and S in different states but not necessarily including crystalline WS2 . A summary of these routes was recently reviewed by Gustavsson and Jacobson [14]. One of these routes relies on the synergy between W-doped DLC coatings lubricated with Scontaining additives [15,16]. In this case the reduction in friction achieved under reciprocating sliding is rather modest (0.24 down to 0.15) and a major drawback is the consumption of the coating during the tribochemical reaction. A recent alternative to overcome this problem relies in the in-situ generation of WS2 using surfaces containing tungsten carbide embedded submicron particles. By these means, WS2 can be formed at the contact interface through tribo-chemical reactions, thus leading to the same low friction values (∼0.05) as those observed in TMD coatings but that are independent of humidity [17]. Furthermore, this method proved to be effective in overcoming the challenge of bringing the nanoparticles into the contact area in the presence of dispersants, since the W-containing particles are already initially embedded on the surface thus making dispersants superfluous. While this approach is promising and can be readily implemented in various engineering applications, the extreme pressure (EP) additives used as sulphur carrier in this study were sulphurised olefins. It has been reported that these type of EP additives may pose environmental concerns, in particular during their manufacturing process that requires the use of sulphur monochloride which leads to the presence of chlorine residues in the final product [18]. For this reason, the current study reports new fundamental insights regarding the in-situ generation of W-S containing tribofilms using lubricant additives that are conformed to the principles of green chemistry [19]. The aim of the present work is the in-situ generation of low friction W-S containing tribofilms using natural organosulphur compounds. The selected natural organosulphur compound, diallyl disulphide, is inexpensive and typically used in the food industry. Further, since food additives are foreseen to be used for human consumption, the toxicity and bioaccumulative properties of such compounds are thoroughly investigated, continuously monitored [20] and strictly regulated [21]. Diallyl disulphide is a non-polar organic compound typically found in garlic that is responsible for its characteristic odour. Diallyl disulphide is soluble in fats, oils and non-polar solvents such as hexane. Diallyl disulphide has been linked with potential health benefits, in particular with a reduction of cardiovascular diseases and some types of cancer [22]. The compound is green to the environment and it can be found on environmentally friendly nematicides [23]. It is also used in the food industry to improve the smell and taste of products. So far, only a single reference could be found in literature, where diallyl-disulphide was used in tribological applications, in particular as one of the components of natural garlic oil proposed as extreme pressure additive [24]. In this work, we exploit the presence of diallyl disulphide for generating in-situ WS2 under the presence of WC functionalised surfaces. The use of natural organosulphur compounds as lubricant additives in combination with WC functionalized surfaces offers a novel potential application as friction modifiers for this environmentally safe compound.
Fig. 1. Molecular structure of diallyl disulphide. Table 1 Summary of the test parameters. Normal load Stroke Frequency Test duration Counterbody Ball diameter
10 N 2 mm 25 Hz 14400 s 100Cr6 Ball 10 mm
mon synthetic base oil that is widely used in industrial components as well as automotive industry. The PAO base oil had a viscosity of 45.5 and 7.9 mm2 /s at 40 and 100 ◦ C, respectively. Its viscosity index (VI) according to ASTM D2270-04 was 146.2 and the density at 15 ◦ C was 0.83 g/cm3 . All the physical properties of the PAO base oil were obtained with a Stabinger viscometer SVM 3000 (Anton Paar GmbH, Austria). The prepared mixtures had a concentration of 1.5 and 10 wt% diallyl disulphide. Since diallyl disulphide is soluble in hydrocarbon oils, the mixture of PAO and diallyl disulphide was stable for al concentrations and no phase separation was observed even after 3 weeks after its initial formulation. Stainless steel AISI 304 surfaces were functionalized by embedding tungsten carbide (WC) particles with a diameter of 0.8 m using a machine hammer peening technique [25]. The WC particles were mixed with non-additivated oil until forming a homogeneous suspension, which was deposited on the surface of the steel samples. Afterwards, the sample was peened at a 90◦ angle with a hammer equipped with a semi-spherical WC ball (8 mm diameter) along a predefined path using an impact frequency of 200 Hz (Fig. 2). The impact energy of the hammer on the oil suspension resulted in the WC particles being mechanically embedded into the surface of the steel. The sample was peened line by line until the complete steel surface was hit by the hammer. The process was repeated twice in order to achieve a final surface coverage by the WC particles of 20%. A thorough description of the embedding process along with a detailed characterisation of the resulting functionalized surfaces can be found elsewhere [26]. 2.2. Tribological evaluation The tribological performance of the lubricant mixtures was ® evaluated under reciprocating sliding conditions using an SRV tribometer (Optimol, Germany). The tests were performed at an oscillating frequency of 25 Hz with a stroke of 2 mm. As counterbody, a 10 mm diameter 100Cr6 bearing steel ball was used. The normal load was set to 10 N. The tribological tests were performed at room temperature, following the testing parameters of our preliminary work (Table 1) [17]. Prior to the tests, tungsten carbide containing surfaces and the bearing steel balls were cleaned for 10 min in ultrasonic bath using toluene and petroleum ether. Afterwards the WC functionalized surfaces were tested for four hours under fully immersed contact conditions using 0.2 ml of lubricant mixture. Throughout the test, the coefficient of friction was recorded.
