- Email: [email protected]
Influence of surface roughness and coating on the friction properties of nanometer-thick liquid lubricant films
Wear 319 (2014) 56–61
Contents lists available at ScienceDirect
Wear journal homepage: www.elsevier.com/locate/wear
Influence of surface roughness and coating on the friction properties of nanometer-thick liquid lubricant films Renguo Lu a, Hedong Zhang a,n, Yasunaga Mitsuya b, Kenji Fukuzawa c, Shintaro Itoh c a
Graduate School of Information Science, Nagoya University, Nagoya 464-8601, Japan Nagoya Industrial Science Research Institute, Nagoya 464-0819, Japan c Graduate School of Engineering, Nagoya University, Nagoya 464-8603, Japan b
art ic l e i nf o
a b s t r a c t
Article history: Received 22 April 2014 Received in revised form 10 July 2014 Accepted 14 July 2014 Available online 19 July 2014
The friction properties of nanometer-thick lubricant films are crucial in the reliability and durability of miniaturized moving mechanical components in micro- and nanoelectromechanical systems and hard disk drives. In this work, gas cluster ion beam treatments were applied to prepare smoothed glass sliding pins and diamond-like carbon (DLC) coated sliding pins for pin-on-disk friction tests to study the effect of the surface roughness and the DLC coating on the friction properties of perfluoropolyether (PFPE) films. The effect of the texture of the solid surface on the friction properties was also investigated. The friction properties were not exclusively determined by the surface energies of the solids. The friction coefficients of the nanometer-thick PFPE films, confined between solid surfaces were sensitive to the surface roughness of the solids. The friction coefficient generally increased with an increasing surface roughness (the composite standard deviation of the surface roughnesses of the pair of solid surfaces). This tendency was reversed when there was a strong longitudinal roughness effect at the contact interface. Moreover, the DLC-coated sliding pins had lower friction coefficients than the glass sliding pins, indicating that the friction properties were noticeably affected by the hardness and Young's modulus of the contact materials. The mechanisms for the effects of the surface roughness, surface texture and contact materials on the friction properties are discussed. & 2014 Elsevier B.V. All rights reserved.
Keywords: Nanometer-thick lubricant film Friction Surface roughness Diamond-like carbon coating Nanoindentation
1. Introduction Friction on the micro/nanometer scale directly affects the performance and reliability of miniaturized moving mechanical components, such as in micro- and nanoelectromechanical systems (MEMS/NEMS) and hard disk drives (HDD) [1,2]. To improve the tribological performance of the sliding components, the surfaces are generally separated by an ultrathin liquid lubricant film. In HDD, 1–2 nm thick perfluoropolyether (PFPE) films are employed as a lubricant to reduce friction and wear [3,4]. As the HDD industry is striving to achieve ultrahigh recording densities, the flying height of a head over a disk (currently less than 4 nm) should be decreased [5,6]. However, doing this will increase the probability of head-disk conflicts and may even cause sustained contact between the head and the disk. Consequently, a better understanding on the friction of nanometer-thick liquid lubricant films confined between two solid surfaces is crucial. The friction is n
Corresponding author. Tel./fax: þ 81 52 789 4803. E-mail address: [email protected] (H. Zhang).
http://dx.doi.org/10.1016/j.wear.2014.07.010 0043-1648/& 2014 Elsevier B.V. All rights reserved.
generally dependent on many factors, such as surface roughness, surface texturing, solid material and lubricant used. Based on the optimization of the interplay of the surface roughness and the lubricant film thickness, the friction coefficient and wear decreased with an increase in the ratio between the film thickness and the surface roughness [7,8]. In addition, the nanotextured surfaces, subject to elastic deformation, reduced the friction coefficient in the presence of liquid lubricants with a linear molecular structure, but not in the presence of branched lubricants [9]. However, the mechanism for how the surface roughness, surface texturing and mechanical properties of the contact materials affect the friction performance on the nanometer scale has not been fully elucidated. Gas cluster ion beam (GCIB) processing is a powerful technique for smoothing surfaces at the atomic level, and it can also assist in the deposition of diamond-like carbon (DLC) with high hardness [10–12]. In this work, GCIB smoothing and GCIB assisted DLC coating were introduced to develop sliding pins with different surface roughnesses and different surface materials. The friction properties of the nanometer-thick PFPE films coated on textured
R. Lu et al. / Wear 319 (2014) 56–61
5 nm
0.0
5.0 μm
0
5 nm
0.0
5.0 μm
0
Fig. 1. Surface morphology of the textured disk (a) and the nontextured disk (b) measured in a 5.0 μm 5.0 μm area.
and nontextured solid surfaces were studied using these sliding pins.
2. Experimental details Textured and nontextured DLC films deposited on 2.5-inchdiameter magnetic disks were used as substrates to discuss the effects of surface texturing on the friction. In addition to the DLC films, the magnetic disks were comprised of a glass substrate and a magnetic layer. The DLC films were about 3.5 nm thick. Fig. 1 shows the surface morphology of the textured and nontextured disks, measured using the tapping mode of an atomic force microscope (AFM; Bruker AXS: Dimension Icon) with a 5.0 μm 5.0 μm area. From Fig. 1(a), many peaks along the circumferential direction of the textured disk were confirmed, which was parallel to the sliding direction in the friction measurements. Thus, this is referred to as the longitudinal roughness. The arithmetic mean roughness (Ra) of the textured disk was 0.36 nm, the root-mean-square roughness (Rq) was 0.46 nm, the maximum peak height (Rp) was 1.99 nm, and the maximum roughness depth (Rmax) was 4.24 nm. For the surface roughness of the nontextured disk, Ra was 0.22 nm, Rq was 0.28 nm, Rp was 1.27 nm, and Rmax was 2.96 nm. PFPE Zdol4000 was used to lubricate the substrates. Its chemical structure is HO–CH2–CF2–O–(C2F4O)p–(CF2O)q–CF2–CH2–OH with p/q¼ 1, and its nominal molecular weight was 4000 g/mol. The lubricant was diluted in an HFE-7200 solvent to a concentration of 0.15 wt% and then applied to the disks using a standard
57
dip-coating method. The lubricant films were adjusted to roughly the same thicknesses: 0.99 nm on the nontextured disks and 1.06 nm on the textured disks. The film thicknesses were measured with a scanning ellipsometer. Using an optical surface analyzer, it was confirmed that the films were distributed uniformly and displayed no evidence of de-wetting of the lubricant. All samples were stored for more than 48 h in a desiccator with 20% humidity to allow the lubricant films to relax prior to the friction measurements. The sliding pins used were optical glass balls (material: LaSFN9) with a diameter of 2 mm. The pins were processed with a GCIB treatment. One set of pins (the sliding pins, Ar1 and Ar8) were treated with different argon-GCIB (Ar-GCIB) doses (Ar1: 1 1016 ions/cm2 and Ar8: 8 1016 ions/cm2). The other set of pins (the sliding pins, DLC_A and DLC_B) were first treated with Ar-GCIB (Ar dose: 1 1016 ions/cm2). Analogous to the magnetic heads whose top surface was coated with DLC, the pins then had 300 nm thick DLC films deposited on them, with the assistance of Ar-GCIB irradiation. The DLC_B sliding pin was further treated with nitrogen-GCIB (N2 dose: 1 1016 ions/cm2). The accelerating voltage for the Ar and N2 cluster ions was 10 kV, the current density was 1 μA/cm2, and the gas pressure was 0.60 MPa. As a reference, an untreated sliding pin was also used. The surface roughness was measured with the tapping mode of an AFM over an area of 5.0 μm 5.0 μm on sliding pins' apexes. The results for all of the sliding pins used in this work are summarized in Table 1. The sliding pins exhibited nearly isotropic surface roughnesses, which were confirmed with the AFM measurements. The sliding pins were adhered to the end of cantilever suspensions and then mounted on a self-developed pin-on-disk tribometer, which features a highly sensitive system for measuring friction under light load conditions [7]. Friction force was determined from the horizontal displacement of a friction transducer consisting of two pairs of parallel leaves. The displacement was measured by a capacitance-type sensor with a resolution of 2.5 nm, giving rise to a resolution of 10 μN for friction force measurements. The sliding pin together with the friction transducer was driven up/down by a high-precision Z piezoelectric stage with a built-in displacement magnification mechanism. The load resolution was 3.5 μN, as determined by the displacement resolution of the Z stage and the spring constant of the suspension. To avoid the influence of solid–solid contact on the measured friction force, the external load should be small enough not to easily break the nanometer-thick lubricant films. Therefore, we used light external loads of 0.4–1.6 mN with an increment of 0.2 mN. We also must prevent vibration of the sliding pin and the friction transducer to ensure reliable and stable friction measurements. In this study, friction measurements were carried out at the 20-mm disk radius. The disk rotational speed was 30 rpm, corresponding to a sliding speed of 0.063 m/s. By monitoring the vertical vibration of the sliding pin and the horizontal vibration of the friction transducer with two laser Doppler vibrometers, we confirmed that stable contact sliding between the sliding pin and the samples was achieved during the friction measurements under the above conditions. At each external load, friction forces were measured for 15 cycles. To eliminate the variation during the running-in Table 1 Surface roughnesses of the sliding pins. No.
