Home Physical Sciences Thermally Induced Superlow Friction of DLC Films in Ambient Air
Article Open Access

Thermally Induced Superlow Friction of DLC Films in Ambient Air

  • Qunfeng Zeng EMAIL logo
Published/Copyright: June 19, 2018

Abstract

Thermally induced superlow friction (0.008) of diamond-like carbon (DLC) films was achieved in ambient air in the present work. Raman and XPS (X-ray Photoelectron Spectroscopy) measurements and analyses show that superlow friction of the annealed DLC films is involved in the transformation of sp3 to sp2 hybridized carbon during annealing and the tribochemical reactions during sliding. The thermally activated graphitization and oxidation of the annealed DLC films in ambient air is beneficial to form the positively charged interface and achieve the stable superlow friction. A friction model was developed and applied to explain superlow friction, which is attributed to Van de Waals force between graphite layers and the repulsive force between hydroxyl group of graphite oxide and hydrogen terminated DLC films surface.

Introduction

Diamond-like carbon (DLC) films have been widely used as solid lubricants to improve the tribological properties of machinery parts such as high-speed roller bearing in space especially under dry friction conditions [1, 2]. In spite of the strong potential space application, the use of DLC films is still limited due to high internal residual stress and low thermal stability in DLC films [3]. Therefore, it is necessary to improve the thermal stability and reduce the internal stress of DLC films by the treatment technique. The annealing treatment is already widely used to reduce the internal stress of DLC films by researchers. Mosaner et al. found that the annealing treatment can reduce the internal stress of DLC films [4]. Zhang et al. also observed the decrease of the internal stress and hardness in the annealed DLC films [5]. Huang et al. found there are no significant changes in the structure and tribological properties below 200 °C, however, above 200 °C there is the graphite-like structure, which results in the poor tribological properties. Therefore, it is concluded that DLC films may retain excellent performances under appropriate annealing temperatures. Silicon nitride is the preferred material for high-speed bearing ball due to its superior hardness and low density. It is expected by combining the advantages of Si3N4 and DLC films. Thus DLC films coated with Si3N4 ball are beneficial to improve the bearing performance of Si3N4 ball and prolong the service lifetime of high-speed bearing. It is well known that the temperature rise is unavoidable even though the oil lubrication or oil–air lubrication is applied for high-speed bearing [7]. The typical temperature rise from room temperature (RT) is around 100 °C. Therefore, it is reasonable and acceptable that the tribotests are performed at the tribotest temperature of 100 °C.

Superlow friction of DLC films are usually related with the hydrogen or hydroxyl group-terminated surface [8]. However, DLC films are sensitive to the oxidizing species of oxygen and water molecular and even oxidized especially in ambient air under high temperature. Therefore, it is necessary to design the terminated surface and reduce the internal stress of DLC films with superlow friction through the annealing treatment and elucidate the microstructure change of DLC films and friction mechanism of the friction pair.

Experiments

First, DLC films were deposited on Si3Ni4 ball with the diameter of 9.5 mm and high-speed steel (HSS) flat with the diameter of 30 mm and the thickness of 5 mm in an unbalanced magnetron sputtering system with direct current bias voltage. The specimens were biased to a negative voltage of 1 kV. The argon ion was etched the specimens to remove the native oxides and contamination on the surface of specimens. The deposition power was 1,000 W. The deposition pressure was 0.3 Pa. The sputtering target is high pure of 99.99 % graphite. The argon and acetylene were introduced as carbon source of DLC films to deposit hydrogen-incorporated DLC films. The thickness of DLC films is about 2 μm. Surface roughness of DLC films is about 0.04 µm.