2. Experimental 2.3. Morphological and chemical analyses of the tribofilms 2.1. Lubricant mixtures and WC functionalized surfaces The lubricant mixtures were prepared using commercially available diallyl disulphide (4,5-dithia-1,7-octadiene, Sigma Aldrich, USA) (Fig. 1). The diallyl disulphide was mixed in poly-alpha-olefin 8 base oil (PAO), which was used as the carrier fluid. PAO is a com-
Scanning electron microscopy (SEM) with energy dispersive xray analyses (EDX) was used for evaluating the morphology and elemental composition of the generated tribofilms. A JEOL JSM 6500 F (Jeol, Japan) was operated using an acceleration voltage of 20 kV. The SEM was used in secondary electron and in backscattered
M. Rodríguez Ripoll et al. / Applied Surface Science 428 (2018) 659–668
661
Fig. 2. Schematic representation of the surface functionalization process along with a top view microscopy image of the resulting surface containing embedded WC particles.
electron mode, as required, in order to visualise the most relevant features of the tribofilms in comparison with the untested surface. The generated tribofilms were analysed at higher resolution using transmission electron microscopy (TEM). To this end, cross section lamellas were cut perpendicular to the sliding direction. The tribofilms were protected with organometallic Pt precursor and subsequently milled using focused ion beam. The TEM analysis were performed with a TECNAI F20 field emission TEM (FEI, Hilsboro, OR, USA) using an acceleration voltage of 200 kV. Scanning transmission images (STEM) were taken using a high angle annular dark field detector operated in dark field mode. EDX analysis were performed using an EDAX Apollo XLTW Silicon Drift Detector. The chemical composition of the tribofilms was investigated using X-ray photoelectron spectroscopy (XPS) with a Thermo Fisher Scientific Theta Probe (East Grinstead, UK) equipped with a monochromatic Al K␣ X-ray source (h = 1486.6 eV) and Ar+ ion gun. During the measurements, the base pressure inside the XPS chamber was kept at values in the range of 10−7 Pa. Prior to the XPS analysis, the tested samples were ultrasonically cleaned during 10 min in toluene followed by 10 additional minutes in petroleum ether (both HPLC grade). Afterwards, the samples were sputtered inside the XPS chamber using soft Ar+ for 20 s, with 3 kV and 1 A sputter current in order to remove the remaining contaminants. The sputtered area was approximately 3 × 3 mm. The elemental composition of the surface was obtained using survey scans acquired at a spot resolution of 400 m and a pass energy of 200 eV. The identified elements were then acquired using high resolution scans (at a pass energy of 50 eV) and the resulting binding energies were referenced to the adventitious carbon at a binding energy of 284.6 eV. XPS chemical mapping was performed both inside and outside the wear scar at a resolution of 100 m step size and 100 m spot diameter. All the acquired spectra were processed with the Average Data System software 5.945 (Thermo Fisher Scientific, East Grinstead, UK), using Gaussian/Lorentzian peak fitting for the high resolution scans. Raman spectroscopy analyses were performed on selected samples using a Micro-Raman spectrometer LabRam ARAMIS-UV Vis (Horiba, Japan). The samples were excited with a 532 nm wavelength laser. The spectra were acquire during 60 s over a frequency range from 200 to 2000 cm−1 .