Ra (nm)
Rq (nm)
Rp (nm)
Rmax (nm)
Untreated Ar1 Ar8 DLC_A DLC_B
1.72 0.86 0.18 0.71 0.46
2.22 1.10 0.22 0.90 0.57
11.8 4.10 0.79 5.56 2.46
20.0 8.36 1.61 8.81 4.65
58
R. Lu et al. / Wear 319 (2014) 56–61
period, the data for the initial 10 cycles were discarded. The friction forces obtained in the last five cycles were averaged. The experiments were repeated three times. All of the measurements were taken at 23 72 1C with a relative humidity of 20 75% in a clean room environment. The surface energy of the sliding pin determines the wettability of liquid lubricant films and may affect the friction properties. Hence, the surface energies were ascertained with a contact angle meter (Kyowa Interface Science Co., LTD. Japan: DropMaster 500). It was difficult to directly measure the contact angles of the sliding pins because of their small size and curved surface, as such, glass plates were used instead. However, the material in the commercially available glass plates was BK7, which was different from that of the sliding pins. An Ar-GCIB treatment effectively decreased the surface roughness without changing the material, and a roughness on the nanometer scale did not significantly change the surface energy for a given material [13,14]. Therefore, considering that the surface energy of an Ar-GCIB treated BK7 glass plate should differ from that of an Ar-GCIB treated LaSFN9 sliding pin, the BK7 glass plates were treated under the same conditions as the DLC_A and DLC_B sliding pins to verify the effect of the DLC coating. Since the DLC coating was 300 nm thick, it should be sufficient to screen out the effect of the underlying substrates, and its surface properties should be independent of the underlying substrates. Water and methyl iodide were used as reference liquids to measure the contact angles. The volume of the liquid droplets was controlled at 1.5 μL. The droplets were brought into contact with the plates and the contact angles were measured within 5 s. A total of five independent measurements were conducted randomly on each surface, and an average value was recorded for each sample. The hardness and Young's modulus of the solid surfaces should affect the friction properties of the intervening liquid lubricants because of their influence on the contact area. Nanoindentation measurements were performed on the sliding pins and the DLCcoated disks using the AFM with a PDNISP tip (a diamond tip mounted on a stainless steel cantilever), to measure their hardness and Young's modulus. The tip had a resonance frequency of 73.6 kHz and a stiff spring constant of 236 N/m. The x-rotation value of this tip, according to the manufacturer was 121. The load for the nanoindentation measurements was set to be 70.8 μN. The AFM images were obtained immediately after the nanoindentations to avoid any time-dependent elastic recovery of the residual impression. The projected contact areas were automatically measured using the software built into the AFM. For each sample, a minimum of five indentations were randomly made within an area of 500 nm 500 nm and the averaged values are presented.
Table 2 Contact angles and surface energies of the glass plates. No.
DLC_A glass plate DLC_B glass plate Untreated glass plate
Contact angle (Deg)
Surface energy (mN/m)
Water
Methyl iodide
Polar
Dispersive
Total
85.5 75.7 51.5
27.5 29.6 50.5
1.3 4.9 25.9
44.0 40.2 23.6
45.3 45.1 49.5
Fig. 2. AFM images of the indentation impressions made on the glass sliding pin (a), the DLC-coated (DLC_A) glass sliding pin (b), the textured disk (c), and the nontextured disk (d).
The untreated glass plates had a water contact angle of 51.51 and a methyl iodide contact angle of 50.51. After the plate was coated with DLC, the water contact angle increased while the methyl iodide contact angle decreased, reflecting a decrease in polar surface energy and an increase in dispersive surface energy. As a result, the total surface energy slightly decreased. Comparing the DLC_A and DLC_B glass plates, their surface energies were almost the same, although the DLC_B glass plate had been further treated with N2-GCIB and its surface roughness was lower. This agrees well with the conclusion that the surface energy is independent of the nanoscale surface roughness [13,14].
3. Experimental results 3.2. Nanoindentation 3.1. Surface energy Young's equation for the contact angle (θ) of a liquid droplet on a surface is cos θ ¼ 1 þ
2
γ dl þ γ pl
½ðγ ds γ dl Þ1=2 þ ðγ ps γ pl Þ1=2 ;
ð1Þ
where γdl and γpl are the dispersive and polar surface energy components of the reference liquid, respectively, γds and γps are the dispersive and polar surface energy components of the sample, respectively. The total surface energy is the sum of the dispersive and polar surface energies. By measuring the contact angles with water and methyl iodide, the surface energies of the DLC-coated glass plates were calculated using Eq. (1) and are presented in Table 2. The results for an untreated BK7 glass plate are also listed in Table 2 as a reference.