DLC films coated with Si3N4 ball were postannealed in a chamber-type electric resistance furnace under different temperatures in ambient air. The annealing time is 1 h. The annealing temperatures are 200°C, 400°C and 600°C, respectively. The specimens are heated from RT to the set temperature, and then held at this temperature for 1 h, and finally cooled to RT. The tribotests of the specimens were carried out by a ball on disc tribometer in ambient air. The tribotest temperatures are RT and 100°C. The disc was heated to 100°C before the tribotests at the tribotest temperature of 100°C. The speed is 0.05 m/s and the normal load is 10 N. Relative humidity is about 50 %. The microstructure of the annealed DLC films was analyzed by Raman and XPS, respectively. Before XPS analysis, argon is used to etch the surface of specimen and remove the contaminations on the top surface of specimen in vacuum. The surface topography was observed by atomic force microscopy (AFM) and optical microscope, respectively.

Results and discussion

Figure 1 shows AFM surface topography of DLC films on the steel substrate flat. It is found that there are lots of spherical particles with nanoscale size on the steel substrate surface. These particles are distributed uniformly on the surface of the steel substrate.

Figure 1: AFM surface topography of DLC films: (a) 2D, (b) 3D.
Figure 1:

AFM surface topography of DLC films: (a) 2D, (b) 3D.

Figure 2 shows the Gaussian fitting Raman spectra of the annealed DLC films coated with Si3N4 ball under different annealing temperatures before the tribotest. Table 1 shows the Raman fitting results of the annealed DLC films coated with Si3N4 ball. Typical Raman spectroscopy of the annealed DLC films is shown in Figure 2(a). D peak becomes obvious; D and G peaks shift to high wave number, FWHM of G peak decreases (from 183 cm−1 to 82 cm−1) and ID/IG (from 0.76 to 1.86) increases from RT to 600 °C of the annealing temperatures. These phenomena mean that the sp2 content increases with increase of the annealing temperature, revealing the graphitization and transformation of sp3 to sp2 hybridized carbon during annealing. At 400 °C of the annealing temperature, there are two obvious bands at 1,397 cm−1 and 1,590 cm−1 in the annealed DLC films. D peak becomes quite broad and high intense, while G peak becomes weak, which indicates that the content of graphite increases with increase of the annealing temperature in the annealed DLC films [9]. The Raman results show that the annealed DLC films are graphitized and even oxidized by oxygen or water in ambient air during annealing [10]. With increase of the annealing temperature, the width of G peak becomes narrow and the intensity of G peak becomes low, which means there occurs the crystalline graphite in annealed DLC films [11]. The width of G peak becomes narrow and FWHM shifts to low wave number (from 183 cm−1 to 82 cm−1) from RT to 600 °C, which also means there occurs the crystallite graphite in annealed DLC films [12]. The annealing treatment especially in high temperature leads to the effusion of hydrogen in the annealed DLC films [6, 13]. Therefore, the appropriate annealing temperature is 400°C for DLC films due to the heating-induced graphitization and dehydrogenation of microstructure changes and oxidation during annealing.

Figure 2: Raman spectroscopy of the annealed DLC films under different annealing temperatures: (a) as-is, (b) 200oC, (c) 400oC, (d) 600oC.
Figure 2:

Raman spectroscopy of the annealed DLC films under different annealing temperatures: (a) as-is, (b) 200oC, (c) 400oC, (d) 600oC.

Table 1:

The Raman Gaussian fitting results of the annealed DLC films.

Annealing temperature (oC)D peak (cm−1)D peak FWHM (cm−1)G peak (cm−1)G peak FWHM (cm−1)ID/IG
RT1,3593401,5431830.76
2001,3912981,5621501.08
4001,3973081,590901.36
6001,3813421,603821.86

Figure 3 shows CoF of the annealed DLC films in ambient air at RT and 100°C of the tribotest temperature. CoF of as-is DLC films increases along with the sliding distance. CoF of DLC films annealed at 200 °C increases to the maximum value and then decreases to around 0.092. CoF of DLC films annealed at 400 °C and 600 °C decreases slightly along with the sliding distance. DLC films annealed at 400°C exhibits the lowest CoF (0.04).

Figure 3: CoF of the annealed DLC films under different tribotest temperature: (a) RT, (b) 100oC.
Figure 3:

CoF of the annealed DLC films under different tribotest temperature: (a) RT, (b) 100oC.

