Simultaneous chemical reduction and surface functionalization of graphene oxide for efficient lubrication of steel–steel contact

Abstract

An improved method for the preparation of graphene oxide (GO) and further exfoliation of GO to graphene is described. Simultaneous surface functionalization and reduction of GO with octadecylamine (ODA) was also achieved without any further reducing agents. The structural aspects of GO, GO-ODA and graphene were analyzed by UV-Vis spectrometry, infra red spectroscopy, Raman spectroscopy and transmission electron microscopy. The wetting behavior of the functionalized particles were accessed by water contact angle measurements. The particle sizes in the oil suspensions which is actually fed to contact for lubrication were estimated using dynamic light scattering technique. Tribological performances and load bearing capacity of the particle suspensions were tested under two different pressure regime on a rotational tribometer with pin/ball-on-disc set up. The increase in hydrophobicity in GO-ODA helped the particles to work much better in low load condition due to the presence of long hydrocarbon chain as a shield between the contact. However the hydrophilic GO was found to perform better in the high load regime due to its adhesion to the substrate when squeezed under high pressure. The work focuses on lubrication mechanism of chemically functionalized graphite particles as the load bearing candidates for efficient lubrication of sliding contact.

---

1. Introduction

Friction and wear are the two major causes for energy and material loss in meso-micro mechanical systems. Lubrication is essential for mechanical assemblies for improved energy efficiency and mechanical durability. The fluid lubricants used in automotive and industrial lubrication perform well as a lubricant in the hydrodynamic regime where the viscosity of the liquid controls friction. The load bearing capacity of most liquids is limited and their performance is impaired in the boundary regime where direct contact occurs between mating parts promoting high friction and wear. Organic boundary lubricant additives may be used to functionalize the mating parts to prevent direct contact. The traditional lubricant additives employed in automotive and industrial lubrication start off many problems of toxicity, gas emission and degradation of additives which lead to environmental problems and loss of good tribological properties. Initiatives are taken for to find substitutes for these additives and develop a new generation of lubricants based on the development of additives for lubrication using solid nanoparticles. Nanoparticles have gained an increasing interest as extreme pressure and anti-wear additive for liquid lubricants due to their high thermal stability and good load bearing property. Several authors investigated the tribological effect of nanoparticles suspended in organic liquids such as oil, grease and aqueous medium.^1–6

Layered materials such as graphite, MoS₂ and WS₂ (platelets of the 2H polytype) which known to have excellent lubricating properties are widely used either as solid lubricant for space applications or as additives dispersed in a lubricating base.^7–10 These materials generally have a layered structure, which shear easily under traction to yield low frictional forces. The low friction of both graphite and metal dichalcogenides is usually due to interplanar mechanical weakness, intrinsic to their crystal structures.¹¹ The weak inter-lamellar bonding facilitates the shear when the direction of sliding is parallel to the planes of the material. Under the action of a shear force, intracrystalline slip occurs in the weak interplanar regions. This mechanism is responsible for the formation of smooth transfer films by wear. When solid nanoparticles are used as lubricant additives the key mechanism of boundary lubrication is clearly different from that when organic surfactants are used as additives. The nanoparticles, suspended in the fluid medium get carried to the region of contact between the asperities of contacting surfaces.¹ The solid particles adhere to the asperities and put up a transfer film on the contacting surface^12,13 while the elastic or plastic strength of the solid particle supports the normal load. A major problems for the above hypothesis is the agglomeration of the particles and transport of these particles to the region of contact. The small particles because of their large surface area to volume ratio are easily attracted to each other to form agglomerates. Such agglomerates are not easily accommodated in the contact zone and the beneficial effect of the layered structure is lost. Therefore functionalization of the particles for better control of particle dispersion and transport is a key challenge. Colloidal stabilization of the nanoparticles using surfactants and chemical functionalization of nanoparticles are the two important methods for the stabilization of nanoparticles.

Graphene following Novoselov and Geim's discovery,¹⁴ has triggered a tremendous amount of scientific interest mainly for its unique electronic, physico-chemical and mechanical properties^15–17 and has received considerable interest for its potential application in many technological fields such as nanoelectronics, sensors, composites, batteries, supercapacitors, lubrication.^6,18–23 Graphene oxide (GO), a nonconductive hydrophilic carbon material²⁴ mainly consists of aromatic lattice of graphene is interrupted by epoxides, hydroxyl, and carboxylic groups is obtained by exfoliation of graphite by strong oxidizing agents. The graphitic structure can be brought back by chemical reduction²⁵ to chemically converted relatively hydrophobic graphene also known as reduced graphene oxide (RGO). The simple and easy synthesis procedure makes GO a promising precursor for graphene and its derivatives.