Fig. 2 shows the representative AFM images of the indentation impressions made on various surfaces. The DLC-coated (DLC_A) sliding pin showed the smallest projected area for the pit that was left by the indent, at the same load as the others. By measuring the projected area (Ap), the hardness (H) can be calculated by the following equation [15]: H ¼ Nmax =Ap ;
ð2Þ
where Nmax is the maximum load exerted by the tip. Fig. 3 shows typical load–displacement curves for the loading and unloading steps of the indentation process. The slope of the unloading curve provides a measure of the elastic modulus. Because the displacement was elastic during unloading, the relationship between the unloading curve and the elastic modulus can be described using elastic contact theory [16]. The reduced
R. Lu et al. / Wear 319 (2014) 56–61
0.8
80 Loading Unloading
70
0.7 Friction froce, F mN
Load, μN
60 50 40 30 20 10 0
0.6 0.5 0.4 0.3 Sliding pin: Ar8 Disk: non-textured Lubricant: Zdol4000 Sliding speed: 0.063 m/s
0.2 0.1
0
5 10 15 Penetration depth, nm
0
20
Fig. 3. Nanoindentation load–displacement curves for the nontextured disk during loading (red open circles) and unloading (blue closed circles) (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article).
30 20
700 600 500 400 300 200
10 0
Friction coefficient
40
2
Untreated Ar1 Ar8 DLC_A DLC_B Glass ball 0.4 DLC-coated glass ball Glass ball DLCcoated 0.3 glass ball
0.2 0.1
100
0
0
Fig. 4. Indentation hardness and Young's modulus of the sliding pins and disks. Error bars indicate the standard deviation of duplicate measurements randomly made within an area of 500 nm 500 nm.
modulus (E*), which accounts for the elastic displacement in both the specimen and the indenter, is evaluated using pffiffiffiffi π S pffiffiffiffiffiffi; En ¼ ð3Þ 2 Ap where S is the stiffness of the contact, determined by fitting the upper region of the unloading curve with a power law function and then measuring the slope of the fitting curve at the peak load. The reduced modulus is related to Young's modulus (Es) of the test specimen through the following relationship: 1 1 ν2i 1 ν2s ¼ þ : Ei Es En
0.5 1 1.5 External load, L mN
0.5
Young's modulus, GPa
50
Glass sliding pin DLC-coated sliding pin Textured disk Non-textured disk
0
Fig. 5. Friction force as a function of the external load.
800
60 Indentation hardness, GPa
59
ð4Þ
Here, Young's modulus of the diamond tip (Ei) was 1140 GPa, Poisson's ratio of the diamond tip (νi) was 0.07 and Poisson's ratio of the sample (νs) was 0.3. The measured nanoindentation hardness and Young's modulus are depicted in Fig. 4. The glass sliding pin had a lower hardness and Young's modulus than both the textured and nontextured disks. After the pins were coated with DLC, the hardness and Young's modulus of the sliding pin increased dramatically, becoming much higher than the disks whose top surfaces were also DLC. This difference was possibly induced by the thickness of the DLC coating: which was 300 nm on the glass balls, but only 3.5 nm on the surface of the disks. The cross-sectional profiles of the nanoindentation areas on the disks show that the indentation
Non-textured disk
Textured disk
Fig. 6. Friction coefficients measured with different sliding pins and samples. Error bars indicate the maximum deviation from the average of three duplicate measurements.
depth exceeded 3.5 nm, indicating that the thin DLC films on the disks were unable to support the indentation load. Hence, the measured hardness and Young's modulus of the disks were most likely affected by the underlying magnetic layers. 3.3. Friction properties Fig. 5 shows the dependence of the friction force on the external load when using an Ar8 sliding pin on a nontextured disk coated with a Zdol4000 film. The solid line in Fig. 5 is the line that was fitted to the data. The friction force increased linearly with an increasing external load. A linear relation was obtained for all of the sliding pins when they were mounted on either the textured disk or the nontextured disk. Therefore, the friction force (F) can be expressed as [17] F ¼ F 0 þ μL;
ð5Þ
where F0 is the friction force when the external load is zero, L is the external load and μ is the friction coefficient. Here the friction coefficient reflects the variations in the friction force with an external load, which is defined as
μ ¼ dF=dL:
ð6Þ
Fig. 6 summarizes the friction coefficients of the Zdol4000 films coated on the textured and nontextured disks that were measured using the glass sliding pins and DLC-coated sliding pins. The surface roughness decreased in the following orders: untreated,
60
R. Lu et al. / Wear 319 (2014) 56–61
Ar1, and Ar8 for the sliding pins that underwent an Ar-GCIB treatment; and untreated, DLC_A, and DLC_B sliding pins after they were coated with a DLC film and underwent a N2-GCIB treatment. Three immediate conclusions follow from the data in Fig. 6. First, the friction coefficient decreased with a decreasing sliding pin surface roughness, except for the textured disk and glass sliding pin combination. Second, the DLC-coated sliding pins had smaller friction coefficients than the glass sliding pins. Third, the friction coefficient of the lubricant film coated on the nontextured disk was higher than that on the textured disk for a given sliding pin.
4. Discussion The surface energy is widely recognized as an important factor that affects the friction behavior. In general, a low surface energy induces a small amount of friction [18]. As seen in Fig. 6, the DLC_A and DLC_B sliding pins gave rise to different friction coefficients, particularly when they were mounted on the nontextured disk. However, the surface energies of the two DLC-coated sliding pins should be almost the same, which was confirmed from the contact angle measurements on the DLC-coated glass plates. Thus, it can be inferred that the friction properties of the liquid lubricant films confined between solid surfaces were not exclusively determined by the surface energies of the solids. The surface roughness is crucial for the friction of nanometerthick liquid lubricant films between solid surfaces. Fig. 7 shows the variations in the friction coefficients with the surface roughness (σ). Here, σ is the composite standard deviation for the surface roughness of the pair of surfaces. σ is defined as qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi σ ¼ R2q1 þ R2q2 ; ð7Þ where Rq1 is the surface roughness of the sliding pin and Rq2 is the surface roughness of the disk. The solid lines in Fig. 7 were inserted as a guide for the eyes. The friction coefficient increased with an increasing surface roughness, except for the textured disk and glass sliding pin combination. However, the friction coefficient generally decreased with an increasing surface roughness for dry friction [19]. The opposite change in the friction coefficient with surface roughness indicated that the interplay between the lubricant film and the surface roughness should be taken into consideration. Let us consider an interface with isotropically distributed contacts, such as that formed between the glass or DLC-coated
0.4 Glass pin
Friction coefficient
0.35
Non-textured disk DLC pin Glass pin
0.3 0.25 0.2
DLC pin
Textured disk (longitudinal)
0.15 0.1
Glass sliding pin on non-textured disk DLCsliding pin on non-textured disk Glass sliding pin on textured disk DLC sliding pin on textured disk
0.05 0
0
0.5 1 1.5 2 2.5 Surface roughness σ, nm
3
Fig. 7. Friction coefficients as a function of the surface roughness. Error bars indicate the maximum deviation from the average of three duplicate measurements.