For 100oC of the tribotest temperature, it is surprisingly found that at stable stage CoF of DLC films annealed at 400 °C is about 0.008, which exhibits superlow friction. CoF is 0.2 at the initial stage, and then decreases to 0.008 after 1,000 s at the stable stage. The whole sliding process is divided into three stages: the running-in stage, low friction stage and superlow friction stage. The first fall of CoF decreases from 0.2 to 0.02, whereas the second fall decreases from 0.02 to 0.008, entering the superlow friction regime.

Superlow friction of DLC films annealed at 400°C is achieved at 100°C of the tribotest temperature, which is closely related with the tribochemical reaction between the contact surfaces and water molecular or oxygen during sliding. Raman is first used to measure the microstructure within the wear track of DLC films annealed at 400°C. Figure 4 shows the Raman spectroscopy of the worn surface in the annealed DLC ​​films. It is found that there are the green-colored films within the wear scar on ball. There are 447 cm−1 and 598 cm−1 bands in nonwear scar; however, the bands shift to 796 cm−1 and 946 cm−1 in wear scar of the annealed DLC films. The peak at 447 cm−1 is Si vibration peak and the peak at 598 cm−1 is Si-O vibration peak [14, 15]. The peak at 796 cm−1 is β-SiC standard peak and the peak at 946 cm−1 corresponds to β-SiC optical phonon mode. It can be inferred that Si is doped into DLC films during annealing especially in the top surface of DLC films. And then Si-O bond cleavage is broken and then reacted with carbon to form Si-C bond due to tribochemical reaction under high frictional heating and high pressure during sliding, which enhances load capacity, improves antifriction behavior and reduces the adhesive wear of the annealed DLC films on the ball.

Figure 4: Raman spectroscopy of DLC films annealed at 400°C on ball.
Figure 4:

Raman spectroscopy of DLC films annealed at 400°C on ball.

Figure 5(a) shows the XPS C1s spectra of worn surface of DLC films annealed at 400°C on ball. XPS bands are first assigned as sp2 carbon (284.6 eV) and C=O (288.0 eV). The peak at 284.6 eV has the same position as that of graphite, indicating aromatic carbon atom. Other species are characteristic of C-Si, aliphatic chains (C-C and C-H) and oxidized carbon chemical groups like hydroxyl (-COH) and carbonyl (-C=O), respectively [16, 17]. A shoulder appears at 283.8 eV, which is assigned to C-Si bonds. These findings agree with Raman measurements of the annealed DLC films on ball. Note an asymmetric C-C peak at 284.6 eV; therefore, it is estimated there is crystalline graphite in the annealed DLC films. The peak at 285.9 eV is assigned to C-OH group or C-N bond. Figure 5(b) shows XPS O 1s photopeak. The oxygen is bonded with C and Si due to tribochemical reactions. The peaks at 528.8 eV and 529.3 eV show NO and the lattice oxygen of iron oxide phase. The peaks at 529.6 eV and 530.6 eV are attributed to O=C-OH and OH groups [18]. The peak at 530.1 eV is attributed to O in water [19]. The peak at 531.2 eV maybe C=O [20]. A peak around 532 eV is associated with Si–O–Si vibration [21]. These results also agreed with the Raman measurements of the annealed DLC films. C-O and C=O bonds are attributed to the oxidation of graphite. It is obvious that DLC films have undergone chemical degradations. Since carbon is present as aliphatic carbon, carbon–oxygen bond indicates the oxidation of annealed DLC films. According to Raman and XPS analyses, it is presumed that the terminated surface of the annealed DLC films is formed by hydroxyl group.

Figure 5: XPS of wear scar of DLC films annealed at 400°C on ball: (a) C1s, (b) O1s.
Figure 5:

XPS of wear scar of DLC films annealed at 400°C on ball: (a) C1s, (b) O1s.

