Home
Search
Collections
Journals
About
Contact us
My IOPscience
Effect of structure on the tribology of ultrathin graphene and graphene oxide films
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2015 Nanotechnology 26 135702 (http://iopscience.iop.org/0957-4484/26/13/135702) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 157.89.65.129 This content was downloaded on 24/05/2015 at 12:07
Please note that terms and conditions apply.
Nanotechnology Nanotechnology 26 (2015) 135702 (11pp)
doi:10.1088/0957-4484/26/13/135702
Effect of structure on the tribology of ultrathin graphene and graphene oxide films Hang Chen and Tobin Filleter Department of Mechanical & Industrial Engineering, 5 King’s College Road, Toronto, ON M5S 3G8 Canada E-mail: fi[email protected] Received 3 October 2014, revised 23 December 2014 Accepted for publication 28 January 2015 Published 9 March 2015 Abstract
The friction and wear properties of graphene and graphene oxide (GO) with varying C/O ratio were investigated using friction force microscopy. When applied as solid lubricants between a sliding contact of a silicon (Si) tip and a SiO2/Si substrate, graphene and ultrathin GO films (as thin as 1–2 atomic layers) were found to reduce friction by ∼6 times and ∼2 times respectively as compared to the unlubricated contact. The differences in measured friction were attributed to different interfacial shear strengths. Ultrathin films of GO with a low C/O ratio of ∼2 were found to wear easily under small normal load. The onset of wear, and the location of wear initiation, is attributed to differences in the local shear strength of the sliding interface as a result of the nonhomogeneous surface structure of GO. While the exhibited low friction of GO as compared to SiO2 makes it an economically viable coating for micro/nano-electro-mechanical systems with the potential to extend the lifetime of devices, its higher propensity for wear may limit its usefulness. To address this limitation, the wear resistance of GO samples with a higher C/O ratio (∼4) was also studied. The higher C/O ratio GO was found to exhibit much improved wear resistance which approached that of the graphene samples. This demonstrates the potential of tailoring the structure of GO to achieve graphene-like tribological properties. Keywords: graphene oxide, graphene, friction, wear, friction force microscopy, x-ray photoelectron spectroscopy, MEMS (Some figures may appear in colour only in the online journal) 1. Introduction
Graphene, a 2D sheet of carbon atoms, has attracted a great deal of interest from scientists most recently for its exceptional material properties including: high electric conductivity and beneficial mechanical properties such as ultrahigh strength [3–5]. Additionally, recent research has demonstrated that graphene exhibits ultralow friction properties. Using friction force microscopy (FFM) Filleter et al demonstrated that single layers and bilayers of graphene can reduce friction on a silicon carbide surface by approximately 10 times [6]. Lee et al also observed via FFM that friction of graphene is reduced as the number of layers increases up to a saturation level consistent to that of bulk graphite at a layer thickness of just 4 atomic layers [7]. This beneficial friction behavior of graphene as an ultrathin solid lubricant makes it a promising candidate for use within the confined geometries of MEMS/NEMS devices. In addition to ultralow friction properties, graphene also exhibits very beneficial adhesion
As the size of devices shrinks to micro and nano scales, surface forces begin to dominate as compared to body forces. This is particularly critical in micro/nano-electro-mechanical systems (MEMS/NEMS). Consequently small devices have an increased likelihood of failure due to tribological issues such as adhesion, friction, and wear when two components in the MEMS/NEMS make contact. Channel gaps between MEMS/NEMS surfaces can range from 1 μm to 100 nm, thus traditional lubrication and wear mitigation methods are ineffective due to their small size [1]. In particular, liquid based lubrication approaches often fail due to high viscosity and surface tension under such confined geometries. Correspondingly, novel nano-engineering approaches have begun to be explored to address tribology improvements between MEMS/NEMS surfaces [2]. 0957-4484/15/135702+11$33.00
1
© 2015 IOP Publishing Ltd Printed in the UK
Nanotechnology 26 (2015) 135702
H Chen and T Filleter
properties. Bunch et al showed that the adhesion energy of monolayer and multilayer graphene in contact with a silicon dioxide substrate is larger than the adhesion energies of typical micromechanical structures, demonstrating that thin coatings of graphene can be easily adhered to MEMS/NEMS substrates [8]. Graphene films, however, have a very high cost which may limit its application to lubricate real systems [9]. Graphene oxide (GO) is a more economical alternative to graphene due to a lower cost of mass production, and has great potential as a solid lubricant for MEMS/NEMS which may also allow for tunable friction properties through controlled surface functionalization [10]. In addition, GO can be chemically and thermally reduced to modify the C/O ratio and allow for tunable electronic and mechanical properties [11, 12]. These potential benefits have motivated initial studies of using GO as a solid lubricant layer. In one of the first studies, Ou et al covalently deposited GO onto a silicon substrate and investigated the friction properties using FFM. This study showed that the relative friction coefficient was reduced by 20% and 70% respectively after coating silicon with GO and reduced GO [13]. Wei et al studied the friction of reduced GO (which exhibits a higher C/O ratio than unreduced GO) by thermal reduction from approximately 100 °C–700 °C and demonstrated that the friction difference of GO exhibits almost a linear trend with increasing temperature [12], however, in this study the specific structure of the GO films at varying thermal reduction temperatures was not reported. These initial studies have demonstrated qualitatively that GO and reduced GO can be effective at reducing friction, however, a quantitative study in which the friction and adhesion forces which correspond to a specific GO structure (i.e. C/O ratio) are measured is still needed. In addition wear of GO at the nanoscale has not yet been comprehensively investigated which is particularly important for the application to MEMS devices over multiple cycles of operation. Previous studies have also primarily focused on thick GO films. Although several existing studies have investigated few layer GO, they have not measured local friction differences between few layers GO using the same scanning FFM tip or applied a quantitative force calibration method. Here we have quantitatively investigated the friction, adhesion, and wear properties of silicon dioxide/silicon, graphene, and GO with different C/O ratios at the nanoscale using FFM to characterize the tribological properties and xray photoelectron spectroscopy (XPS) to characterize the specific film structure.
cleaning in methanol using an ultrasonic bath. The wafers were then blown dry with nitrogen gas. Graphene was deposited onto the silicon substrates by mechanically exfoliating highly oriented pyrolytic graphite (SPI Supplies) via the scotch tape method [14]. GO aqueous solutions were obtained by dissolving 1.0 mg of GO flake with either low C/ O ratio (ACS Materials LLC) or high C/O ratio (CheapTubes Inc.) prepared by the modified Hummers method [15, 16] into 20 mL of deionized water (resistivity > 18 MΩ cm, MP Biomedicals, LLC) by sonication for 1 h in an ultrasonic bath. The as-prepared light brown GO solution was then diluted to 0.01 mg ml−1. The obtained solution was then centrifuged at 4000 rpm for 30 min to remove unexfoliated GO using an Eppendorf MiniSpin centrifuge to produce the supernatant. Prior to GO deposition the silicon wafers were treated with aqueous potassium hydroxide (KOH) solution (weight concentration: 50%) for 15 min to make the surface more hydrophilic. Using this surface treatment the GO films were found to be more uniformly coated on the surfaces as compared to similar GO depositions on untreated surfaces. The improved coating is in part a result of the reduction of the contact angle of the solution deposited on the treated silicon wafer [17]. Control measurements using FFM conducted on untreated and KOH treated surfaces showed no significant difference in friction or adhesion properties. The silicon wafers were then cleaned with the same method of substrate cleaning described previously for the graphene sample preparation. To prepare all GO samples for FFM measurements GO solutions (8 μl) were drop cast onto silicon wafers with a micropipette and afterwards the solution was dried in air. 2.2. Film structure characterization and FFM
Thin films of graphene were identified via an optical microscope (Zeiss, Germany) [18]. Raman spectroscopy (Renishaw, 532 nm laser excitation) was performed to characterize the microstructure of the graphene and GO samples, and the surface morphologies of graphene and GO samples were observed by an atomic force microscope (AFM) (Asylum MFP3D) in tapping mode. XPS was also conducted on the GO samples to identify the specific chemical structure of the functional groups as well as measure the C/O ratio. The nanotribological properties of graphene and GO samples were measured using an AFM in FFM mode using rectangular cantilevers (Nanoworld FMR10, normal spring constant: ∼1.3 nN nm−1) in ambient conditions of 25 °C and 45% relative humidity. For each FFM measurement first the adhesion force between the AFM tip and film was measured by recording a force distance curve and measuring the adhesion force as the maximum force required to pull the AFM tip out of contact with the surface. The error of the adhesion force was calculated as the standard deviation of several (minimum of ten measurements) pull-off force measurements recorded on different regions of the surface in the near vicinity of the area used for friction measurements. Adhesion forces measured for different samples were measured under the same ambient condition to minimize any influence on changes in the meniscus [19].