sliding pins and the nontextured disk. As the surface roughness decreased, the mean contact pressure between a pair of contacting asperities decreased [19], allowing more lubricant molecules to intervene between the contacting asperities. These intervening lubricant molecules resulted in decrease of the friction coefficient. When a lubricant film was coated on the textured disk that had a longitudinal roughness, the anisotropic effect of the disk surface became more noticeable at the contact interface with the glass sliding pins, as the pin surface roughness decreased. Thus, it can be suggested that the longitudinal anisotropy of the composite surface roughness gave rise to a high friction coefficient [7]. However, once the glass sliding pins were coated with a DLC film, the hardness and Young's modulus became much higher than that of the textured disk, as seen from Fig. 6. During the sliding contact measurements, the textured disk was more likely to become deformed than the DLC-coated sliding pins, suppressing the effect of the longitudinal roughness on the friction. This helps to explain the experimental results where, similar to the sliding pins on the nontextured disk, the friction coefficient of the DLC-coated sliding pins on the textured disk increased with an increasing surface roughness. There are two conceivable reasons for why the DLC-coated sliding pins had smaller friction coefficients than the glass sliding pins. One is that surface graphitization or surfaces with graphite inclusions on the DLC coatings contribute to the low friction performance [20]. The other is the change in the contact status, caused by the DLC coatings. When coated with a DLC film, the sliding pins became much harder than the disks, as described above. Moreover, the high Young's modulus of the DLC coating led to less elastic deformation of the sliding pins during contact with the disk. Therefore, in comparison with the glass sliding pins, the contact area between the DLC-coated sliding pins and the disk was less sensitive to changes in the external load, resulting in low friction coefficients. A similar explanation can be applied to the different friction coefficients for a given sliding pin on the nontextured disk and on the textured disk. Namely, as compared with the textured disks, the smaller hardness of the nontextured disk resulted in larger deformation of the disk during sliding contact, leading to a larger contact area and thereby a higher friction coefficient.
5. Conclusions By using glass and DLC-coated glass sliding pins with different surface roughnesses, the friction properties of textured and nontextured disks coated with nanometer-thick PFPE films were investigated. The friction properties of nanometer-thick liquid lubricant films, confined between two solid surfaces, were not exclusively determined by the surface energies of the solids and were sensitive to the surface roughness of the solid surfaces. Regardless of the materials in contact, the friction coefficient generally tended to increase with an increasing surface roughness (the composite standard deviation for the surface roughness of the pair of contacting surfaces). However, the friction coefficient decreased with an increasing surface roughness for the textured disk and glass sliding pin combination. This result suggested that the strong longitudinal roughness effect at the contact interface gave rise to the high friction coefficient. Moreover, the DLC-coated sliding pins had lower friction coefficients than the glass sliding pins, indicating that the hardness and Young's modulus of the contact materials affected the area that was contacted, and consequently, the friction properties of the confined lubricant films.
R. Lu et al. / Wear 319 (2014) 56–61
Acknowledgments This work was supported in part by the Funding Program for Next Generation World-Leading Researchers (No. GR053), Grantin-Aid for Scientific Research (C No. 22560153), the Industrial Technology Research Grant Program in 2006 from NEDO (No. 06A12003A), and a grant from the Storage Research Consortium. The authors would like to thank Dr Teruyuki Kitagawa from Nomura Plating Co., Ltd., for his help in the GCIB treatment. References [1] O.M. Braun, A.G. Naumovets, Nanotribology: microscopic mechanisms of friction, Surf. Sci. Rep. 60 (2006) 79–158. [2] L. Li, P.M. Jones, Y. Hsia, Characterization of a nanometer-thick sputtered polytetrafluoroethylene film, Appl. Surf. Sci. 257 (2011) 4478–4485. [3] H. Zhang, T. Takimoto, K. Fukuzawa, S. Itoh, Effect of ultraviolet irradiation on adhesion of nanometer-thick lubricant films coated on magnetic disk surface, IEEE Trans. Magn. 47 (2011) 94–99. [4] B. Zhang, H. Chiba, A. Nakajima, Blow-off flow of nano PFPE liquid film at hard disk surfaces, Tribol. Lett. 39 (2010) 193–199. [5] N. Li, Y. Meng, D.B. Bogy, Effect of PFPE lubricant properties on the critical clearance and rate of the lubricant transfer from disk surface to slider, Tribol. Lett. 43 (2011) 275–286. [6] U. Boettcher, H. Li, R.A. de Callafon, F.E. Talke, Dynamic flying height adjustment in hard disk drives through feed forward control, IEEE Trans. Magn. 47 (2011) 1823–1829. [7] R. Lu, H. Zhang, Y. Mitsuya, K. Fukuzawa, S. Itoh, Friction measurements of nanometer-thick lubricant films using ultra-smooth sliding pins treated with gas cluster ion beam, Appl. Surf. Sci. 280 (2013) 619–625. [8] H. Lee, B. Bhushan, Role of surface roughness and lubricant film thickness in nanolubrication of sliding components in adaptive optics, J. Colloid Interface Sci. 353 (2011) 574–581.
61
[9] A.A. Al-Azizi, O. Eryilmaz, A. Erdemir, S.H. Kim, Effect of nanoscale surface texture and lubricant molecular structure on boundary lubrication in liquid, Langmuir 29 (2013) 13419–13426. [10] I. Yamada, J. Matsuo, N. Toyoda, A. Kirkpatrick, Material processing by gas cluster ion beams, Mater. Sci. Eng. Rep. 34 (2001) 231–295. [11] T. Kitagawa, I. Yamada, N. Toyoda, H. Tsubakino, J. Matsuo, G. Takaoka, A. Kirkpatrick, Hard DLC film formation by gas cluster ion beam assisted deposition, Nucl. Instrum. Methods Phys. Res. B 201 (2003) 405–412. [12] A. Gandopadhyay, K. Sinha, D. Uy, D.G. Mcwatt, R.J. Zdrodowski, S.J. Simko, Friction, wear, and surface film formation characteristics of diamond-like carbon thin coating in valvetrain application, Tribol. Trans. 54 (2011) 104–114. [13] E.P. Ivanova, V.K. Truong, J.Y. Wang, C.C. Berndt, R.T. Jones, I.I. Yusuf, I. Peake, H. W. Schmidt, C. Fluke, D. Barnes, R.J. Crawford, Impact of nanoscale roughness of titanium thin film surfaces on bacterial retention, Langmuir 26 (2010) 1973–1982. [14] P.M. Hansson, L. Skedung, P.M. Claesson, A. Swerin, J. Schoelkopf, P.A.C. Gane, M.W. Rutland, E. Thormann, Robust hydrophobic surface displaying different surface roughness scales while maintaining the same wettability, Langmuir 27 (2011) 8153–8159. [15] X. Li, H. Gao, C.J. Murphy, K.K. Caswell, Nanoindentaion of silver nanowires, Nano Lett. 3 (2003) 1495–1498. [16] W.C. Oliver, G.M. Pharr, Measurement of hardness and elastic modulus by instrumented indentation: advances in understanding and refinements to methodology, J. Mater. Res. 19 (2004) 3–20. [17] J. Gao, W.D. Luedtke, D. Gourdon, M. Ruths, J.N. Israelachvili, U. Landman, Frictional forces and Amontons's law: from the molecular to the macroscopic scale, J. Phys. Chem. B 108 (2004) 3410–3425. [18] Y. Kim, F. Limanto, D.H. Lee, R.S. Fearing, R. Maboudian, Role of countersubstrate surface energy in macroscale friction of nanofiber arrays, Langmuir 28 (2012) 2922–2927. [19] J.A. Ogilvy, Numerical simulation of friction between contacting rough surfaces, J. Phys. D: Appl. Phys. 24 (1991) 2098–2109. [20] B. Vengudusamy, R.A. Mufti, G.D. Lamb, J.H. Green, H.A. Spikes, Friction properties of DLC/DLC contacts in base oil, Tribol. Int. 44 (2011) 922–932.