Figure 6(a) shows the microstructure of wear scar on the flat at 100°C of the tribotest temperature. D peak is obvious, G peak is 1,587 cm−1 and ID/IG is 1.7. Figure 6(b) shows XPS C 1s photopeak. These species are aliphatic chains (C-C and C-H) and oxidized carbon chemical groups like ether (-C-O), hydroxyl (-COH), carbonyl (-C=O) and ester/acid C(O)OH, respectively. The bands at 284.2 and 284.8 eV are C-C bond and C-H bond, respectively. The peak at 286.9 eV is carbon of oxidation. C-C or C-H has transitions in aromatic systems. The peak at 291.8 eV is graphite or carbon oxygen double bond. The peak of 293.6 eV is carboxylate carbon. There exits the terminated surface with –COH or C(O)OH on the surface of DLC films on the flat. These results are beneficial to understand the tribochemical reaction and superlow friction mechanism of DLC films.

Figure 6: The microstructure of wear scar of DLC films on disc: (a) Raman, (b) XPS.
Figure 6:

The microstructure of wear scar of DLC films on disc: (a) Raman, (b) XPS.

Superlow friction of DLC films is usually correlated with hydrogen or hydroxyl group. Therefore, it is supposed that superlow friction is probably associated with the graphitization and oxidation of carbon induced by the environmental energy and friction heating. In the light of the experimental results obtained here, the graphitization and oxidation process can be explained as follows: (1) dehydrogenation, (2) graphitization. The sp3 carbon bond is transferred to sp2 carbon bond in annealed DLC films. (3) The formation of the terminated surface. Graphite oxide is formed on the top surface of the annealed DLC films and the top surface is terminated by hydroxyl group in ambient air under high annealing temperature and friction heating. Oxidative cleavage of C-C bond to carbonyl compounds is an essential operation in organic synthesis. First, at the initial stage of oxidation, carbon atom at the edge and defect of graphite is easily oxidized to graphite oxide under high annealing temperature. Oxygen and hydroxyl group are reacted chemically with π bond in graphite, forming C-O bond at the edge of graphite layer. Due to C-O and hydroxyl group at edge of graphite, the gap between graphite layers would become high, which enhances the oxidation of graphite and intercalation of hydroxyl group between graphite layers. Then, the intercalated graphite layers are gradually formed on the top surface of the annealed DLC films. Due to the tribochemical reaction, hydroxyl group is slightly moved to the center of graphite layer, which separates graphite layers largely and reduces the Van de Waals force between graphite layers. Figure 7 shows the oxidation process of the annealed DLC films during the annealing treatment.

Figure 7: The schematic illustrations of the oxidation of graphite in the annealed DLC films.
Figure 7:

The schematic illustrations of the oxidation of graphite in the annealed DLC films.

The published models cannot be used to explain superlow friction behavior in the present case. Therefore, a new model is proposed to explain superlow friction of the annealed DLC films. Before the tribotest, the graphite oxide and hydrated graphite oxide were already formed on the top surface of the 400°C annealed DLC films [22]. For RT tribotest of the friction pair, CoF is high at the initial stage, which causes high friction heating and improves the graphitization and oxidation of DLC films on disc. For 100°C tribotest, DLC films on disc are heated to 100°C before tribotest. CoF is also high at the initial stage but decreases rapidly to superlow friction. The top surface of DLC films on disc may be oxidized partially to form the terminated surface with hydroxyl group under high friction heating. Hydroxyl groups increases the space of graphite layers on ball simultaneously, which results in low van der Waals force during sliding. It is found from Figure 3 that CoF at the initial stage is high and it takes long time to achieve superlow and stable CoF (0.008) at the stable stage. Under low-temperature annealing treatment, local internal stress can be released in DLC films and the excellent mechanical properties like high hardness and Young’s modulus of DLC films are retained. However, the transformation of sp3 to sp2 and the realignment of carbon structure are appeared by the effusion of hydrogen under high temperature and pressure. Raman and XPS measurements support the formation of the terminated surface with hydroxyl group of the annealed DLC films. Therefore, we propose a new model to explain superlow friction of the friction pair. Figure 8 shows the schematic illustrations of partially dihydrated hydrogen in DLC films and graphite oxide of the annealed DLC films on interfaces. The sliding of the friction pair occurs between the intercalation hydroxyl groups with graphite layers and the shielding of hydrogen or hydroxyl group on the top surface of DLC films on disc, forming the superlow friction interface. On the basis of the surface characterization, it can be inferred that Si is doped into DLC films on ball under high temperature during annealing. During sliding, DLC films on Si3N4 ball were partially worn out and under high friction heating and pressure Si3N4 can react with water to form few SiO2, which leads to high CoF and high friction heat at the initial stage in ambient air. Subsequently, Si-O bond cleavage is broken and reacted with carbon to form Si-C bond, which results in low CoF and enhances the load capacity and reduces wear resistance of the annealed DLC films. DLC films on flat were partially oxidized by oxygen or water molecules and hydrogen or hydroxyl groups of the chemical products are terminated carbon, which leads to superlow friction of the annealed DLC films. Therefore, there is the presence of van der Waals forces between graphite layers and the dipole–dipole effects that form an interfacial Coulomb repulsive force between two contact surfaces with hydroxyl groups or hydrogen-terminated graphite oxide and DLC films, which results in superlow friction, as shown in Figure 8.