2. Experimental section 2.1. Materials and sample preparation
Graphene and GO films were prepared on silicon (Si) substrates with a silicon dioxide (SiO2) surface layer. The silicon wafers (N-doped, 100 orientation, 285 nm SiO2 thickness) were first cleaned for 10 min in acetone followed by a 10 min 2
Nanotechnology 26 (2015) 135702
H Chen and T Filleter
3. Results
Quantitative FFM was used to measure lateral force signals simultaneously over surface regions containing two surfaces within the same scan line (either SiO2 and graphene, or SiO2 and GO) to make a direct comparison of lateral force differences between the two surfaces. As the AFM tip was scanned laterally across a surface, lateral force signals in both the forward scan and backward scan directions were recorded while sliding. The friction force was then calculated as half the difference between the forward and backward lateral force. The normal and lateral forces acting on the tip are given by [20]: FN = k N SN Va − b,
FL =
1 k T SL Vc− d , h2
3.1. Structure of graphene and GO films
The thickness of graphene was characterized using AFM in tapping mode (see figure 1(c) and line profile in figure 1(e)). From the AFM topography image in figures 1(c) and (e) for example, we observed a multilayer graphene film with a thickness of approximately 2.9 nm. A similar approach was used to observe, and measure the thickness of GO films (see figure 1(d) and line profile in figure 1(f)). The presence of graphene and GO were then verified using Raman spectroscopy [22, 23] as shown in figure 2. The Raman spectrum of graphene indicates the G peak located at 1582 cm−1 and 2D peak at 2675 cm−1, which are signatures of the in-plane vibration and second-order zone boundary phonons respectively. The absence of D peak in the Raman spectrum of graphene indicates that the exfoliated graphene in this study has a low defect density. The Raman spectrum of the GO shows a prominent peak at 1347 cm−1 with intensity comparable to the G peak at 1599 cm−1, which is characteristic of GO, and is indicative of significant structural disorder due to the defects in the GO plane. The thickness of single layer GO was found to vary from 0.6 to 1.2 nm, which is in agreement with literature values, and can be understood by its hygroscopic nature [24]. Thicker GO films were also observed. For example, the GO flake in figure 1(d) has thickness of ∼3.3 nm and thus it has a layer thickness of approximately three to four layers. XPS survey data indicated that the low C/O ratio GO studied has a C/O ratio of ∼2 while the high C/O ratio GO has a ratio of ∼3.7. Differences between the structures of the two samples were investigated with high resolution XPS measurement as shown in figure 3, which clearly indicate different concentrations of oxygen functional groups in the two samples studied. In addition to the chemical states associated to the oxygen functional groups, peaks for both sp3 and sp2 bonding of carbon were also included for XPS fitting.
(1)
(2)
where Va-b and Vc-d are the normal and lateral deflection signals from the four quadrant photodetector recorded during FFM scanning, kN and kT are normal and torsional spring constant respectively, Sn and Sl are the normal and lateral deflection sensitivity of the photodetector respectively, and h is the tip height. To quantify lateral force, the spring constant of the cantilever and the deflection sensitivity of photodetector were first measured. Normal and torsional spring constants were measured using the Sader method [21]. Both the normal and torsional spring constant require the resonant frequency and quality factor, which were measured from the thermal noise spectra of the normal and lateral deflection signals. The deflection sensitivity of the photodetector was determined from the slope of the cantilever deflection versus distance curve. The lateral deflection sensitivity of the photodetector, similarly, is determined from lateral deflection signal as a function of lateral displacement. Measurement of lateral deflection sensitivity is in general more difficult than the normal deflection sensitivity due to the higher lateral stiffness of the cantilever and possible in-plane bending in addition to the torsional response. The test probe method was used under ambient conditions to measure the lateral deflection sensitivity which avoids such issues, and has been shown to yield accurate measurements of the deflection sensitivity [20]. For this method a 70 μm diameter colloidal glass sphere was attached to a silicon cantilever (the same type of cantilever used for FFM) by epoxy and then the lateral signal versus distance was obtained via pushing the colloidal sphere against a freshly cleaved side wall (100) of a KBr crystal. The slope of the contact region in the resulting force plot gives the lateral deflection sensitivity (SL,test). The lateral deflection sensitivity of the photo detector with the sharp integrated tips used for FFM measurements (SL) was then calculated from SL,test considering the differences of the torsional arm length and lateral in-plane bending parameters between the colloid probe and sharp integrated tip. The resulted SL was then used for the calculation of lateral force from the torsional signals.
3.2. Friction and wear of graphene on SiO2/Si
Figure 4 shows the height topography and lateral force map of graphene on a SiO2/Si surface, and figure 5 shows friction versus normal force curves. Friction on multilayer graphene (4 layers) was found to be much lower than that of SiO2, as shown in figure 5, where friction measured on graphene was ∼6 times lower than on SiO2 under a normal load condition of ∼200 nN. Root mean square (rms) roughness of a 1.5 nm layer of multilayer graphene film and the SiO2 substrate was measured to be 0.39 ± 0.07 and 0.22 ± 0.03 nm respectively. The graphene has a larger rms roughness than SiO2, which suggests that the lower friction of graphene as compared to SiO2 is not due to its roughness. Despite the large friction difference, the force of adhesion (pull-off force between tip and sample) of graphene and SiO2/Si was found to be 25 ± 1 and 27.1 ± 0.9 nN respectively, suggesting that the friction difference is an intrinsic difference in the energy difference due to sliding. This is in agreement with the observation of a 3
Nanotechnology 26 (2015) 135702
H Chen and T Filleter
Figure 1. (a) Optical image of exfoliated graphene on a SiO2/Si substrate (white arrow identifies the graphene film); (b) GO aqua-solution in centrifuge tube after centrifuge (precipitates are below the red dashed line and black arrow identifies the precipitates); (c) AFM tapping mode image of multilayer graphene film on a SiO2/Si substrate and corresponding line profile (e) along the black dashed line showing a multilayer graphene flake on SiO2/Si substrate; (d) AFM tapping mode image of GO film on SiO2/Si substrate and corresponding line profile (f) showing a multilayer graphene oxide flake on SiO2/Si substrate.
modulus and Poisson’s ratio of 70 GPa and 0.17 [26] were used in the model for silicon dioxide (both silicon dioxide substrate and silicon dioxide tip) and graphene samples. Given the ultrathin nature of the graphene films we have assumed that the modulus of the film/substrate surface is dominated by the underlying thick silicon dioxide substrate [27]. Alternatively the modulus of graphite (∼30 GPa) could have been used to estimate the multilayered graphene modulus loaded out-of plane. The tip radius was assumed to be 15 nm in this model, it should be noted that while this is an estimate of the tip radius, the same tip was used for
change in the slope of the friction versus normal force curves in figure 5. In order to estimate the interfacial shear strength between the two contacts under study, the friction versus normal load curves were fit using the generalized Maugis–Dugdale model [25], which allows determination of an interface’s contact behavior that is intermediate between the Derjaguin−Mueller −Toporov (DMT) assumption of a ‘hard’ contact and the Johnson−Kendall−Roberts (JKR) assumption of a ‘soft’ contact. The generalized Maugis–Dugdale model is referred to as ‘DMT–JKR transition model’ in this study. A Young’s 4
Nanotechnology 26 (2015) 135702
H Chen and T Filleter
1000
(b)
2D
lntensity (a.u.)
lntensity (a.u.)
(a)
G
2000
Raman shift
(cm-1)
1000
3000
D
G
2D
2000
Raman shift
(cm-1)
3000
Figure 2. (a) Raman spectrum of exfoliated graphene on a SiO2/Si substrate: G peak at 1582 cm−1 and 2D peak at 2675 cm−1; (b) Raman
spectrum of GO films on a SiO2/Si substrate: D peak at 1347 cm−1 and G peak at 1599 cm−1 (the Raman spectra of graphene and GO were both taken at a laser wavelength of 532 nm).
Figure 3. XPS measurements of (a) low C/O ratio (∼2) and (b) high C/O ratio (∼3.7) graphene oxide samples.
simultaneous measurements on the graphene and silicon dioxide surfaces. From the fits the interfacial shear strength between the silicon tip and graphene, and the silicon tip and silicon dioxide surfaces was found to be 173 ± 13 MPa (∼125 MPa if the modulus of graphite is instead used) and 1800 ± 44 MPa respectively. This measured interfacial shear strength between the tip and the silicon dioxide surface is on the same order of magnitude as compared to that reported in previous literature with a similar contact system [27]. This again confirms that the difference in friction behavior is due to different interfacial shear strength and not any significant difference in adhesion. The observed friction difference between graphene and silicon dioxide is also in good agreement with previous reported results in the literature [27] where friction forces were found to decrease by ∼90% on supported monolayer graphene relative to SiO2. No wear was observed on the graphene surface after the friction measurements shown in figure 5, conducted at up to normal forces of ∼200 nN.