3. Results 3.1. Friction and wear behaviour The frictional behaviour of the lubricant mixtures containing diallyl disulphide was investigated against WC functionalized surfaces under reciprocating sliding conditions (Fig. 3). The base PAO oil led to a steady coefficient of friction (COF) value of around 0.13, after an initial running in period. When adding a concentration of 1.5 wt% of diallyl disulphide to the PAO base oil, the coefficient of friction steadily decreased after an initial running in period at a friction level of 0.22. The final COF measured (∼0.07) was significantly lower than the values achieved with pure PAO, but no steady-state conditions were achieved throughout the four hours duration of the test. Additional experiments revealed that about 5 h are required to reach a steady-state COF under these conditions. Therefore, in order to enhance the performance of the lubricant mixture, the diallyl disulphide concentration was raised up to 10 wt%. In this case, the lubricant mixture led to an initial coefficient of friction of 0.19, but shortly afterwards, the COF started rapidly to decrease until a COF of 0.04 was reached in less than two hours after starting the rubbing process. Afterwards, the COF remained steady throughout the entire test. These results suggest that after an initial running-in period, the observed drop in friction might be due to the progressive formation of a tribofilm at the contact interface between the WC functionalized surface and the bearing steel ball. The performance of the tribofilm depends on the concentration of diallyl disulphide but in the best case is able to dramatically reduce friction to values below 0.05. This friction value lies within the value range reported in literature for WS2 in boundary lubricated contacts [27] and for W-S containing coatings in dry air contact conditions [13]. The role of tungsten at the contact interface was addressed by testing the best performing lubricant mixture (10 wt% diallyl disulphide) against conventional AISI 304 steel samples, using the same experimental conditions as described in Table 1. The results show that throughout the test, the coefficient of friction remains fairly stable at a value of about 0.15 (see supplementary material). In this case, the relatively high friction together with the lack of friction reduction as the experiment progresses is a hint that indicates the absence of low friction tribofilm formation during the test.
662
M. Rodríguez Ripoll et al. / Applied Surface Science 428 (2018) 659–668
Fig. 3. Representative friction curves of all tested lubricant mixtures against WC functionalized surfaces.
Fig. 4. Surface topography using secondary electrons of WC functionalized surfaces after being tested using a lubricant mixture of 10 wt% diallyl disulphide in PAO (a and b). Detail of the tribofilm formed using backscattered electron images (c and d).
The morphologies of the wear scars formed on the WC functionalized surfaces revealed the presence of a dark tribofilm under light microscopy (see supplementary material). No visible wear could be observed on the investigated functionalized surfaces. It is known that tungsten carbide particles have a superior wear resistance when compared to conventional steel substrates, even when tested under PAO (supplementary material) [26]. The low coefficient of friction values around 0.04 reported within the present section accompanied by lack of significant mechanical wear can only be achieved by the simultaneous combination of a sulphur car-
rier compound and the tungsten-containing functionalized surface. Under the absence of tungsten at the contact interface, the results showed a clearly deficient friction and wear performance and the values of the coefficient of friction obtained are by a factor of three higher compared to the values achieved when tungsten is present on the surface (supplementary material). These results show that diallyl disulphide per se is not able to provide any benefits in terms of friction under the selected testing conditions and emphasises its synergistic effect with the WC functionalized surfaces.
M. Rodríguez Ripoll et al. / Applied Surface Science 428 (2018) 659–668
663
Fig. 5. High resolution TEM images and their corresponding EDX maps of the low friction tribofilm formed using 10 wt% diallyl disulphide as friction modifier.
Table 2 Wear scar diameter as a function of diallyl disulphide concentration. Lubricant mixture
Wear scar diameter [m]
PAO 1.5 wt% diallyl disulphide in PAO 10 wt% diallyl disulphide in PAO
455 ± 39 523 ± 69 858 ± 14
Wear of the 100Cr6 bearing steel ball used as counterbody was addressed by measuring the wear scar diameters using light microscopy. The results are summarised in Table 2 and show that the wear scar increases for higher concentrations of diallyl disulphide. 3.2. Morphology and composition of the low friction tribofilms Hereinafter, the analyses of the tribofilms will be focused mainly on the solution of diallyl disulphide at the concentration of 10 wt%, since it showed the best friction performance and will serve to illustrate better the underlying friction mechanism. SEM micrographs using secondary electrons reveal the presence of a dark tribofilm covering the wear scar (Fig. 4a). Within the tribofilm, darker regions can be identified, which correspond to sulphur-rich areas (Fig. 4b). These dark dots homogeneously scattered throughout the surface are more clearly visible using backscattered electrons (Fig. 4c). By this means, it can be distinguish a tribofilm composed by the embedded tungsten carbide particles seen as white dots due to the large atomic mass, regions of light grey colour formed mainly by iron and oxygen and the dark spots found to be sulphur rich areas (Fig. 4d). Moreover, no signs of abrasion or material transfer could be observed in any of the SEM micrographs.