Contents lists available at ScienceDirect
Wear journal homepage: www.elsevier.com/locate/wear
Influence of surface roughness and coating on the friction properties of nanometer-thick liquid lubricant films Renguo Lu a, Hedong Zhang a,n, Yasunaga Mitsuya b, Kenji Fukuzawa c, Shintaro Itoh c a
Graduate School of Information Science, Nagoya University, Nagoya 464-8601, Japan Nagoya Industrial Science Research Institute, Nagoya 464-0819, Japan c Graduate School of Engineering, Nagoya University, Nagoya 464-8603, Japan b
art ic l e i nf o
a b s t r a c t
Article history: Received 22 April 2014 Received in revised form 10 July 2014 Accepted 14 July 2014 Available online 19 July 2014
The friction properties of nanometer-thick lubricant films are crucial in the reliability and durability of miniaturized moving mechanical components in micro- and nanoelectromechanical systems and hard disk drives. In this work, gas cluster ion beam treatments were applied to prepare smoothed glass sliding pins and diamond-like carbon (DLC) coated sliding pins for pin-on-disk friction tests to study the effect of the surface roughness and the DLC coating on the friction properties of perfluoropolyether (PFPE) films. The effect of the texture of the solid surface on the friction properties was also investigated. The friction properties were not exclusively determined by the surface energies of the solids. The friction coefficients of the nanometer-thick PFPE films, confined between solid surfaces were sensitive to the surface roughness of the solids. The friction coefficient generally increased with an increasing surface roughness (the composite standard deviation of the surface roughnesses of the pair of solid surfaces). This tendency was reversed when there was a strong longitudinal roughness effect at the contact interface. Moreover, the DLC-coated sliding pins had lower friction coefficients than the glass sliding pins, indicating that the friction properties were noticeably affected by the hardness and Young's modulus of the contact materials. The mechanisms for the effects of the surface roughness, surface texture and contact materials on the friction properties are discussed. & 2014 Elsevier B.V. All rights reserved.
Keywords: Nanometer-thick lubricant film Friction Surface roughness Diamond-like carbon coating Nanoindentation
1. Introduction Friction on the micro/nanometer scale directly affects the performance and reliability of miniaturized moving mechanical components, such as in micro- and nanoelectromechanical systems (MEMS/NEMS) and hard disk drives (HDD) [1,2]. To improve the tribological performance of the sliding components, the surfaces are generally separated by an ultrathin liquid lubricant film. In HDD, 1–2 nm thick perfluoropolyether (PFPE) films are employed as a lubricant to reduce friction and wear [3,4]. As the HDD industry is striving to achieve ultrahigh recording densities, the flying height of a head over a disk (currently less than 4 nm) should be decreased [5,6]. However, doing this will increase the probability of head-disk conflicts and may even cause sustained contact between the head and the disk. Consequently, a better understanding on the friction of nanometer-thick liquid lubricant films confined between two solid surfaces is crucial. The friction is n
Corresponding author. Tel./fax: þ 81 52 789 4803. E-mail address: [email protected] (H. Zhang).
http://dx.doi.org/10.1016/j.wear.2014.07.010 0043-1648/& 2014 Elsevier B.V. All rights reserved.
generally dependent on many factors, such as surface roughness, surface texturing, solid material and lubricant used. Based on the optimization of the interplay of the surface roughness and the lubricant film thickness, the friction coefficient and wear decreased with an increase in the ratio between the film thickness and the surface roughness [7,8]. In addition, the nanotextured surfaces, subject to elastic deformation, reduced the friction coefficient in the presence of liquid lubricants with a linear molecular structure, but not in the presence of branched lubricants [9]. However, the mechanism for how the surface roughness, surface texturing and mechanical properties of the contact materials affect the friction performance on the nanometer scale has not been fully elucidated. Gas cluster ion beam (GCIB) processing is a powerful technique for smoothing surfaces at the atomic level, and it can also assist in the deposition of diamond-like carbon (DLC) with high hardness [10–12]. In this work, GCIB smoothing and GCIB assisted DLC coating were introduced to develop sliding pins with different surface roughnesses and different surface materials. The friction properties of the nanometer-thick PFPE films coated on textured
R. Lu et al. / Wear 319 (2014) 56–61
5 nm
0.0
5.0 μm
0
5 nm
0.0
5.0 μm
0
Fig. 1. Surface morphology of the textured disk (a) and the nontextured disk (b) measured in a 5.0 μm 5.0 μm area.
and nontextured solid surfaces were studied using these sliding pins.
2. Experimental details Textured and nontextured DLC films deposited on 2.5-inchdiameter magnetic disks were used as substrates to discuss the effects of surface texturing on the friction. In addition to the DLC films, the magnetic disks were comprised of a glass substrate and a magnetic layer. The DLC films were about 3.5 nm thick. Fig. 1 shows the surface morphology of the textured and nontextured disks, measured using the tapping mode of an atomic force microscope (AFM; Bruker AXS: Dimension Icon) with a 5.0 μm 5.0 μm area. From Fig. 1(a), many peaks along the circumferential direction of the textured disk were confirmed, which was parallel to the sliding direction in the friction measurements. Thus, this is referred to as the longitudinal roughness. The arithmetic mean roughness (Ra) of the textured disk was 0.36 nm, the root-mean-square roughness (Rq) was 0.46 nm, the maximum peak height (Rp) was 1.99 nm, and the maximum roughness depth (Rmax) was 4.24 nm. For the surface roughness of the nontextured disk, Ra was 0.22 nm, Rq was 0.28 nm, Rp was 1.27 nm, and Rmax was 2.96 nm. PFPE Zdol4000 was used to lubricate the substrates. Its chemical structure is HO–CH2–CF2–O–(C2F4O)p–(CF2O)q–CF2–CH2–OH with p/q¼ 1, and its nominal molecular weight was 4000 g/mol. The lubricant was diluted in an HFE-7200 solvent to a concentration of 0.15 wt% and then applied to the disks using a standard
57
dip-coating method. The lubricant films were adjusted to roughly the same thicknesses: 0.99 nm on the nontextured disks and 1.06 nm on the textured disks. The film thicknesses were measured with a scanning ellipsometer. Using an optical surface analyzer, it was confirmed that the films were distributed uniformly and displayed no evidence of de-wetting of the lubricant. All samples were stored for more than 48 h in a desiccator with 20% humidity to allow the lubricant films to relax prior to the friction measurements. The sliding pins used were optical glass balls (material: LaSFN9) with a diameter of 2 mm. The pins were processed with a GCIB treatment. One set of pins (the sliding pins, Ar1 and Ar8) were treated with different argon-GCIB (Ar-GCIB) doses (Ar1: 1 1016 ions/cm2 and Ar8: 8 1016 ions/cm2). The other set of pins (the sliding pins, DLC_A and DLC_B) were first treated with Ar-GCIB (Ar dose: 1 1016 ions/cm2). Analogous to the magnetic heads whose top surface was coated with DLC, the pins then had 300 nm thick DLC films deposited on them, with the assistance of Ar-GCIB irradiation. The DLC_B sliding pin was further treated with nitrogen-GCIB (N2 dose: 1 1016 ions/cm2). The accelerating voltage for the Ar and N2 cluster ions was 10 kV, the current density was 1 μA/cm2, and the gas pressure was 0.60 MPa. As a reference, an untreated sliding pin was also used. The surface roughness was measured with the tapping mode of an AFM over an area of 5.0 μm 5.0 μm on sliding pins' apexes. The results for all of the sliding pins used in this work are summarized in Table 1. The sliding pins exhibited nearly isotropic surface roughnesses, which were confirmed with the AFM measurements. The sliding pins were adhered to the end of cantilever suspensions and then mounted on a self-developed pin-on-disk tribometer, which features a highly sensitive system for measuring friction under light load conditions [7]. Friction force was determined from the horizontal displacement of a friction transducer consisting of two pairs of parallel leaves. The displacement was measured by a capacitance-type sensor with a resolution of 2.5 nm, giving rise to a resolution of 10 μN for friction force measurements. The sliding pin together with the friction transducer was driven up/down by a high-precision Z piezoelectric stage with a built-in displacement magnification mechanism. The load resolution was 3.5 μN, as determined by the displacement resolution of the Z stage and the spring constant of the suspension. To avoid the influence of solid–solid contact on the measured friction force, the external load should be small enough not to easily break the nanometer-thick lubricant films. Therefore, we used light external loads of 0.4–1.6 mN with an increment of 0.2 mN. We also must prevent vibration of the sliding pin and the friction transducer to ensure reliable and stable friction measurements. In this study, friction measurements were carried out at the 20-mm disk radius. The disk rotational speed was 30 rpm, corresponding to a sliding speed of 0.063 m/s. By monitoring the vertical vibration of the sliding pin and the horizontal vibration of the friction transducer with two laser Doppler vibrometers, we confirmed that stable contact sliding between the sliding pin and the samples was achieved during the friction measurements under the above conditions. At each external load, friction forces were measured for 15 cycles. To eliminate the variation during the running-in Table 1 Surface roughnesses of the sliding pins. No.