Figure 8: Schematic illustrations of partially dihydrated hydrogen of the friction pair on interface.
Figure 8:

Schematic illustrations of partially dihydrated hydrogen of the friction pair on interface.

Conclusions

In summary, the increase in ID/IG ratios from 0.76 at RT to 1.86 at 600°C of the annealing temperature and the wave number of G peak shifting high are experimentally observed. This is associated with graphitization and oxidation of the annealed DLC films. It is found that 400°C is the appropriate annealing temperature for DLC films because DLC films annealed at 400°C exhibit superlow friction of 0.008 at 100°C of the tribotest temperature. Superlow friction mechanism is proposed to the synergism of Van de Waals force between graphite layers and the strong repulsive force between the terminated surfaces involving in hydroxyl group of graphite oxide and hydrogen atom of DLC films. Our observations show that thermally induced structural changes of DLC films are capable of achieving superlow friction in ambient air. These findings are beneficial to design the terminated surface of the DLC film friction system with superlow friction behavior in engineering application.

Funding statement: This work was supported by the National Natural Science Foundation of China; [51305331]; [51675409].

Acknowledgements

We acknowledge the financial supports from theNational Natural Science Foundation of China (51305331 and 51675409) and China Scholarship Council.

References

[1] A. Erdemir and C. Donnet, J. Phys. D Appl. Phys., 39 (2006) R311–R327.10.1088/0022-3727/39/18/R01Search in Google Scholar

[2] Q. Zeng, O. Eryilmaz and A. Erdemir, Thin Solid Films, 519 (2011) 3203–3212.10.1016/j.tsf.2011.01.102Search in Google Scholar

[3] D.R. Tallant, J.E. Parmeter, M.P. Siegal and R.L. Simpson, Diam. Relat. Mater., 4 (1995) 191–199.10.1016/0925-9635(94)00243-6Search in Google Scholar

[4] P. Mosaner, M. Bonelli and A. Miotello, Appl. Surf. Sci., 208-209 (2003) 561–565.10.1016/S0169-4332(02)01383-1Search in Google Scholar

[5] W. Zhang, A. Tanaka, K. Wazumi, Y. Koga and B. Xu, Diam. Relat. Mater., 13 (2004) 2166–2169.10.1016/j.diamond.2004.06.041Search in Google Scholar

[6] L. Huang, J. Yuan, C. Li, Z. Wang, T. Zhou and Z. Yin, Vacuum, 127 (2016) 96–102.10.1016/j.vacuum.2016.02.017Search in Google Scholar

[7] Q. Zeng, J. Zhang, J. Hong and C. Liu, Industr. Lubr. Tribol, 68 (2016) 325–335.10.1108/ILT-05-2015-0066Search in Google Scholar

[8] Q. Zeng, O. Eryilmaz and A. Erdemir, RSC Adv., 5 (2015) 93147–93154.10.1039/C5RA16084GSearch in Google Scholar