A cross sectional lamella obtained from the tribofilm was analysed using transmission electron microscopy (Fig. 5). The bright field images show the presence of discontinuous tribofilm with a thickness oscillating between 50 and 200 nm. Occasionally, tungsten carbide particles embedded in the surface can be observed in the vicinity of this tribofilm. EDX analyses reveal that the chemical composition of the tribofilm consists mainly of iron, oxygen, tungsten and sulphur. The sulphur and tungsten rich areas are overlapping, thus suggesting the formation of W-S compounds. The rest of the tribofilm is mostly iron rich. The homogenous oxygen distribution along both, the iron and the tungsten rich area indicates the possible presence of iron and tungsten oxides. The dark field images verify the complexity of the tribofilm. A top layer with a thickness of 5–10 nm can be identified, which is formed by lamellar structures. This thin top layer covers most of the surface and can be even found in regions where no tribofilm is present. The layer was too thin for making an accurate EDX analysis but its morphology and structure coincide with the tribofilms found by Polcar et al. on counterparts sliding against W-S-C/Cr coatings in humid air [28] and consequently, could be attributed to WS2 . Beneath the layer, a tribofilm composed of amorphous, crystalline and lamellar structures of different interspacing can be found indicating a complex mixture of iron and tungsten oxides and sulphides. Similar complex tribofilms containing iron and tungsten oxides and sulphides were observed in point contacts lubricated by dispersions containing PAO base oil with WS2 nanoparticles [29]. An additional TEM lamella obtained from the ball used as counteracting body reveals the occasional presence of a thin sulphur-rich area close to the surface (Fig. 6). The presence of a thin top layer covering most of the functionalized surface along with the
664
M. Rodríguez Ripoll et al. / Applied Surface Science 428 (2018) 659–668 Table 3 Relative surface atomic concentrations inside the wear scars as detected by the XPS analysis. Atomic orbital
Atomic concentration (%) 10 wt% diallyl disulphide in PAO
W4f S2p C1s O1s Fe2p
2.3 6.9 59.9 27.3 2.9
3.3. Chemical composition of the low friction tribofilms Fig. 6. High resolution TEM image of the tribofilm formed on the steel ball used as the counterbody.
presence of sulphur rich inclusions in the counteracting body suggest that the W-S tribofilm is formed on scattered spots across the functionalized surface and during sliding the counterbody smears it, enabling low friction.
The chemical composition of the tribofilms formed on the WCfunctionalized surface lubricated with 10 wt% diallyl disulphide in PAO is shown in Table 3. The high resolution W4f spectrum showed that the main component was WC, as outlined by the welldefined doublet W4f7/2 and W4f5/2 at binding energies of 31.4 eV and 33.6 eV, respectively (Fig. 7, top). Additionally, the spectrum reveals the presence of WS2 binding states (31.8 eV and 34.2 eV). The other binding states include WO2 and WO3 [17,28]. The presence of WS2 inside the wear scar generated on the WCfunctionalized surface lubricated with 10 wt% diallyl disulphide in
Fig. 7. High resolution XPS spectra of W4f (top) and S2p (bottom) for 10 wt% diallyl disulphide in PAO. The XPS analyses were performed inside the wear scars generated on the WC functionalized disc surfaces.