Ra (nm)
Rq (nm)
Rp (nm)
Rmax (nm)
Untreated Ar1 Ar8 DLC_A DLC_B
1.72 0.86 0.18 0.71 0.46
2.22 1.10 0.22 0.90 0.57
11.8 4.10 0.79 5.56 2.46
20.0 8.36 1.61 8.81 4.65
58
R. Lu et al. / Wear 319 (2014) 56–61
period, the data for the initial 10 cycles were discarded. The friction forces obtained in the last five cycles were averaged. The experiments were repeated three times. All of the measurements were taken at 23 72 1C with a relative humidity of 20 75% in a clean room environment. The surface energy of the sliding pin determines the wettability of liquid lubricant films and may affect the friction properties. Hence, the surface energies were ascertained with a contact angle meter (Kyowa Interface Science Co., LTD. Japan: DropMaster 500). It was difficult to directly measure the contact angles of the sliding pins because of their small size and curved surface, as such, glass plates were used instead. However, the material in the commercially available glass plates was BK7, which was different from that of the sliding pins. An Ar-GCIB treatment effectively decreased the surface roughness without changing the material, and a roughness on the nanometer scale did not significantly change the surface energy for a given material [13,14]. Therefore, considering that the surface energy of an Ar-GCIB treated BK7 glass plate should differ from that of an Ar-GCIB treated LaSFN9 sliding pin, the BK7 glass plates were treated under the same conditions as the DLC_A and DLC_B sliding pins to verify the effect of the DLC coating. Since the DLC coating was 300 nm thick, it should be sufficient to screen out the effect of the underlying substrates, and its surface properties should be independent of the underlying substrates. Water and methyl iodide were used as reference liquids to measure the contact angles. The volume of the liquid droplets was controlled at 1.5 μL. The droplets were brought into contact with the plates and the contact angles were measured within 5 s. A total of five independent measurements were conducted randomly on each surface, and an average value was recorded for each sample. The hardness and Young's modulus of the solid surfaces should affect the friction properties of the intervening liquid lubricants because of their influence on the contact area. Nanoindentation measurements were performed on the sliding pins and the DLCcoated disks using the AFM with a PDNISP tip (a diamond tip mounted on a stainless steel cantilever), to measure their hardness and Young's modulus. The tip had a resonance frequency of 73.6 kHz and a stiff spring constant of 236 N/m. The x-rotation value of this tip, according to the manufacturer was 121. The load for the nanoindentation measurements was set to be 70.8 μN. The AFM images were obtained immediately after the nanoindentations to avoid any time-dependent elastic recovery of the residual impression. The projected contact areas were automatically measured using the software built into the AFM. For each sample, a minimum of five indentations were randomly made within an area of 500 nm 500 nm and the averaged values are presented.
Table 2 Contact angles and surface energies of the glass plates. No.
DLC_A glass plate DLC_B glass plate Untreated glass plate
Contact angle (Deg)
Surface energy (mN/m)
Water
Methyl iodide
Polar
Dispersive
Total
85.5 75.7 51.5
27.5 29.6 50.5
1.3 4.9 25.9
44.0 40.2 23.6
45.3 45.1 49.5
Fig. 2. AFM images of the indentation impressions made on the glass sliding pin (a), the DLC-coated (DLC_A) glass sliding pin (b), the textured disk (c), and the nontextured disk (d).
The untreated glass plates had a water contact angle of 51.51 and a methyl iodide contact angle of 50.51. After the plate was coated with DLC, the water contact angle increased while the methyl iodide contact angle decreased, reflecting a decrease in polar surface energy and an increase in dispersive surface energy. As a result, the total surface energy slightly decreased. Comparing the DLC_A and DLC_B glass plates, their surface energies were almost the same, although the DLC_B glass plate had been further treated with N2-GCIB and its surface roughness was lower. This agrees well with the conclusion that the surface energy is independent of the nanoscale surface roughness [13,14].
3. Experimental results 3.2. Nanoindentation 3.1. Surface energy Young's equation for the contact angle (θ) of a liquid droplet on a surface is cos θ ¼ 1 þ
2
γ dl þ γ pl
½ðγ ds γ dl Þ1=2 þ ðγ ps γ pl Þ1=2 ;
ð1Þ
where γdl and γpl are the dispersive and polar surface energy components of the reference liquid, respectively, γds and γps are the dispersive and polar surface energy components of the sample, respectively. The total surface energy is the sum of the dispersive and polar surface energies. By measuring the contact angles with water and methyl iodide, the surface energies of the DLC-coated glass plates were calculated using Eq. (1) and are presented in Table 2. The results for an untreated BK7 glass plate are also listed in Table 2 as a reference.
Fig. 2 shows the representative AFM images of the indentation impressions made on various surfaces. The DLC-coated (DLC_A) sliding pin showed the smallest projected area for the pit that was left by the indent, at the same load as the others. By measuring the projected area (Ap), the hardness (H) can be calculated by the following equation [15]: H ¼ Nmax =Ap ;
ð2Þ
where Nmax is the maximum load exerted by the tip. Fig. 3 shows typical load–displacement curves for the loading and unloading steps of the indentation process. The slope of the unloading curve provides a measure of the elastic modulus. Because the displacement was elastic during unloading, the relationship between the unloading curve and the elastic modulus can be described using elastic contact theory [16]. The reduced
R. Lu et al. / Wear 319 (2014) 56–61
0.8
80 Loading Unloading
70
0.7 Friction froce, F mN
Load, μN
60 50 40 30 20 10 0
0.6 0.5 0.4 0.3 Sliding pin: Ar8 Disk: non-textured Lubricant: Zdol4000 Sliding speed: 0.063 m/s
0.2 0.1
0
5 10 15 Penetration depth, nm
0
20
Fig. 3. Nanoindentation load–displacement curves for the nontextured disk during loading (red open circles) and unloading (blue closed circles) (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article).