3.3. Friction and wear of low C/O ratio GO on SiO2/Si
Figures 6(a) and (b) shows the height topography of a 1–2 layer GO film (∼1.2 nm) with a C/O ratio of ∼2 on a SiO2/Si surface. Figure 6(c) shows the lateral force of the forward and backward scan along the black dashed line in figure 6(a) at a normal load of −13.8 nN. From figure 6(c) it can be seen that friction on the SiO2/Si surface is approximately 2.3 times larger than on the 1–2 layer GO film at a normal force of −13.8 nN. This demonstrates that, as is also true for graphene films, just one or two atomic layers of GO is effective at reducing friction on silicon dioxide surfaces that are often used for MEMS. Figure 6(d) shows a zoom out AFM topography image of the FFM scan area, which indicates that some of the GO area was worn (e.g. the red dashed rectangle area in figure 6(d)) while other areas are still effectively lubricating the SiO2 surface. In addition, worn GO particles were found to be piled up at the edge of the scan area after the FFM scan as shown in figure 6(d). In order to correlate this wear behavior with friction force, friction versus normal load 5
Nanotechnology 26 (2015) 135702
H Chen and T Filleter
Figure 4. (a)AFM height topography (contact mode) image and corresponding line profile (c) along the while dashed line showing a 4 layer graphene film on a SiO2/Si substrate; (b) lateral force map of forward scan and corresponding lateral force line profile (d) of both forward and backward scan along the same line.
higher than ∼0 nN. The onset of wear can be identified by the discontinuous jump in the friction versus normal load curve [28]. From the different friction and wear behavior on these areas it is clear that some areas of the GO flake are able to resist wear at relatively higher normal load. This behavior can be attributed to the heterogeneous nature of the GO surface [29] which will be discussed in detail later in the context of the proposed wear mechanism for thin GO films. A similar transition in the friction versus normal load curves that identify the onset of wear was also observed for multilayer GO films. Figures 7(a) and (b) shows the height topography and lateral force map of a multilayer GO film (∼3 nm) on a SiO2/Si surface. From figure 7 the friction of SiO2/Si is found to be ∼1.9 times higher than that of multilayer GO at a normal force of −3.6 nN. Adhesion between the silicon tip/silicon dioxide and the silicon tip/GO were measured to be 4.4 ± 0.3 and 4.1 ± 0.2 nN respectively. It should be noted that the rms roughness of the GO film covered areas was measured using AFM to be the same as the uncoated SiO2 areas within measurement error, which rules out the possibility of differences in rms roughness as the dominant mechanism for the observed friction difference between GO and SiO2. From figure 7(c) a discontinuity in the friction force was observed when the normal force was higher than ∼9 nN. From figures 8(a)–(c), it can be observed that the GO flake became smaller and smaller with increasing normal load during the FFM scan. This onset of wear was found to initiate
Figure 5. Friction force versus normal force of silicon dioxide and
graphene (4 layers) on a SiO2/Si substrate. The solid black and red lines are fits using the DMT–JKR transition model.
curves for FFM scans on GO (both on the worn area and unworn area) and the SiO2/Si surface were plotted in figure 6(e) to give a direct comparison between the three surface areas. From figure 6(e), there is an indication that the GO films started to wear when the normal force reached 6
Nanotechnology 26 (2015) 135702
H Chen and T Filleter
Figure 6. (a)AFM height topography image (contact mode) under normal force of −13.8 nN; (b) line profile along the black dashed line in the topography (a) showing a ∼1.2 nm GO film on a SiO2/Si substrate; (c) lateral force of both forward and backward scan along the same black dashed line in (a); (d) zoom out image (tapping mode) of the FFM scan area (black dashed rectangle area corresponds to the area of image (a) scanned in FFM mode) and the red dashed rectangle indicates the worn area of GO during FFM scanning; (e) friction force versus normal force of SiO2 surface area, worn GO surface area, and unworn GO surface area.
at a normal load of ∼10 nN which is consistent with the discontinuity observed in the friction force plot. GO appears to be suddenly displaced in the lateral direction resulting in a jump in friction force and subsequently removed gradually from the substrate, potentially arising from plastic deformation. Repetition of the experiment with another GO flake under the same conditions exhibited worn GO particles accumulated at the edge of the scan area. Similar to the transition fit used for friction load curves of graphene data, the friction versus normal load curves of GO (before wear occurred) were selected and fit using the DMT– JKR transition model. The Young’s modulus and Poisson’s ratio of 70 GPa and 0.17 [26] were used in the model for silicon dioxide (both silicon dioxide substrate and silicon dioxide tip) and GO sample. The tip radius was also assumed to be 15 nm in the transition model. The interfacial shear strength between silicon tip and GO, and the silicon tip and silicon dioxide surface was found to be 1200 ± 180 and 1930 ± 282 MPa respectively.
3.4. Wear resistance of high C/O ratio GO films on SiO2/Si
In the previous section low C/O ratio (∼2) GO was found to exhibit lower friction as compared to silicon dioxide surfaces, however it was also found to be easily worn under a few nN normal load. Graphene, however, was found to resist wear under normal loads of up to several hundred nN. As was discussed earlier, reduced GO (which exhibits a higher C/O ratio than unreduced GO) exhibits lower friction than GO [12, 13]. The effect of increasing the C/O ratio of GO may also play a role in improving the wear behavior of GO at the nanoscale as it yields a more graphene-like surface structure. FFM combined with XPS was used to characterize the effect of chemical composition on the wear properties of the high C/O ratio GO samples. Figure 9(a) shows a tapping mode AFM image of a region of a high C/O ratio (∼3.7) GO film (∼1.1 nm thickness) on a SiO2 substrate. The region in the white area in figure 9(a) was scanned with normal loads ranging from 0 to 270 nN. From the topography image of figures 9(b)–(d), it was observed that the GO was wear 7
Nanotechnology 26 (2015) 135702
H Chen and T Filleter
Figure 7. (a) AFM contact mode topography image of multilayer GO a SiO2/Si substrate and (c) line profile along the black dashed line in the topography image; (b) lateral force map and (d) line profile of lateral force signal along the same line as shown in (d); (e) friction force versus normal force plot.
4. Discussion
resistant up to a normal load of up to 270 nN. This behavior was consistently observed for several similar high C/O ratio samples. This high wear resistance is quite different from the easy tendency of wear of the low C/O ratio GO samples which always began to wear at normal loads of just a few nN. This correlation between the high C/O ratio structure of the GO films and wear resistance properties similar to that of the graphene films demonstrates the clear advantages of the tailorability of the GO surface structures.
4.1. Friction
If we compare the results found on GO to graphene, we find that friction of 1–2 layer GO is approximately 3 times higher. Considering that the relative adhesion for GO and graphene and the substrate is similar, this suggests a higher interfacial shear strength between GO and silicon dioxide than that
8
Nanotechnology 26 (2015) 135702
H Chen and T Filleter
Figure 8. AFM contact mode topography images of GO on a SiO2 surface at normal force of −3.6 nN (a), 10.8 nN (b) and 21.6 nN (c) respectively.
between graphene and silicon dioxide. The different interfacial shear strength is likely due to the presence of the dangling functional groups and higher defect density in the 2D GO plane as revealed by high resolution TEM studies [29]. This is consistent with molecular dynamics simulation which have predicted that functional groups, such as epoxide and hydroxl groups, on the GO plane increase the shear strenght as compared to a pristine graphene structure [30]. The friction difference between GO and SiO2 reported here (∼1.9–2.3) is slightly lower as compared to that previously reported by Ou et al (∼2–3 times) [13]. This difference may be attributed to the fact that the silicon wafer studied by Ou et al were treated with pirahna which may increase the roughness of the silicon surface resulting in higher friction in the absence of the GO film. In addition, previous studies of thick GO films (∼50–400 nm) by Liang et al [31] exhibit the same general finding of reduced friction on the GO surfaces as compared to SiO2 as reported here where they observed a friction reduction of up to ∼6 times for the GO coated surfaces, however, it is difficult to quantitativly compare the two studies as they utlize very different contact sizes, GO sample structures and compositions, and loading conditions. One major difference between the findings reported here and those reported by Liang et al stems from the size scale of the contacts under study. Liang et al found that for macroscopic sliding contacts the thick GO films exhibited an increasing trend between film thickness and friciton coefficient for their micro-tribometer experiments of thick GO films on SiO2 that was attributed to differences in film roughness [31]. Here we have found instead that when the sliding contact is on the nanoscale there is no significant friction dependance on thickness. This is a very important observation as it suggests that the previously reported thickness dependance by Liang et al which is a result of varrying surface roughness, is a macroscale phenomena which emerges for multi asperity contacts and is not exhibited for nanoscale contacts.