M. Rodríguez Ripoll et al. / Applied Surface Science 428 (2018) 659–668
Fig. 8. XPS chemical mapping of sulphur at a binding energy of 162.5 eV. Table 4 Relative surface atomic concentrations inside the wear scars as detected by the XPS analysis. Atomic orbital
Compound
Binding energy [eV]
Atomic%
1.5 wt% diallyl disulphide in PAO
10 wt% diallyl disulphide in PAO
C 1s
Graphite WC C O C O
284.6 283.0 286.2 289.2
25.8 1.7 4.2 2.4
14.0 1.2 – –
Fe 2p
Fe2 O3 Fe metallic Fe metallic FeSx Fe3 O4
710.8 706.8 707.9 713.9 709.2
8.7 – – 2.1 5.0
9.8 3.8 5.7 3.1 9.2
O 1s
Fe2 O3 WOx FeSO4
530.1 531.4 532.3
18.6 9.1 2.9
29.8 14.4 –
S 2p
FeS FeS2 SOx
161.4 162.7 167.5
8.1 6.1 1.6
5.7 2.7 –
W 4f
WC WS2 WO3 WO2
31.5 31.8 35.1 36.5
1.2 0.2 0.4 0.1
0.3 – 0.2 –
PAO is further outlined in the high resolution S2p peak from Fig. 7 (bottom). A doublet S2p3/2 and S2p1/2 could be detected at 162.5 eV and 163.7 eV, respectively and was ascribed to W-S bonds, according to prior findings [17,28,30]. The intensity of the WS2 peak is given by its relatively high surface atomic concentration (4.5 at.%). To further evidence the existence of WS2 type tribofilms on the WC-functionalized surfaces lubricated with 10 wt% diallyl disulphide in PAO, a XPS mapping was performed at one of the turning points of the shallow worn area as shown in Fig. 8. The distribution of sulphur inside and outside the worn area clearly showed a high intensity of S2p3/2 at a binding energy of 162.5 eV which is related to WS2 and was exclusively detected inside the worn area. This clearly implies that WS2 can only be formed by the frictional energy in the tribocontact via an in-situ tribo-chemical reaction. Moreover, it should be noted that the intensity is fairly homogeneous inside the worn area, indicating the presence of a continuous tribofilm. XPS analyses were also performed on the balls used as counterbody and are summarised in Table 4 for the two concentrations of diallyl disulphide used. The tribofilms formed on the surface of the balls mainly consisted of a combination of iron oxides and iron
665
sulphides mixed with tungsten oxides and carbides, probably due to the transfer of debris originated at the WC embedded particles. In this case, the presence of WS2 on the ball could not be detected by XPS. Raman spectroscopy measurements were performed inside the disc wear scars in order to gain a better insight into the formation of WS2 tribofilms and their role on the friction reduction mechanism. In this case, the selected samples were the surfaces lubricated with 1.5 wt% diallyl disulphide and tested for 4 and 8 h. As reported in Section 3.1 and shown in Fig. 3, the samples tested during 4 h using 1.5 wt% diallyl disulphide were able to substantially reduce friction but a steady coefficient of friction was not reached. The second sample tested during 8 h, reached a steady-state friction value of 0.075, slightly higher than the value obtained for 10 wt%. The Raman spectrum after 4 h test shows the presence of two peaks at 337 and 370 cm−1 and a small peak at 421 cm−1 (Fig. 9). The characteristic peaks for WS2 are at 352 and 421 cm−1 [31,32] even though some authors reported the first peak at 340 cm−1 when investigating W-S-N coatings [33]. The peaks corresponding to FeS2 are located at 340 and 375 cm−1 [34]. Regardless of the surface investigated, it can be clearly seen that the peak at 352 cm−1 is overlapping with the peaks at 337 and 370 cm−1 . After 4 h test, due to the low intensity of the peak at 421 cm−1 , the spectrum indicates that the tribofilm is mainly composed of FeS2 . After 8 h test, and coincident with a steady low friction, the peaks found in the Raman spectrum are identical, and the main highlight is the higher intensity of the peak at 421 cm−1 , which indicates a higher presence of WS2 inside the wear track. The peak found at 1550 cm−1 could not be fully identified, but it may be related to the presence of WC, whose peaks are located at 1340 and 1580 cm−1 [35]. 