30 20
700 600 500 400 300 200
10 0
Friction coefficient
40
2
Untreated Ar1 Ar8 DLC_A DLC_B Glass ball 0.4 DLC-coated glass ball Glass ball DLCcoated 0.3 glass ball
0.2 0.1
100
0
0
Fig. 4. Indentation hardness and Young's modulus of the sliding pins and disks. Error bars indicate the standard deviation of duplicate measurements randomly made within an area of 500 nm 500 nm.
modulus (E*), which accounts for the elastic displacement in both the specimen and the indenter, is evaluated using pffiffiffiffi π S pffiffiffiffiffiffi; En ¼ ð3Þ 2 Ap where S is the stiffness of the contact, determined by fitting the upper region of the unloading curve with a power law function and then measuring the slope of the fitting curve at the peak load. The reduced modulus is related to Young's modulus (Es) of the test specimen through the following relationship: 1 1 ν2i 1 ν2s ¼ þ : Ei Es En
0.5 1 1.5 External load, L mN
0.5
Young's modulus, GPa
50
Glass sliding pin DLC-coated sliding pin Textured disk Non-textured disk
0
Fig. 5. Friction force as a function of the external load.
800
60 Indentation hardness, GPa
59
ð4Þ
Here, Young's modulus of the diamond tip (Ei) was 1140 GPa, Poisson's ratio of the diamond tip (νi) was 0.07 and Poisson's ratio of the sample (νs) was 0.3. The measured nanoindentation hardness and Young's modulus are depicted in Fig. 4. The glass sliding pin had a lower hardness and Young's modulus than both the textured and nontextured disks. After the pins were coated with DLC, the hardness and Young's modulus of the sliding pin increased dramatically, becoming much higher than the disks whose top surfaces were also DLC. This difference was possibly induced by the thickness of the DLC coating: which was 300 nm on the glass balls, but only 3.5 nm on the surface of the disks. The cross-sectional profiles of the nanoindentation areas on the disks show that the indentation
Non-textured disk
Textured disk
Fig. 6. Friction coefficients measured with different sliding pins and samples. Error bars indicate the maximum deviation from the average of three duplicate measurements.
depth exceeded 3.5 nm, indicating that the thin DLC films on the disks were unable to support the indentation load. Hence, the measured hardness and Young's modulus of the disks were most likely affected by the underlying magnetic layers. 3.3. Friction properties Fig. 5 shows the dependence of the friction force on the external load when using an Ar8 sliding pin on a nontextured disk coated with a Zdol4000 film. The solid line in Fig. 5 is the line that was fitted to the data. The friction force increased linearly with an increasing external load. A linear relation was obtained for all of the sliding pins when they were mounted on either the textured disk or the nontextured disk. Therefore, the friction force (F) can be expressed as [17] F ¼ F 0 þ μL;
ð5Þ
where F0 is the friction force when the external load is zero, L is the external load and μ is the friction coefficient. Here the friction coefficient reflects the variations in the friction force with an external load, which is defined as
μ ¼ dF=dL:
ð6Þ
Fig. 6 summarizes the friction coefficients of the Zdol4000 films coated on the textured and nontextured disks that were measured using the glass sliding pins and DLC-coated sliding pins. The surface roughness decreased in the following orders: untreated,
60
R. Lu et al. / Wear 319 (2014) 56–61
Ar1, and Ar8 for the sliding pins that underwent an Ar-GCIB treatment; and untreated, DLC_A, and DLC_B sliding pins after they were coated with a DLC film and underwent a N2-GCIB treatment. Three immediate conclusions follow from the data in Fig. 6. First, the friction coefficient decreased with a decreasing sliding pin surface roughness, except for the textured disk and glass sliding pin combination. Second, the DLC-coated sliding pins had smaller friction coefficients than the glass sliding pins. Third, the friction coefficient of the lubricant film coated on the nontextured disk was higher than that on the textured disk for a given sliding pin.
4. Discussion The surface energy is widely recognized as an important factor that affects the friction behavior. In general, a low surface energy induces a small amount of friction [18]. As seen in Fig. 6, the DLC_A and DLC_B sliding pins gave rise to different friction coefficients, particularly when they were mounted on the nontextured disk. However, the surface energies of the two DLC-coated sliding pins should be almost the same, which was confirmed from the contact angle measurements on the DLC-coated glass plates. Thus, it can be inferred that the friction properties of the liquid lubricant films confined between solid surfaces were not exclusively determined by the surface energies of the solids. The surface roughness is crucial for the friction of nanometerthick liquid lubricant films between solid surfaces. Fig. 7 shows the variations in the friction coefficients with the surface roughness (σ). Here, σ is the composite standard deviation for the surface roughness of the pair of surfaces. σ is defined as qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi σ ¼ R2q1 þ R2q2 ; ð7Þ where Rq1 is the surface roughness of the sliding pin and Rq2 is the surface roughness of the disk. The solid lines in Fig. 7 were inserted as a guide for the eyes. The friction coefficient increased with an increasing surface roughness, except for the textured disk and glass sliding pin combination. However, the friction coefficient generally decreased with an increasing surface roughness for dry friction [19]. The opposite change in the friction coefficient with surface roughness indicated that the interplay between the lubricant film and the surface roughness should be taken into consideration. Let us consider an interface with isotropically distributed contacts, such as that formed between the glass or DLC-coated
0.4 Glass pin
Friction coefficient
0.35
Non-textured disk DLC pin Glass pin
0.3 0.25 0.2
DLC pin
Textured disk (longitudinal)
0.15 0.1
Glass sliding pin on non-textured disk DLCsliding pin on non-textured disk Glass sliding pin on textured disk DLC sliding pin on textured disk
0.05 0
0
0.5 1 1.5 2 2.5 Surface roughness σ, nm
3
Fig. 7. Friction coefficients as a function of the surface roughness. Error bars indicate the maximum deviation from the average of three duplicate measurements.