4.2. Wear
High C/O ratio GO was found to exhibit a high wear resistance as compared to the easy tendency of wear of the low C/ O ratio GO samples. This correlation between the high C/O ratio structure of the GO films and wear resistance properties similar to that of the graphene films demonstrates the clear advantages of the tailorability of the GO surface structures. The difference in wear behavior for graphene, high C/O ratio GO, and low C/O ratio GO can be understood based on the varying surface structures of the three specimen. In the case of the low C/O ratio GO specimens the first stages of wear are likely to initiate at lower normal loads than for graphene due to differences in the local shear strength at different locations on the GO surface with different functional groups [30] and at defects in the GO plane which act as nucleation sites for wear to initiate. Wang et al have shown through DFT calculations that epoxide and hydroxyl functional groups that exist in GO lead to a local higher energy corrugation and shear strength for sliding between small flakes of GO as compared to graphene [30]. Microscopic GO films (in the lateral dimension), such as those that are under investigation here, have a different structure than the very small flakes studied in the DFT studies. Larger experimental GO flakes are instead composed of a mixture of non-homogenous regions of pristine graphene-like and functionalized GO-like patches. This has been demonstrated recently through high resolution TEM imaging of GO surfaces [29]. Therefore we expect that there will be local differences in the shear strength between the sliding tip and the GO surface at these different surface regions. XPS experiments conducted on the low C/O ratio GO samples studied here show the presence of ∼35 atomic% of epoxide, ∼15 atomic% of carbonyl, and ∼9 atomic% of carboxyl groups that make up these functionalized regions of the surface. We believe it is at these local regions on the GO surface with higher shear strength that wear initiates when the normal force reaches a critical level and the in plane stress is such that the fracture strength of the GO films is reached. Wear then progresses in the form of the local fracture and removal of small regions of 9
Nanotechnology 26 (2015) 135702
H Chen and T Filleter
Figure 9. (a) Tapping mode AFM image of high C/O ratio GO on a SiO2 substrate. (b), (c) and (d) are topography images for scanning under normal loads of 0, 60 and 270 nN respectively from FFM scans in the white dashed rectangular area in (a).
the GO films. In addition, the non-homogenous nature of the GO surface can explain why some regions wear at lower normal loads than others. This is significantly different behavior than that observed for graphene and high C/O ratio films which exhibited wear free sliding for normal forces of ∼200 nN, almost two orders of magnitude higher. High C/O ratio GO has a significantly reduced density of functional groups on its surface and therefore exhibits a more graphenelike surface structure. This points to a major limitation in using GO with a high density of functional groups as a solid lubricant under larger applied stress, as it contains many surface regions that tend to wear more easily than graphene.
friction of 1–2 layer and few-layer (3–4 layers) GO were found to range between 1.9–2.3 times lower than that of SiO2/ Si under similar normal loading conditions with no significant dependence on film thickness, suggesting that GO film thickness dose not play a primary role for nanoscale sliding contacts. A major difference in the wear behavior was observed between graphene and GO films. Graphene was found to resist wear with normal loads of up to ∼200 nN. In contrast, atomic thin GO sheet with C/O ratios of ∼2 tend to wear easily under normal loads of a few nN with an onset of wear that is attributed to higher local shear strengths at the nonhomogeneous surface regions on the GO films as compared to graphene. The wear behavior of GO was also found to be spatially dependant on the non-homogeneous surface regions. This limitation can be overcome by modifying the structure of the GO films. GO films with higher C/O ratios, which exhibit structures more similar to pristine graphene, were found to be wear resistant up to normal loads of ∼270 nN. These findings indicate the great importance of the local surface structures of GO films on their tribological behavior, and motivate additional future studies which focus on how variations in this
5. Summary and conclusion The tribological properties of SiO2/Si, graphene and GO with different C/O ratios were investigated using quantitative FFM with nanoscale sliding contacts. Friction of few-layer graphene was found to be approximately 6 times lower than that of SiO2/Si, which is attributed to lower interfacial shear strength between graphene and silicon dioxide. In contrast 10
Nanotechnology 26 (2015) 135702
H Chen and T Filleter
surface structure can lead to tunable approaches for lubrication and wear resistant coatings. While the lower friction of GO as compared to SiO2/Si makes it a potential economical coating for M/NEMS devices to extend the lifetime of devices and reduce energy dissipation, careful considerations in controlling its structure must be taken to limit its higher propensity for wear in the case of low C/O ratio structures as compared to graphene when applying it as a solid lubricant.
[13] Ou J et al 2010 Tribology study of reduced graphene oxide sheets on silicon substrate synthesized via covalent assembly Langmuir 26 15830–6 [14] Novoselov K S et al 2005 Two-dimensional atomic crystals Proc. Natl Acad. Sci. USA 102 10451–3 [15] Hummers W S and Offeman R E 1958 Preparation of graphitic oxide J. Am. Chem. Soc. 80 1339–1339 [16] Kovtyukhova N I et al 1999 Layer-by-layer assembly of ultrathin composite films from micron-sized graphite oxide sheets and polycations Chem. Mater. 11 771–8 [17] Franssila S 2010 Wafer cleaning and surface preparation Introduction to Microfabrication (Chichester: Wiley) chapter 12 pp 143–52 [18] Blake P et al 2007 Making graphene visible Appl. Phys. Lett. 91 063124-3 [19] Bhushan B and Dandavate C 2000 Thin-film friction and adhesion studies using atomic force microscopy J. Appl. Phys. 87 1201–10 [20] Cannara R J, Eglin M and Carpick R W 2006 Lateral force calibration in atomic force microscopy: a new lateral force calibration method and general guidelines for optimization Rev. Sci. Instrum. 77 053701–11 [21] Sader J E, Chon J W M and Mulvaney P 1999 Calibration of rectangular atomic force microscope cantilevers Rev. Sci. Instrum. 70 3967–9 [22] Ferrari A C et al 2006 Raman spectrum of graphene and graphene layers Phys. Rev. Lett. 97 187401 [23] Kudin K N, Ozbas B, Schniepp H C, Prud’homme R K, Aksay I A and Car R 2007 Raman spectra of graphite oxide and functionalized graphene sheets Nano Lett. 8 36–41 [24] Stankovich S et al 2007 Synthesis of graphene-based nanosheets via chemical reduction of exfoliated graphite oxide Carbon 45 1558–65 [25] Carpick R W, Ogletree D F and Salmeron M 1999 A general equation for fitting contact area and friction vs load measurements J. Colloid Interface Sci. 211 395–400 [26] Kim M T 1996 Influence of substrates on the elastic reaction of films for the microindentation tests Thin Solid Films 283 12–6 [27] Deng Z, Klimov N N, Solares S D, Li T, Xu H and Cannara R J 2012 Nanoscale interfacial friction and adhesion on supported versus suspended monolayer and multilayer graphene Langmuir 29 235–43 [28] Steiner P et al 2010 Atomic friction investigations on ordered superstructures Tribol. Lett. 39 321–7 [29] Erickson K, Erni R, Lee Z, Alem N, Gannett W and Zettl A 2010 Determination of the local chemical structure of graphene oxide and reduced graphene oxide Adv. Mater. 22 4467–72 [30] Wang L-F, Ma T-B, Hu Y-Z and Wang H 2012 Atomic-scale friction in graphene oxide: an interfacial interaction perspective from first-principles calculations Phys. Rev. B 86 125436 [31] Liang H, Bu Y, Zhang J, Cao Z and Liang A 2013 Graphene oxide film as solid lubricant ACS Appl. Mater. Interfaces 5 6369–75
Acknowledgments This work was supported by the Canada Foundation for Innovation and NSERC. The authors would like to thank Changhong Cao for assistance in conducting Raman spectroscopy experiments. The authors would also like to thank Rana Sodhi and Peter Broderson for conducting XPS data collection and analysis of the GO samples.
References [1] Chandross M, Lorenz C, Grest G, Stevens M and Webb E III 2005 Nanotribology of anti-friction coatings in MEMS JOM 57 55–61 [2] Kim S H, Asay D B and Dugger M T 2007 Nanotribology and MEMS Nano Today 2 22–9 [3] Castro Neto A H, Guinea F, Peres N M R, Novoselov K S and Geim A K 2009 The electronic properties of graphene Rev. Mod. Phys. 81 109–62 [4] Lee C, Wei X, Kysar J W and Hone J 2008 Measurement of the elastic properties and intrinsic strength of monolayer graphene Science 321 385–8 [5] Lee G-H et al 2013 High-strength chemical-vapor–deposited graphene and grain boundaries Science 340 1073–6 [6] Filleter T et al 2009 Friction and dissipation in epitaxial graphene films Phys. Rev. Lett. 102 086102 [7] Lee C et al 2010 Frictional characteristics of atomically thin sheets Science 328 76–80 [8] Koenig S P, Boddeti N G, Dunn M L and Bunch J S 2011 Ultrastrong adhesion of graphene membranes Nat. Nanotechnology 6 543–6 [9] Graphene Supermarket 2015 http://graphene-supermarket.com [10] Ou J, Wang Y, Wang J, Liu S, Li Z and Yang S 2011 Selfassembly of octadecyltrichlorosilane on graphene oxide and the tribological performances of the resultant film J. Phys. Chem. C 115 10080–6 [11] Compton O C, Jain B, Dikin D A, Abouimrane A, Amine K and Nguyen S T 2011 Chemically active reduced graphene oxide with tunable C/O ratios ACS Nano 5 4380–91 [12] Wei Z et al 2010 Nanoscale tunable reduction of graphene oxide for graphene electronics Science 328 1373–6
11
Search
Collections
Journals
About
Contact us
My IOPscience
Effect of structure on the tribology of ultrathin graphene and graphene oxide films
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2015 Nanotechnology 26 135702 (http://iopscience.iop.org/0957-4484/26/13/135702) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 157.89.65.129 This content was downloaded on 24/05/2015 at 12:07
Please note that terms and conditions apply.