3.4. Reaction mechanism between diallyl disulphide and tungsten carbide It has been reported that dialkyl disulphides react with iron at rubbing surface and yield iron sulphide [36]. The detailed mechanism of iron sulphide formation is discussed in S. Plaza [37]. Although the previous papers focus on Fe alloys, we expect analogous reaction for WC at rubbing contact, because both Fe and W are transition elements whose reactivity is determined by the availability of vacant d-atomic orbitals. In particular, the electron configuration of W is [Xe]4f14 5d4 6s2 , so that W is ready to accept maximum 6 electrons on the 5d-atomic orbital. Diallyl disulphide can donate non-bonding electron pair (2 electrons) on S and - electrons (2 electrons) on double C C bond. W and the half moiety of diallyl disulphide can form a bidentate type intermediate through coordination of electrons, depicted as dotted lines (Fig. 10). A coordination of both moieties of diallyl disulphide to W C W is possible. It should be pointed out that disulphides with allyl, benzyl, or phenyl substituent exhibit better anti-wear properties compared to saturated hydrocarbons [38]. The reactions of transition metals with allylic compounds are well-known for allylation of nucleophiles [39]. A rearrangement of chemical bonds through radical reaction is likely. The reaction yields “allylated” WC (WC CH2 CH CH2 ) and thio-radical species accompanied with W S bond formation. The intra-molecular rearrangement of the radical species forms another W S bond (Fig. 10). The resultant radical species can further react with diallyl disulphide to form other W S bonds sequentially. The following friction reduction mechanism can be thus inferred from the results obtained (Fig. 11). During the rubbing process, the sulphur present in the lubricant mixtures reacts with the embedded WC particles as described. This tribo-chemical reaction, results in the localised formation of S rich spots containing WS2 , which are found scattered over the contact area. Thanks to the characteristic hexagonal crystal structure with chemical bonds between W
666
M. Rodríguez Ripoll et al. / Applied Surface Science 428 (2018) 659–668
Fig. 9. Raman spectra measured on functionalized surfaces tested using a lubricant mixture containing 1.5 wt% of diallyl disulphide during 4 and 8 h.
Fig. 10. Proposed reaction mechanism. Step 1: Interaction of W with Diallyl disulphide moiety and rearrangement of chemical bonds through radical reaction Step 2: Intra-molecular rearrangement and W S W S bond formation
and S atoms, WS2 has a 2D flat sheet structure [40]. The repulsion between layers results in easy slip between the lamellas that at the macroscopic level is evidenced by a low friction coefficient of about 0.04 [27]. Due to the high polarisation of the lamellas, WS2 has a good adhesion to metallic surfaces so that during rubbing, there is transfer of material to the counterbody, even this could not be measured by XPS in the present work. By this means the WS2 lamellas generated in individual spots are smeared over the contacting surface, as evidenced in the XPS mapping. Similar kinds of tribofilms have been reported in literature for contacts lubricated by base oil containing MoS2 nanoparticles [41]. In this case, the authors reported the presence of Mo in places of the contact interface, where no tribofilm was apparently visible, probably as a consequence of local exfoliation of MoS2 nanoparticles and smearing over the surface. In their case, this mechanism analogous to the one observed in the present work could not be proved, since the Mo signal may also indicate the presence of nanoparticles due to poor
cleaning of the surface. In our case, since no TMD were explicitly added to the base lubricants, this local generation of lamellas and smearing over the surface seems to be more plausible. Regarding the counterbody, the formation of a tribofilm that consisted mainly of iron oxides and sulphides as well as tungsten oxides and carbides, leads to a higher ball wear for higher concentrations of diallyl disulphide. This could be attributed to the higher relative atomic concentration of iron sulphides as well as to the presence of iron sulphates detected inside the wear scars generated on the surfaces lubricated with 1.5 wt% diallyl disulphide. Conversely, in case of 10 wt% diallyl disulphide, no iron sulphates could be detected inside the wears scars and the relative surface atomic concentration of iron sulphide was lower when compared to the 1.5 wt% diallyl disulphide. According to previous studies, both iron sulphide and iron sulphate are known to improve tribological properties of counteracting metal surfaces [42].
M. Rodríguez Ripoll et al. / Applied Surface Science 428 (2018) 659–668
667
Fig. 11. Schematic representation of the proposed lubrication mechanism.