sliding pins and the nontextured disk. As the surface roughness decreased, the mean contact pressure between a pair of contacting asperities decreased [19], allowing more lubricant molecules to intervene between the contacting asperities. These intervening lubricant molecules resulted in decrease of the friction coefficient. When a lubricant film was coated on the textured disk that had a longitudinal roughness, the anisotropic effect of the disk surface became more noticeable at the contact interface with the glass sliding pins, as the pin surface roughness decreased. Thus, it can be suggested that the longitudinal anisotropy of the composite surface roughness gave rise to a high friction coefficient [7]. However, once the glass sliding pins were coated with a DLC film, the hardness and Young's modulus became much higher than that of the textured disk, as seen from Fig. 6. During the sliding contact measurements, the textured disk was more likely to become deformed than the DLC-coated sliding pins, suppressing the effect of the longitudinal roughness on the friction. This helps to explain the experimental results where, similar to the sliding pins on the nontextured disk, the friction coefficient of the DLC-coated sliding pins on the textured disk increased with an increasing surface roughness. There are two conceivable reasons for why the DLC-coated sliding pins had smaller friction coefficients than the glass sliding pins. One is that surface graphitization or surfaces with graphite inclusions on the DLC coatings contribute to the low friction performance [20]. The other is the change in the contact status, caused by the DLC coatings. When coated with a DLC film, the sliding pins became much harder than the disks, as described above. Moreover, the high Young's modulus of the DLC coating led to less elastic deformation of the sliding pins during contact with the disk. Therefore, in comparison with the glass sliding pins, the contact area between the DLC-coated sliding pins and the disk was less sensitive to changes in the external load, resulting in low friction coefficients. A similar explanation can be applied to the different friction coefficients for a given sliding pin on the nontextured disk and on the textured disk. Namely, as compared with the textured disks, the smaller hardness of the nontextured disk resulted in larger deformation of the disk during sliding contact, leading to a larger contact area and thereby a higher friction coefficient.
5. Conclusions By using glass and DLC-coated glass sliding pins with different surface roughnesses, the friction properties of textured and nontextured disks coated with nanometer-thick PFPE films were investigated. The friction properties of nanometer-thick liquid lubricant films, confined between two solid surfaces, were not exclusively determined by the surface energies of the solids and were sensitive to the surface roughness of the solid surfaces. Regardless of the materials in contact, the friction coefficient generally tended to increase with an increasing surface roughness (the composite standard deviation for the surface roughness of the pair of contacting surfaces). However, the friction coefficient decreased with an increasing surface roughness for the textured disk and glass sliding pin combination. This result suggested that the strong longitudinal roughness effect at the contact interface gave rise to the high friction coefficient. Moreover, the DLC-coated sliding pins had lower friction coefficients than the glass sliding pins, indicating that the hardness and Young's modulus of the contact materials affected the area that was contacted, and consequently, the friction properties of the confined lubricant films.
R. Lu et al. / Wear 319 (2014) 56–61
Acknowledgments This work was supported in part by the Funding Program for Next Generation World-Leading Researchers (No. GR053), Grantin-Aid for Scientific Research (C No. 22560153), the Industrial Technology Research Grant Program in 2006 from NEDO (No. 06A12003A), and a grant from the Storage Research Consortium. The authors would like to thank Dr Teruyuki Kitagawa from Nomura Plating Co., Ltd., for his help in the GCIB treatment. References [1] O.M. Braun, A.G. Naumovets, Nanotribology: microscopic mechanisms of friction, Surf. Sci. Rep. 60 (2006) 79–158. [2] L. Li, P.M. Jones, Y. Hsia, Characterization of a nanometer-thick sputtered polytetrafluoroethylene film, Appl. Surf. Sci. 257 (2011) 4478–4485. [3] H. Zhang, T. Takimoto, K. Fukuzawa, S. Itoh, Effect of ultraviolet irradiation on adhesion of nanometer-thick lubricant films coated on magnetic disk surface, IEEE Trans. Magn. 47 (2011) 94–99. [4] B. Zhang, H. Chiba, A. Nakajima, Blow-off flow of nano PFPE liquid film at hard disk surfaces, Tribol. Lett. 39 (2010) 193–199. [5] N. Li, Y. Meng, D.B. Bogy, Effect of PFPE lubricant properties on the critical clearance and rate of the lubricant transfer from disk surface to slider, Tribol. Lett. 43 (2011) 275–286. [6] U. Boettcher, H. Li, R.A. de Callafon, F.E. Talke, Dynamic flying height adjustment in hard disk drives through feed forward control, IEEE Trans. Magn. 47 (2011) 1823–1829. [7] R. Lu, H. Zhang, Y. Mitsuya, K. Fukuzawa, S. Itoh, Friction measurements of nanometer-thick lubricant films using ultra-smooth sliding pins treated with gas cluster ion beam, Appl. Surf. Sci. 280 (2013) 619–625. [8] H. Lee, B. Bhushan, Role of surface roughness and lubricant film thickness in nanolubrication of sliding components in adaptive optics, J. Colloid Interface Sci. 353 (2011) 574–581.
61
[9] A.A. Al-Azizi, O. Eryilmaz, A. Erdemir, S.H. Kim, Effect of nanoscale surface texture and lubricant molecular structure on boundary lubrication in liquid, Langmuir 29 (2013) 13419–13426. [10] I. Yamada, J. Matsuo, N. Toyoda, A. Kirkpatrick, Material processing by gas cluster ion beams, Mater. Sci. Eng. Rep. 34 (2001) 231–295. [11] T. Kitagawa, I. Yamada, N. Toyoda, H. Tsubakino, J. Matsuo, G. Takaoka, A. Kirkpatrick, Hard DLC film formation by gas cluster ion beam assisted deposition, Nucl. Instrum. Methods Phys. Res. B 201 (2003) 405–412. [12] A. Gandopadhyay, K. Sinha, D. Uy, D.G. Mcwatt, R.J. Zdrodowski, S.J. Simko, Friction, wear, and surface film formation characteristics of diamond-like carbon thin coating in valvetrain application, Tribol. Trans. 54 (2011) 104–114. [13] E.P. Ivanova, V.K. Truong, J.Y. Wang, C.C. Berndt, R.T. Jones, I.I. Yusuf, I. Peake, H. W. Schmidt, C. Fluke, D. Barnes, R.J. Crawford, Impact of nanoscale roughness of titanium thin film surfaces on bacterial retention, Langmuir 26 (2010) 1973–1982. [14] P.M. Hansson, L. Skedung, P.M. Claesson, A. Swerin, J. Schoelkopf, P.A.C. Gane, M.W. Rutland, E. Thormann, Robust hydrophobic surface displaying different surface roughness scales while maintaining the same wettability, Langmuir 27 (2011) 8153–8159. [15] X. Li, H. Gao, C.J. Murphy, K.K. Caswell, Nanoindentaion of silver nanowires, Nano Lett. 3 (2003) 1495–1498. [16] W.C. Oliver, G.M. Pharr, Measurement of hardness and elastic modulus by instrumented indentation: advances in understanding and refinements to methodology, J. Mater. Res. 19 (2004) 3–20. [17] J. Gao, W.D. Luedtke, D. Gourdon, M. Ruths, J.N. Israelachvili, U. Landman, Frictional forces and Amontons's law: from the molecular to the macroscopic scale, J. Phys. Chem. B 108 (2004) 3410–3425. [18] Y. Kim, F. Limanto, D.H. Lee, R.S. Fearing, R. Maboudian, Role of countersubstrate surface energy in macroscale friction of nanofiber arrays, Langmuir 28 (2012) 2922–2927. [19] J.A. Ogilvy, Numerical simulation of friction between contacting rough surfaces, J. Phys. D: Appl. Phys. 24 (1991) 2098–2109. [20] B. Vengudusamy, R.A. Mufti, G.D. Lamb, J.H. Green, H.A. Spikes, Friction properties of DLC/DLC contacts in base oil, Tribol. Int. 44 (2011) 922–932.