Nanotechnology Nanotechnology 26 (2015) 135702 (11pp)
doi:10.1088/0957-4484/26/13/135702
Effect of structure on the tribology of ultrathin graphene and graphene oxide films Hang Chen and Tobin Filleter Department of Mechanical & Industrial Engineering, 5 King’s College Road, Toronto, ON M5S 3G8 Canada E-mail: fi[email protected] Received 3 October 2014, revised 23 December 2014 Accepted for publication 28 January 2015 Published 9 March 2015 Abstract
The friction and wear properties of graphene and graphene oxide (GO) with varying C/O ratio were investigated using friction force microscopy. When applied as solid lubricants between a sliding contact of a silicon (Si) tip and a SiO2/Si substrate, graphene and ultrathin GO films (as thin as 1–2 atomic layers) were found to reduce friction by ∼6 times and ∼2 times respectively as compared to the unlubricated contact. The differences in measured friction were attributed to different interfacial shear strengths. Ultrathin films of GO with a low C/O ratio of ∼2 were found to wear easily under small normal load. The onset of wear, and the location of wear initiation, is attributed to differences in the local shear strength of the sliding interface as a result of the nonhomogeneous surface structure of GO. While the exhibited low friction of GO as compared to SiO2 makes it an economically viable coating for micro/nano-electro-mechanical systems with the potential to extend the lifetime of devices, its higher propensity for wear may limit its usefulness. To address this limitation, the wear resistance of GO samples with a higher C/O ratio (∼4) was also studied. The higher C/O ratio GO was found to exhibit much improved wear resistance which approached that of the graphene samples. This demonstrates the potential of tailoring the structure of GO to achieve graphene-like tribological properties. Keywords: graphene oxide, graphene, friction, wear, friction force microscopy, x-ray photoelectron spectroscopy, MEMS (Some figures may appear in colour only in the online journal) 1. Introduction
Graphene, a 2D sheet of carbon atoms, has attracted a great deal of interest from scientists most recently for its exceptional material properties including: high electric conductivity and beneficial mechanical properties such as ultrahigh strength [3–5]. Additionally, recent research has demonstrated that graphene exhibits ultralow friction properties. Using friction force microscopy (FFM) Filleter et al demonstrated that single layers and bilayers of graphene can reduce friction on a silicon carbide surface by approximately 10 times [6]. Lee et al also observed via FFM that friction of graphene is reduced as the number of layers increases up to a saturation level consistent to that of bulk graphite at a layer thickness of just 4 atomic layers [7]. This beneficial friction behavior of graphene as an ultrathin solid lubricant makes it a promising candidate for use within the confined geometries of MEMS/NEMS devices. In addition to ultralow friction properties, graphene also exhibits very beneficial adhesion
As the size of devices shrinks to micro and nano scales, surface forces begin to dominate as compared to body forces. This is particularly critical in micro/nano-electro-mechanical systems (MEMS/NEMS). Consequently small devices have an increased likelihood of failure due to tribological issues such as adhesion, friction, and wear when two components in the MEMS/NEMS make contact. Channel gaps between MEMS/NEMS surfaces can range from 1 μm to 100 nm, thus traditional lubrication and wear mitigation methods are ineffective due to their small size [1]. In particular, liquid based lubrication approaches often fail due to high viscosity and surface tension under such confined geometries. Correspondingly, novel nano-engineering approaches have begun to be explored to address tribology improvements between MEMS/NEMS surfaces [2]. 0957-4484/15/135702+11$33.00
1
© 2015 IOP Publishing Ltd Printed in the UK
Nanotechnology 26 (2015) 135702
H Chen and T Filleter
properties. Bunch et al showed that the adhesion energy of monolayer and multilayer graphene in contact with a silicon dioxide substrate is larger than the adhesion energies of typical micromechanical structures, demonstrating that thin coatings of graphene can be easily adhered to MEMS/NEMS substrates [8]. Graphene films, however, have a very high cost which may limit its application to lubricate real systems [9]. Graphene oxide (GO) is a more economical alternative to graphene due to a lower cost of mass production, and has great potential as a solid lubricant for MEMS/NEMS which may also allow for tunable friction properties through controlled surface functionalization [10]. In addition, GO can be chemically and thermally reduced to modify the C/O ratio and allow for tunable electronic and mechanical properties [11, 12]. These potential benefits have motivated initial studies of using GO as a solid lubricant layer. In one of the first studies, Ou et al covalently deposited GO onto a silicon substrate and investigated the friction properties using FFM. This study showed that the relative friction coefficient was reduced by 20% and 70% respectively after coating silicon with GO and reduced GO [13]. Wei et al studied the friction of reduced GO (which exhibits a higher C/O ratio than unreduced GO) by thermal reduction from approximately 100 °C–700 °C and demonstrated that the friction difference of GO exhibits almost a linear trend with increasing temperature [12], however, in this study the specific structure of the GO films at varying thermal reduction temperatures was not reported. These initial studies have demonstrated qualitatively that GO and reduced GO can be effective at reducing friction, however, a quantitative study in which the friction and adhesion forces which correspond to a specific GO structure (i.e. C/O ratio) are measured is still needed. In addition wear of GO at the nanoscale has not yet been comprehensively investigated which is particularly important for the application to MEMS devices over multiple cycles of operation. Previous studies have also primarily focused on thick GO films. Although several existing studies have investigated few layer GO, they have not measured local friction differences between few layers GO using the same scanning FFM tip or applied a quantitative force calibration method. Here we have quantitatively investigated the friction, adhesion, and wear properties of silicon dioxide/silicon, graphene, and GO with different C/O ratios at the nanoscale using FFM to characterize the tribological properties and xray photoelectron spectroscopy (XPS) to characterize the specific film structure.
cleaning in methanol using an ultrasonic bath. The wafers were then blown dry with nitrogen gas. Graphene was deposited onto the silicon substrates by mechanically exfoliating highly oriented pyrolytic graphite (SPI Supplies) via the scotch tape method [14]. GO aqueous solutions were obtained by dissolving 1.0 mg of GO flake with either low C/ O ratio (ACS Materials LLC) or high C/O ratio (CheapTubes Inc.) prepared by the modified Hummers method [15, 16] into 20 mL of deionized water (resistivity > 18 MΩ cm, MP Biomedicals, LLC) by sonication for 1 h in an ultrasonic bath. The as-prepared light brown GO solution was then diluted to 0.01 mg ml−1. The obtained solution was then centrifuged at 4000 rpm for 30 min to remove unexfoliated GO using an Eppendorf MiniSpin centrifuge to produce the supernatant. Prior to GO deposition the silicon wafers were treated with aqueous potassium hydroxide (KOH) solution (weight concentration: 50%) for 15 min to make the surface more hydrophilic. Using this surface treatment the GO films were found to be more uniformly coated on the surfaces as compared to similar GO depositions on untreated surfaces. The improved coating is in part a result of the reduction of the contact angle of the solution deposited on the treated silicon wafer [17]. Control measurements using FFM conducted on untreated and KOH treated surfaces showed no significant difference in friction or adhesion properties. The silicon wafers were then cleaned with the same method of substrate cleaning described previously for the graphene sample preparation. To prepare all GO samples for FFM measurements GO solutions (8 μl) were drop cast onto silicon wafers with a micropipette and afterwards the solution was dried in air. 2.2. Film structure characterization and FFM
Thin films of graphene were identified via an optical microscope (Zeiss, Germany) [18]. Raman spectroscopy (Renishaw, 532 nm laser excitation) was performed to characterize the microstructure of the graphene and GO samples, and the surface morphologies of graphene and GO samples were observed by an atomic force microscope (AFM) (Asylum MFP3D) in tapping mode. XPS was also conducted on the GO samples to identify the specific chemical structure of the functional groups as well as measure the C/O ratio. The nanotribological properties of graphene and GO samples were measured using an AFM in FFM mode using rectangular cantilevers (Nanoworld FMR10, normal spring constant: ∼1.3 nN nm−1) in ambient conditions of 25 °C and 45% relative humidity. For each FFM measurement first the adhesion force between the AFM tip and film was measured by recording a force distance curve and measuring the adhesion force as the maximum force required to pull the AFM tip out of contact with the surface. The error of the adhesion force was calculated as the standard deviation of several (minimum of ten measurements) pull-off force measurements recorded on different regions of the surface in the near vicinity of the area used for friction measurements. Adhesion forces measured for different samples were measured under the same ambient condition to minimize any influence on changes in the meniscus [19].