4. Discussion The main potential advantages of the in-situ formation of low friction tribofilms when compared to previous approaches is that in contrast to TMD nanoparticles there is no need for producing stable suspensions, since in our case, the WC particles are embedded and homogeneously distributed across the surface. This concept requires the use of sulphur carrier additives in the base oil, but compared to nanoparticles, avoids the need of using dispersants for preparing stable emulsions. The latter point is crucial since dispersants have been reported to have an antagonistic effect with TMD nanoparticles by preventing adhesion of TMD lamellas at the contact interface, thus resulting in the lack of low friction tribofilm [12]. A further advantage of the in-situ generation of low friction tribofilms is that the tribofilm is formed on site and it allows achieving very low friction levels under humid air conditions. As highlighted in the present work, low friction tribofilms can be formed from natural products that are not able to provide friction modifier properties per se, but can reach this goal via a tribo-chemical reaction in synergy with a functionalized substrate. Conventionally, in many engineering applications where low friction is desired, this goal is achieved either by using suitable oil additives, not a few of them have the risk of environmental concern and that often rely on the presence of ferrous surfaces for being functional, or by using low friction TMD coatings that are not able to perform under humid environments. The performance of diallyl disulphide as friction modifier via the in-situ formation of WS2 is promising despite that the concentrations required (up to 10 wt%) are considerably higher than the concentrations of sulfurized olefin polysulfide EP additive used in our previous work. Further in the latter case, the low friction state was typically achieved between 1.5 and 2 h after rubbing, whereas for 10 wt% diallyl disulphide, the time required was slightly longer (between 2 and 2.5 h). The concept of functionalizing surfaces to react with specific additives opens up a wide range of possibilities, which allows tuning surfaces to target additive interactions. Our new approach deliberately exploits the synergy between lubricant and surface treatment and opens the door for the design and construction of tribologically effective systems. 5. Conclusions The present work has shown the feasibility of using diallyl disulphide as natural organosulphur friction modifier via the
in-situ formation of low friction W-S tribofilms, which are tribochemically formed at the contact interface on demand. The low friction tribofilms formed under the presence of diallyl disulphide and tungsten carbide have a chemical composition formed by iron and tungsten oxides and sulphides with a thickness of 50–200 m. On top of this tribofilm there is a thin layer of 5–10 nm formed by WS2 . The formation of these tribofilms requires the simultaneous presence of both, a sulphur carrier additive and a surface functionalized with tungsten carbide particles. Under the absence of either element, friction remains at much higher values. The in-situ generation mechanism of WS2 suggests that the tribo-chemical reaction occurs at individual spots scattered around the surface where WS2 lamellas are transferred to the counterbody and smeared over the surface, which results in a coefficient of friction of 0.04. The presented approached relies in the synergy between the lubricant additives and the material surface and enable the use of environmentally friendly compounds as efficient friction modifiers. Acknowledgments This work was funded by the Austrian COMET Programme (Project K2 XTribology. No. 849109) and carried out at the “Excellence Centre of Tribology”. The authors would like to thank Mr. Lukas Spiller for running the tribological tests. The functionalized surfaces were manufactured by Dr. Christoph Lechner (Technische Universität Wien). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.apsusc.2017.09. 100. References [1] W.J. Bartz, Tribol. Int. 31 (1998) 35–47. [2] R. Tenne, L. Margulis, M. Genut, G. Hodes, Nature 360 (1992) 444–446. [3] Y. Feldman, E. Wasserman, D.J. Srolovitz, R. Tenne, Science 267 (1995) 222–225. [4] L. Rapoport, Y. Bilik, Y. Feldman, M. Homyonfer, S.R. Cohen, R. Tenne, Nature 387 (1997) 791–793. ˇ [5] M. Remˇskar, Z. Skraba, M. Regula, C. Ballif, R. Sanjinés, F. Lévy, Adv. Mater. 10 (1998) 246–249. [6] J. Kogovˇsek, M. Remˇskar, A. Mrzel, M. Kalin, Tribol. Int. 61 (2013) 40–47. [7] M. Kalin, J. Kogovˇsek, M. Remˇskar, Wear 303 (2013) 480–485. [8] M. Pardo, T. Shuster-Meiseles, S. Levin-Zaidman, A. Rudich, Y. Rudich, Environ. Sci. Technol. 48 (2014) 3457–3466.