[9] D.C. Elias, R.R. Nair, T.M.G. Mohiuddin, S.V. Morozov, P. Blake, M.P. Halsall, A.C. Ferrari, D.W. Boukhvalov, M.I. Katsnelson, A.K. Geim and K.S. Novoselov, Science, 323 (2009) 610–613.10.1126/science.1167130Search in Google Scholar PubMed

[10] A. Pasquarello and R. Car, Phys. Rev. Lett., 80 (1998) 5145.10.1103/PhysRevLett.80.5145Search in Google Scholar

[11] L.G. Cancado, K. Takai, T. Enoki, M. Endo, Y.A. Kim, H. Mizusaki, N.L. Speziali, A. Jorioc and M.A. Pimenta, Carbon, 46 (2008) 272–275.10.1016/j.carbon.2007.11.015Search in Google Scholar

[12] J. Narayan and A. Bhaumik, J. Appl. Phys., 118 (2015) 215303.10.1063/1.4936595Search in Google Scholar

[13] J. Schwan, S. Ulrich, V. Batori, H. Ehrhard and S. Silva, J. Appl. Phys., 80 (1996) 440–447.10.1063/1.362745Search in Google Scholar

[14] O. Jantschner, S. Fieldb, D. Holeca, A. Fianc, D. Musicd, J. Schneiderd, K. Zorne and C. Mitterer, Acta. Mater., 82 (2015) 437446.10.1016/j.actamat.2014.09.030Search in Google Scholar

[15] S. Bhowmick, A. Banerji, M.J. Lukitsch and A.T. Alpas, Wear, 330 (2015) 261–271.10.1016/j.wear.2015.01.072Search in Google Scholar

[16] H. Pan, J.B. Yi, L. Shen, R.Q. Wu, J.H. Yang, J.Y. Lin, Y.P. Feng, J. Ding, L.H. Van and J.H. Yin, Phys. Rev. Lett., 99 (2007) 127201.10.1103/PhysRevLett.99.127201Search in Google Scholar PubMed

[17] Z. Fu, C. Wang, W. Zhang, W. Wang, W. Yue, X. Yu, Z. Peng, S. Lin and M. Dai, Mater. Des., 51 (2013) 775–779.10.1016/j.matdes.2013.04.098Search in Google Scholar

[18] D. Yang, A. Velamakanni, G. Bozoklu, S. Park, M. Stoller, R.D. Piner, S. Stankovich, I. Jung, D.A. Field, C.A. Ventric Jr and R.S. Ruoff, Carbon, 47 (2009) 145–152.10.1016/j.carbon.2008.09.045Search in Google Scholar

[19] H.S. Casalongue, S. Kaya, V. Viswanathan, D.J. Miller, D. Friebel, H. Hansen, J.K. Nørskov, A. Nilsson and H. Ogasawara, Nature Commun., 4 (2013) 1.10.1038/ncomms3817Search in Google Scholar

[20] A. Bagri, C. Mattevi, M. Acik, Y.J. Chabal, M. Chhowalla and V.B. Shenoy, Nature Chem., 2 (2010) 581–587.10.1038/nchem.686Search in Google Scholar PubMed

[21] J. Serra, P. Gonzalez, S. Liste, C. Serra, S. Chiussi, B. Leon, M. Perez-Amor, H.O. Ylanen and M. Hupa, J. Non-Cryst. Solids, 332 (2003) 20–27.10.1016/j.jnoncrysol.2003.09.013Search in Google Scholar

[22] D.Y. Wang, C.L. Chang and W.Y. Ho, Surf. Coat. Tech., 120 (1999) 138–144.10.1016/S0257-8972(99)00350-3Search in Google Scholar

Received: 2017-03-30
Accepted: 2017-09-18
Published Online: 2018-06-19
Published in Print: 2018-08-28

© 2018 Walter de Gruyter GmbH, Berlin/Boston

This article is distributed under the terms of the Creative Commons Attribution Non-Commercial License, which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Downloaded on 12.12.2025 from https://www.degruyterbrill.com/document/doi/10.1515/htmp-2017-0045/html
Scroll to top button