2. Experimental section 2.1. Materials and sample preparation
Graphene and GO films were prepared on silicon (Si) substrates with a silicon dioxide (SiO2) surface layer. The silicon wafers (N-doped, 100 orientation, 285 nm SiO2 thickness) were first cleaned for 10 min in acetone followed by a 10 min 2
Nanotechnology 26 (2015) 135702
H Chen and T Filleter
3. Results
Quantitative FFM was used to measure lateral force signals simultaneously over surface regions containing two surfaces within the same scan line (either SiO2 and graphene, or SiO2 and GO) to make a direct comparison of lateral force differences between the two surfaces. As the AFM tip was scanned laterally across a surface, lateral force signals in both the forward scan and backward scan directions were recorded while sliding. The friction force was then calculated as half the difference between the forward and backward lateral force. The normal and lateral forces acting on the tip are given by [20]: FN = k N SN Va − b,
FL =
1 k T SL Vc− d , h2
3.1. Structure of graphene and GO films
The thickness of graphene was characterized using AFM in tapping mode (see figure 1(c) and line profile in figure 1(e)). From the AFM topography image in figures 1(c) and (e) for example, we observed a multilayer graphene film with a thickness of approximately 2.9 nm. A similar approach was used to observe, and measure the thickness of GO films (see figure 1(d) and line profile in figure 1(f)). The presence of graphene and GO were then verified using Raman spectroscopy [22, 23] as shown in figure 2. The Raman spectrum of graphene indicates the G peak located at 1582 cm−1 and 2D peak at 2675 cm−1, which are signatures of the in-plane vibration and second-order zone boundary phonons respectively. The absence of D peak in the Raman spectrum of graphene indicates that the exfoliated graphene in this study has a low defect density. The Raman spectrum of the GO shows a prominent peak at 1347 cm−1 with intensity comparable to the G peak at 1599 cm−1, which is characteristic of GO, and is indicative of significant structural disorder due to the defects in the GO plane. The thickness of single layer GO was found to vary from 0.6 to 1.2 nm, which is in agreement with literature values, and can be understood by its hygroscopic nature [24]. Thicker GO films were also observed. For example, the GO flake in figure 1(d) has thickness of ∼3.3 nm and thus it has a layer thickness of approximately three to four layers. XPS survey data indicated that the low C/O ratio GO studied has a C/O ratio of ∼2 while the high C/O ratio GO has a ratio of ∼3.7. Differences between the structures of the two samples were investigated with high resolution XPS measurement as shown in figure 3, which clearly indicate different concentrations of oxygen functional groups in the two samples studied. In addition to the chemical states associated to the oxygen functional groups, peaks for both sp3 and sp2 bonding of carbon were also included for XPS fitting.
(1)
(2)
where Va-b and Vc-d are the normal and lateral deflection signals from the four quadrant photodetector recorded during FFM scanning, kN and kT are normal and torsional spring constant respectively, Sn and Sl are the normal and lateral deflection sensitivity of the photodetector respectively, and h is the tip height. To quantify lateral force, the spring constant of the cantilever and the deflection sensitivity of photodetector were first measured. Normal and torsional spring constants were measured using the Sader method [21]. Both the normal and torsional spring constant require the resonant frequency and quality factor, which were measured from the thermal noise spectra of the normal and lateral deflection signals. The deflection sensitivity of the photodetector was determined from the slope of the cantilever deflection versus distance curve. The lateral deflection sensitivity of the photodetector, similarly, is determined from lateral deflection signal as a function of lateral displacement. Measurement of lateral deflection sensitivity is in general more difficult than the normal deflection sensitivity due to the higher lateral stiffness of the cantilever and possible in-plane bending in addition to the torsional response. The test probe method was used under ambient conditions to measure the lateral deflection sensitivity which avoids such issues, and has been shown to yield accurate measurements of the deflection sensitivity [20]. For this method a 70 μm diameter colloidal glass sphere was attached to a silicon cantilever (the same type of cantilever used for FFM) by epoxy and then the lateral signal versus distance was obtained via pushing the colloidal sphere against a freshly cleaved side wall (100) of a KBr crystal. The slope of the contact region in the resulting force plot gives the lateral deflection sensitivity (SL,test). The lateral deflection sensitivity of the photo detector with the sharp integrated tips used for FFM measurements (SL) was then calculated from SL,test considering the differences of the torsional arm length and lateral in-plane bending parameters between the colloid probe and sharp integrated tip. The resulted SL was then used for the calculation of lateral force from the torsional signals.
3.2. Friction and wear of graphene on SiO2/Si
Figure 4 shows the height topography and lateral force map of graphene on a SiO2/Si surface, and figure 5 shows friction versus normal force curves. Friction on multilayer graphene (4 layers) was found to be much lower than that of SiO2, as shown in figure 5, where friction measured on graphene was ∼6 times lower than on SiO2 under a normal load condition of ∼200 nN. Root mean square (rms) roughness of a 1.5 nm layer of multilayer graphene film and the SiO2 substrate was measured to be 0.39 ± 0.07 and 0.22 ± 0.03 nm respectively. The graphene has a larger rms roughness than SiO2, which suggests that the lower friction of graphene as compared to SiO2 is not due to its roughness. Despite the large friction difference, the force of adhesion (pull-off force between tip and sample) of graphene and SiO2/Si was found to be 25 ± 1 and 27.1 ± 0.9 nN respectively, suggesting that the friction difference is an intrinsic difference in the energy difference due to sliding. This is in agreement with the observation of a 3
Nanotechnology 26 (2015) 135702
H Chen and T Filleter
Figure 1. (a) Optical image of exfoliated graphene on a SiO2/Si substrate (white arrow identifies the graphene film); (b) GO aqua-solution in centrifuge tube after centrifuge (precipitates are below the red dashed line and black arrow identifies the precipitates); (c) AFM tapping mode image of multilayer graphene film on a SiO2/Si substrate and corresponding line profile (e) along the black dashed line showing a multilayer graphene flake on SiO2/Si substrate; (d) AFM tapping mode image of GO film on SiO2/Si substrate and corresponding line profile (f) showing a multilayer graphene oxide flake on SiO2/Si substrate.
modulus and Poisson’s ratio of 70 GPa and 0.17 [26] were used in the model for silicon dioxide (both silicon dioxide substrate and silicon dioxide tip) and graphene samples. Given the ultrathin nature of the graphene films we have assumed that the modulus of the film/substrate surface is dominated by the underlying thick silicon dioxide substrate [27]. Alternatively the modulus of graphite (∼30 GPa) could have been used to estimate the multilayered graphene modulus loaded out-of plane. The tip radius was assumed to be 15 nm in this model, it should be noted that while this is an estimate of the tip radius, the same tip was used for
change in the slope of the friction versus normal force curves in figure 5. In order to estimate the interfacial shear strength between the two contacts under study, the friction versus normal load curves were fit using the generalized Maugis–Dugdale model [25], which allows determination of an interface’s contact behavior that is intermediate between the Derjaguin−Mueller −Toporov (DMT) assumption of a ‘hard’ contact and the Johnson−Kendall−Roberts (JKR) assumption of a ‘soft’ contact. The generalized Maugis–Dugdale model is referred to as ‘DMT–JKR transition model’ in this study. A Young’s 4
Nanotechnology 26 (2015) 135702
H Chen and T Filleter
1000
(b)
2D
lntensity (a.u.)
lntensity (a.u.)
(a)
G
2000
Raman shift
(cm-1)
1000
3000
D
G
2D
2000
Raman shift
(cm-1)
3000
Figure 2. (a) Raman spectrum of exfoliated graphene on a SiO2/Si substrate: G peak at 1582 cm−1 and 2D peak at 2675 cm−1; (b) Raman
spectrum of GO films on a SiO2/Si substrate: D peak at 1347 cm−1 and G peak at 1599 cm−1 (the Raman spectra of graphene and GO were both taken at a laser wavelength of 532 nm).
Figure 3. XPS measurements of (a) low C/O ratio (∼2) and (b) high C/O ratio (∼3.7) graphene oxide samples.
simultaneous measurements on the graphene and silicon dioxide surfaces. From the fits the interfacial shear strength between the silicon tip and graphene, and the silicon tip and silicon dioxide surfaces was found to be 173 ± 13 MPa (∼125 MPa if the modulus of graphite is instead used) and 1800 ± 44 MPa respectively. This measured interfacial shear strength between the tip and the silicon dioxide surface is on the same order of magnitude as compared to that reported in previous literature with a similar contact system [27]. This again confirms that the difference in friction behavior is due to different interfacial shear strength and not any significant difference in adhesion. The observed friction difference between graphene and silicon dioxide is also in good agreement with previous reported results in the literature [27] where friction forces were found to decrease by ∼90% on supported monolayer graphene relative to SiO2. No wear was observed on the graphene surface after the friction measurements shown in figure 5, conducted at up to normal forces of ∼200 nN.