668
M. Rodríguez Ripoll et al. / Applied Surface Science 428 (2018) 659–668
[9] L. Chang, H. Yang, W. Fu, N. Yang, J. Chen, M. Li, G. Zou, J. Li, Mater. Res. Bull. 41 (2006) 1242–1248. [10] P.U. Aldana, B. Vacher, T. Le Mogne, M. Belin, B. Thiebaut, F. Dassenoy, Tribol. Lett. 56 (2014) 249–258. [11] A. Tomala, B. Vengudusamy, M. Rodríguez Ripoll, A. Naveira Suarez, M. Remˇskar, R. Rosentsveig, Tribol. Lett. 59 (2015) 1–18. [12] P. Rabaso, F. Dassenoy, F. Ville, M. Diaby, B. Vacher, T. Le Mogne, M. Belin, J. Cavoret, Tribol. Lett. 55 (2014) 503–516. [13] T. Polcar, A. Cavaleiro, Surf. Coatings Technol. 206 (2011) 686–695. [14] F. Gustavsson, S. Jacobson, Tribol. Int. 101 (2016) 340–347. [15] B. Podgornik, D. Hren, J. Viˇzintin, Thin Solid Films 476 (2005) 92–100. [16] B. Podgornik, D. Hren, J. Viˇzintin, S. Jacobson, N. Stavlid, S. Hogmark, Wear 261 (2006) 32–40. [17] V. Totolin, M. Rodríguez Ripoll, M. Jech, B. Podgornik, Tribol. Int. 94 (2016) 269–278. [18] L.R. Rudnick, Lubricant Additives: Chemistry and Applications, 2nd ed., CRC Press, 2009. [19] P. Anastas, N. Eghbali, Green Chemistry: Principles and Practice, 2010. [20] Commission Regulation (EU), No 257/2010 of 25 March 2010 Setting up a Programme for the Re-evaluation of Approved Food Additives in Accordance with Regulation (EC) No 1333/2008 of the European Parliament and of the Council on Food Additives, 2010. [21] Regulation (EC), No 1331/2008 of the European Parliament and of the Council of 16 December 2008 Establishing a Common Authorisation Procedure for Food Additives, Food Enzymes and Food Flavourings, 2008. [22] L. Yi, Q. Su, Food Chem. Toxicol. 57 (2013) 362–370. [23] I.-K. Park, J.-Y. Park, K.-H. Kim, K.-S. Choi, I.-H. Choi, C.-S. Kim, S.-C. Shin, Nematology 7 (2005) 767–774.
[24] W. Li, C. Jiang, M. Chao, X. Wang, ACS Sustain. Chem. Eng. 2 (2014) 798–803. [25] F. Bleicher, C. Lechner, C. Habersohn, M. Obermair, F. Heindl, M. Rodriguez Ripoll, CIRP Ann. Manuf. Technol. 62 (2013) 239–242. [26] M.R. Ripoll, F. Heindl, C. Lechner, V. Totolin, M. Jech, F. Bleicher, Tribol. Trans. 60 (2017) 479–489. [27] V.B. Niste, M. Ratoi, Mater. Today Commun. 8 (2016) 1–11. [28] T. Polcar, F. Gustavsson, T. Thersleff, S. Jacobson, A. Cavaleiro, Faraday Discuss. 156 (2012) 383–401. [29] M. Ratoi, V.B. Niste, J. Walker, J. Zekonyte, Tribol. Lett. 52 (2013) 81–91. [30] NIST X-Ray Photoelectron Spectroscopy Database, 2016 http://srdata.nist. gov/xps/. [31] L. Joly-Pottuz, F. Dassenoy, M. Belin, B. Vacher, J.M. Martin, N. Fleischer, Tribol. Lett. 18 (2005) 477–485. [32] T.W. Scharf, A. Rajendran, R. Banerjee, F. Sequeda, Thin Solid Films 517 (2009) 5666–5675. [33] J. Sundberg, H. Nyberg, E. Särhammar, T. Nyberg, S. Jacobson, U. Jansson, Surf. Coatings Technol. 258 (2014) 86–94. [34] M. Miyajima, K. Kitamura, K. Matsumoto, Tribol. Online 11 (2016) 382–388. [35] X. Zhou, Y. Qiu, J. Yu, J. Yin, S. Gao, Int. J. Hydrogen Energy 36 (2011) 7398–7404. [36] J. Tannous, B.M.I. de Bouchet, T. Le-Mogne, P. Charles, J.M. Martin, Tribol. Mater. Surfaces Interfaces 1 (2007) 98–104. [37] S. Plaza, A S L E Trans. 30 (1986) 493–500. [38] K.G. Allum, E.S. Forbes, J. Inst. Pet. 53 (1967) 173–185. [39] J. Tsuji, I. Minami, Acc. Chem. Res. 20 (1987) 140–145. [40] E. Gibney, Nature 522 (2015) 274–276. [41] J. Kogovˇsek, M. Kalin, Tribol. Lett. 53 (2014) 585–597. [42] V. Totolin, N. Ranetcaia, V. Hamciuc, N. Shore, N. Dörr, C. Ibanescu, B.C. Simionescu, V. Harabagiu, Tribol. Int. 67 (2013) 1–10.