3.3. Friction and wear of low C/O ratio GO on SiO2/Si
Figures 6(a) and (b) shows the height topography of a 1–2 layer GO film (∼1.2 nm) with a C/O ratio of ∼2 on a SiO2/Si surface. Figure 6(c) shows the lateral force of the forward and backward scan along the black dashed line in figure 6(a) at a normal load of −13.8 nN. From figure 6(c) it can be seen that friction on the SiO2/Si surface is approximately 2.3 times larger than on the 1–2 layer GO film at a normal force of −13.8 nN. This demonstrates that, as is also true for graphene films, just one or two atomic layers of GO is effective at reducing friction on silicon dioxide surfaces that are often used for MEMS. Figure 6(d) shows a zoom out AFM topography image of the FFM scan area, which indicates that some of the GO area was worn (e.g. the red dashed rectangle area in figure 6(d)) while other areas are still effectively lubricating the SiO2 surface. In addition, worn GO particles were found to be piled up at the edge of the scan area after the FFM scan as shown in figure 6(d). In order to correlate this wear behavior with friction force, friction versus normal load 5
Nanotechnology 26 (2015) 135702
H Chen and T Filleter
Figure 4. (a)AFM height topography (contact mode) image and corresponding line profile (c) along the while dashed line showing a 4 layer graphene film on a SiO2/Si substrate; (b) lateral force map of forward scan and corresponding lateral force line profile (d) of both forward and backward scan along the same line.
higher than ∼0 nN. The onset of wear can be identified by the discontinuous jump in the friction versus normal load curve [28]. From the different friction and wear behavior on these areas it is clear that some areas of the GO flake are able to resist wear at relatively higher normal load. This behavior can be attributed to the heterogeneous nature of the GO surface [29] which will be discussed in detail later in the context of the proposed wear mechanism for thin GO films. A similar transition in the friction versus normal load curves that identify the onset of wear was also observed for multilayer GO films. Figures 7(a) and (b) shows the height topography and lateral force map of a multilayer GO film (∼3 nm) on a SiO2/Si surface. From figure 7 the friction of SiO2/Si is found to be ∼1.9 times higher than that of multilayer GO at a normal force of −3.6 nN. Adhesion between the silicon tip/silicon dioxide and the silicon tip/GO were measured to be 4.4 ± 0.3 and 4.1 ± 0.2 nN respectively. It should be noted that the rms roughness of the GO film covered areas was measured using AFM to be the same as the uncoated SiO2 areas within measurement error, which rules out the possibility of differences in rms roughness as the dominant mechanism for the observed friction difference between GO and SiO2. From figure 7(c) a discontinuity in the friction force was observed when the normal force was higher than ∼9 nN. From figures 8(a)–(c), it can be observed that the GO flake became smaller and smaller with increasing normal load during the FFM scan. This onset of wear was found to initiate
Figure 5. Friction force versus normal force of silicon dioxide and
graphene (4 layers) on a SiO2/Si substrate. The solid black and red lines are fits using the DMT–JKR transition model.
curves for FFM scans on GO (both on the worn area and unworn area) and the SiO2/Si surface were plotted in figure 6(e) to give a direct comparison between the three surface areas. From figure 6(e), there is an indication that the GO films started to wear when the normal force reached 6
Nanotechnology 26 (2015) 135702
H Chen and T Filleter
Figure 6. (a)AFM height topography image (contact mode) under normal force of −13.8 nN; (b) line profile along the black dashed line in the topography (a) showing a ∼1.2 nm GO film on a SiO2/Si substrate; (c) lateral force of both forward and backward scan along the same black dashed line in (a); (d) zoom out image (tapping mode) of the FFM scan area (black dashed rectangle area corresponds to the area of image (a) scanned in FFM mode) and the red dashed rectangle indicates the worn area of GO during FFM scanning; (e) friction force versus normal force of SiO2 surface area, worn GO surface area, and unworn GO surface area.
at a normal load of ∼10 nN which is consistent with the discontinuity observed in the friction force plot. GO appears to be suddenly displaced in the lateral direction resulting in a jump in friction force and subsequently removed gradually from the substrate, potentially arising from plastic deformation. Repetition of the experiment with another GO flake under the same conditions exhibited worn GO particles accumulated at the edge of the scan area. Similar to the transition fit used for friction load curves of graphene data, the friction versus normal load curves of GO (before wear occurred) were selected and fit using the DMT– JKR transition model. The Young’s modulus and Poisson’s ratio of 70 GPa and 0.17 [26] were used in the model for silicon dioxide (both silicon dioxide substrate and silicon dioxide tip) and GO sample. The tip radius was also assumed to be 15 nm in the transition model. The interfacial shear strength between silicon tip and GO, and the silicon tip and silicon dioxide surface was found to be 1200 ± 180 and 1930 ± 282 MPa respectively.
3.4. Wear resistance of high C/O ratio GO films on SiO2/Si
In the previous section low C/O ratio (∼2) GO was found to exhibit lower friction as compared to silicon dioxide surfaces, however it was also found to be easily worn under a few nN normal load. Graphene, however, was found to resist wear under normal loads of up to several hundred nN. As was discussed earlier, reduced GO (which exhibits a higher C/O ratio than unreduced GO) exhibits lower friction than GO [12, 13]. The effect of increasing the C/O ratio of GO may also play a role in improving the wear behavior of GO at the nanoscale as it yields a more graphene-like surface structure. FFM combined with XPS was used to characterize the effect of chemical composition on the wear properties of the high C/O ratio GO samples. Figure 9(a) shows a tapping mode AFM image of a region of a high C/O ratio (∼3.7) GO film (∼1.1 nm thickness) on a SiO2 substrate. The region in the white area in figure 9(a) was scanned with normal loads ranging from 0 to 270 nN. From the topography image of figures 9(b)–(d), it was observed that the GO was wear 7
Nanotechnology 26 (2015) 135702
H Chen and T Filleter
Figure 7. (a) AFM contact mode topography image of multilayer GO a SiO2/Si substrate and (c) line profile along the black dashed line in the topography image; (b) lateral force map and (d) line profile of lateral force signal along the same line as shown in (d); (e) friction force versus normal force plot.
4. Discussion
resistant up to a normal load of up to 270 nN. This behavior was consistently observed for several similar high C/O ratio samples. This high wear resistance is quite different from the easy tendency of wear of the low C/O ratio GO samples which always began to wear at normal loads of just a few nN. This correlation between the high C/O ratio structure of the GO films and wear resistance properties similar to that of the graphene films demonstrates the clear advantages of the tailorability of the GO surface structures.
4.1. Friction
If we compare the results found on GO to graphene, we find that friction of 1–2 layer GO is approximately 3 times higher. Considering that the relative adhesion for GO and graphene and the substrate is similar, this suggests a higher interfacial shear strength between GO and silicon dioxide than that
8
Nanotechnology 26 (2015) 135702
H Chen and T Filleter
Figure 8. AFM contact mode topography images of GO on a SiO2 surface at normal force of −3.6 nN (a), 10.8 nN (b) and 21.6 nN (c) respectively.
between graphene and silicon dioxide. The different interfacial shear strength is likely due to the presence of the dangling functional groups and higher defect density in the 2D GO plane as revealed by high resolution TEM studies [29]. This is consistent with molecular dynamics simulation which have predicted that functional groups, such as epoxide and hydroxl groups, on the GO plane increase the shear strenght as compared to a pristine graphene structure [30]. The friction difference between GO and SiO2 reported here (∼1.9–2.3) is slightly lower as compared to that previously reported by Ou et al (∼2–3 times) [13]. This difference may be attributed to the fact that the silicon wafer studied by Ou et al were treated with pirahna which may increase the roughness of the silicon surface resulting in higher friction in the absence of the GO film. In addition, previous studies of thick GO films (∼50–400 nm) by Liang et al [31] exhibit the same general finding of reduced friction on the GO surfaces as compared to SiO2 as reported here where they observed a friction reduction of up to ∼6 times for the GO coated surfaces, however, it is difficult to quantitativly compare the two studies as they utlize very different contact sizes, GO sample structures and compositions, and loading conditions. One major difference between the findings reported here and those reported by Liang et al stems from the size scale of the contacts under study. Liang et al found that for macroscopic sliding contacts the thick GO films exhibited an increasing trend between film thickness and friciton coefficient for their micro-tribometer experiments of thick GO films on SiO2 that was attributed to differences in film roughness [31]. Here we have found instead that when the sliding contact is on the nanoscale there is no significant friction dependance on thickness. This is a very important observation as it suggests that the previously reported thickness dependance by Liang et al which is a result of varrying surface roughness, is a macroscale phenomena which emerges for multi asperity contacts and is not exhibited for nanoscale contacts.
4.2. Wear
High C/O ratio GO was found to exhibit a high wear resistance as compared to the easy tendency of wear of the low C/ O ratio GO samples. This correlation between the high C/O ratio structure of the GO films and wear resistance properties similar to that of the graphene films demonstrates the clear advantages of the tailorability of the GO surface structures. The difference in wear behavior for graphene, high C/O ratio GO, and low C/O ratio GO can be understood based on the varying surface structures of the three specimen. In the case of the low C/O ratio GO specimens the first stages of wear are likely to initiate at lower normal loads than for graphene due to differences in the local shear strength at different locations on the GO surface with different functional groups [30] and at defects in the GO plane which act as nucleation sites for wear to initiate. Wang et al have shown through DFT calculations that epoxide and hydroxyl functional groups that exist in GO lead to a local higher energy corrugation and shear strength for sliding between small flakes of GO as compared to graphene [30]. Microscopic GO films (in the lateral dimension), such as those that are under investigation here, have a different structure than the very small flakes studied in the DFT studies. Larger experimental GO flakes are instead composed of a mixture of non-homogenous regions of pristine graphene-like and functionalized GO-like patches. This has been demonstrated recently through high resolution TEM imaging of GO surfaces [29]. Therefore we expect that there will be local differences in the shear strength between the sliding tip and the GO surface at these different surface regions. XPS experiments conducted on the low C/O ratio GO samples studied here show the presence of ∼35 atomic% of epoxide, ∼15 atomic% of carbonyl, and ∼9 atomic% of carboxyl groups that make up these functionalized regions of the surface. We believe it is at these local regions on the GO surface with higher shear strength that wear initiates when the normal force reaches a critical level and the in plane stress is such that the fracture strength of the GO films is reached. Wear then progresses in the form of the local fracture and removal of small regions of 9
Nanotechnology 26 (2015) 135702
H Chen and T Filleter
Figure 9. (a) Tapping mode AFM image of high C/O ratio GO on a SiO2 substrate. (b), (c) and (d) are topography images for scanning under normal loads of 0, 60 and 270 nN respectively from FFM scans in the white dashed rectangular area in (a).
the GO films. In addition, the non-homogenous nature of the GO surface can explain why some regions wear at lower normal loads than others. This is significantly different behavior than that observed for graphene and high C/O ratio films which exhibited wear free sliding for normal forces of ∼200 nN, almost two orders of magnitude higher. High C/O ratio GO has a significantly reduced density of functional groups on its surface and therefore exhibits a more graphenelike surface structure. This points to a major limitation in using GO with a high density of functional groups as a solid lubricant under larger applied stress, as it contains many surface regions that tend to wear more easily than graphene.
friction of 1–2 layer and few-layer (3–4 layers) GO were found to range between 1.9–2.3 times lower than that of SiO2/ Si under similar normal loading conditions with no significant dependence on film thickness, suggesting that GO film thickness dose not play a primary role for nanoscale sliding contacts. A major difference in the wear behavior was observed between graphene and GO films. Graphene was found to resist wear with normal loads of up to ∼200 nN. In contrast, atomic thin GO sheet with C/O ratios of ∼2 tend to wear easily under normal loads of a few nN with an onset of wear that is attributed to higher local shear strengths at the nonhomogeneous surface regions on the GO films as compared to graphene. The wear behavior of GO was also found to be spatially dependant on the non-homogeneous surface regions. This limitation can be overcome by modifying the structure of the GO films. GO films with higher C/O ratios, which exhibit structures more similar to pristine graphene, were found to be wear resistant up to normal loads of ∼270 nN. These findings indicate the great importance of the local surface structures of GO films on their tribological behavior, and motivate additional future studies which focus on how variations in this
5. Summary and conclusion The tribological properties of SiO2/Si, graphene and GO with different C/O ratios were investigated using quantitative FFM with nanoscale sliding contacts. Friction of few-layer graphene was found to be approximately 6 times lower than that of SiO2/Si, which is attributed to lower interfacial shear strength between graphene and silicon dioxide. In contrast 10
Nanotechnology 26 (2015) 135702
H Chen and T Filleter
surface structure can lead to tunable approaches for lubrication and wear resistant coatings. While the lower friction of GO as compared to SiO2/Si makes it a potential economical coating for M/NEMS devices to extend the lifetime of devices and reduce energy dissipation, careful considerations in controlling its structure must be taken to limit its higher propensity for wear in the case of low C/O ratio structures as compared to graphene when applying it as a solid lubricant.
[13] Ou J et al 2010 Tribology study of reduced graphene oxide sheets on silicon substrate synthesized via covalent assembly Langmuir 26 15830–6 [14] Novoselov K S et al 2005 Two-dimensional atomic crystals Proc. Natl Acad. Sci. USA 102 10451–3 [15] Hummers W S and Offeman R E 1958 Preparation of graphitic oxide J. Am. Chem. Soc. 80 1339–1339 [16] Kovtyukhova N I et al 1999 Layer-by-layer assembly of ultrathin composite films from micron-sized graphite oxide sheets and polycations Chem. Mater. 11 771–8 [17] Franssila S 2010 Wafer cleaning and surface preparation Introduction to Microfabrication (Chichester: Wiley) chapter 12 pp 143–52 [18] Blake P et al 2007 Making graphene visible Appl. Phys. Lett. 91 063124-3 [19] Bhushan B and Dandavate C 2000 Thin-film friction and adhesion studies using atomic force microscopy J. Appl. Phys. 87 1201–10 [20] Cannara R J, Eglin M and Carpick R W 2006 Lateral force calibration in atomic force microscopy: a new lateral force calibration method and general guidelines for optimization Rev. Sci. Instrum. 77 053701–11 [21] Sader J E, Chon J W M and Mulvaney P 1999 Calibration of rectangular atomic force microscope cantilevers Rev. Sci. Instrum. 70 3967–9 [22] Ferrari A C et al 2006 Raman spectrum of graphene and graphene layers Phys. Rev. Lett. 97 187401 [23] Kudin K N, Ozbas B, Schniepp H C, Prud’homme R K, Aksay I A and Car R 2007 Raman spectra of graphite oxide and functionalized graphene sheets Nano Lett. 8 36–41 [24] Stankovich S et al 2007 Synthesis of graphene-based nanosheets via chemical reduction of exfoliated graphite oxide Carbon 45 1558–65 [25] Carpick R W, Ogletree D F and Salmeron M 1999 A general equation for fitting contact area and friction vs load measurements J. Colloid Interface Sci. 211 395–400 [26] Kim M T 1996 Influence of substrates on the elastic reaction of films for the microindentation tests Thin Solid Films 283 12–6 [27] Deng Z, Klimov N N, Solares S D, Li T, Xu H and Cannara R J 2012 Nanoscale interfacial friction and adhesion on supported versus suspended monolayer and multilayer graphene Langmuir 29 235–43 [28] Steiner P et al 2010 Atomic friction investigations on ordered superstructures Tribol. Lett. 39 321–7 [29] Erickson K, Erni R, Lee Z, Alem N, Gannett W and Zettl A 2010 Determination of the local chemical structure of graphene oxide and reduced graphene oxide Adv. Mater. 22 4467–72 [30] Wang L-F, Ma T-B, Hu Y-Z and Wang H 2012 Atomic-scale friction in graphene oxide: an interfacial interaction perspective from first-principles calculations Phys. Rev. B 86 125436 [31] Liang H, Bu Y, Zhang J, Cao Z and Liang A 2013 Graphene oxide film as solid lubricant ACS Appl. Mater. Interfaces 5 6369–75
Acknowledgments This work was supported by the Canada Foundation for Innovation and NSERC. The authors would like to thank Changhong Cao for assistance in conducting Raman spectroscopy experiments. The authors would also like to thank Rana Sodhi and Peter Broderson for conducting XPS data collection and analysis of the GO samples.
References [1] Chandross M, Lorenz C, Grest G, Stevens M and Webb E III 2005 Nanotribology of anti-friction coatings in MEMS JOM 57 55–61 [2] Kim S H, Asay D B and Dugger M T 2007 Nanotribology and MEMS Nano Today 2 22–9 [3] Castro Neto A H, Guinea F, Peres N M R, Novoselov K S and Geim A K 2009 The electronic properties of graphene Rev. Mod. Phys. 81 109–62 [4] Lee C, Wei X, Kysar J W and Hone J 2008 Measurement of the elastic properties and intrinsic strength of monolayer graphene Science 321 385–8 [5] Lee G-H et al 2013 High-strength chemical-vapor–deposited graphene and grain boundaries Science 340 1073–6 [6] Filleter T et al 2009 Friction and dissipation in epitaxial graphene films Phys. Rev. Lett. 102 086102 [7] Lee C et al 2010 Frictional characteristics of atomically thin sheets Science 328 76–80 [8] Koenig S P, Boddeti N G, Dunn M L and Bunch J S 2011 Ultrastrong adhesion of graphene membranes Nat. Nanotechnology 6 543–6 [9] Graphene Supermarket 2015 http://graphene-supermarket.com [10] Ou J, Wang Y, Wang J, Liu S, Li Z and Yang S 2011 Selfassembly of octadecyltrichlorosilane on graphene oxide and the tribological performances of the resultant film J. Phys. Chem. C 115 10080–6 [11] Compton O C, Jain B, Dikin D A, Abouimrane A, Amine K and Nguyen S T 2011 Chemically active reduced graphene oxide with tunable C/O ratios ACS Nano 5 4380–91 [12] Wei Z et al 2010 Nanoscale tunable reduction of graphene oxide for graphene electronics Science 328 1373–6
11