In recent years, chemical modification of graphene oxide (GO) sheets has become a popular research subject as presence of reactive oxygen functional groups allows GO to undergo chemical and structural modifications.^26,27 Moreover the presence of carbon oxygen functional groups allows its dissolution in water without further hydrophilic treatments. However in non-aqueous medium its dissolution is somehow poor due to its hydrophilic nature. However the RGO can also be obtained by simultaneous surface functionalization and reduction of GO with octadecylamine (ODA) without the use of any reducing agents. Chemicals with long alkyl chains, such as ODA, have been used to make hydrophilic GO to hydrophobic RGO.^28–30 It is noted that the dispersibility of alkylated graphene in non-polar hydrocarbon solvents increases on increasing the chain length of the alkyl group attached to the GO.³¹ Graphene as well as graphene oxide and ODA-functionalized graphene oxide are the chemically functionalized building block of the most common macroscopic solid lubricant graphite which exhibits low friction and wear rates. Therefore besides its excellent electronic and mechanical properties, the lubricating property of graphene, graphene oxide and ODA-functionalized graphene oxide is also of interest for application in future.

The methods of application of functionalized graphitic particles as a lubricating medium are limited^32–36 and no attempt has been made to discuss them in details in the fundamental level. It has been used as an additive agent in both oils and water medium. It can even be used as a surface coating in the form of assembled monolayer for MEMS lubrication. Even though few reports were already present on liquid/surface-based lubrication, they suffer from low efficiency. Hence, it is always essential to tune the efficiency of lubricants by modifying their chemical and structural properties. The purpose of the present work is to come to an understanding the performance of these lubricant particles by optimizing the materials and process requirements. One of the main issue involved in the use of nanoparticles in liquid suspensions, as a lubricant, is the ability of these particles to migrate into the region of contact. The size of the particles, their adhesion to the mating solid surface, the viscosity of the fluid, the normal load, and sliding velocity are the factors, which may be expected to play roles in the transport of these particles to the region of contact, hence their lubricity. While in contact if these particles roll or shear easily the expected frictional forces are low. Once sheared there may be a process of transfer film formation on both the counterfaces and sliding takes place between films transferred to the substrate and the counterface. The deposit of particles acts as a third body reducing asperity interactions and thus increasing load-bearing capacity of the rubbing pairs.

The particle size distribution in the suspension plays an important role in controlling the friction. While very big particles may have problems getting into the narrow confiners of the contact, the small particles can infiltrate to the tiniest spaces between contacting surfaces and also have high chemical and physical stability even under extreme conditions and increased load-bearing capacity. We have chosen in this work graphene oxide and their reduced counterparts, RGO with ODA (GO-ODA) and ultrathin graphene as lubricant additive in heavy paraffin oil to improve the frictional and antiwear characteristics and load bearing capacity of oil. The graphitic functionalized particles were synthesized and their tribological properties and friction reduction mechanisms were established. The performances of the nanoparticle in the lubricating fluid was assessed in a rotational tribometer with a pin-on-disc and ball-on-disc set up under various pressure regime. We investigate in this paper the deformation and distribution of these functionalized particles in oil suspension and show how such functionalization of the particles, when used as lubricants in steel on steel sliding contact, control the friction process with respect to contact pressures.

2. Experimental

2.1 Materials and methods

2.1.1 Materials. Natural flake graphite (60 mesh) was purchased from Loba Chemie, India. Hydrochloric acid (37%, AR grade, Merck), sulfuric acid (98%, SD Fine Chem, India), potassium permanganate (AR grade, Merck) and hydrogen peroxide (30%, AR grade, Merck) were purchased and used as oxidation of graphitic skeleton. Hydrazine monohydrate (99% AR grade, Merck) was used as reducing agent for the preparation of graphene. Octadecylamine (90%, Technical grade) was purchased from SRL chemicals India and used for the simultaneous reducing and surface modifying agent without further purification.

Hardened 440C stainless steel (0.15% carbon, 1% manganese, 0.04% phosphorus, 0.03% sulphur, 1% silicon, 11.5–13.5% chromium, 50 HRC) is used as disc and pin materials in the tribological experiments. The steel disc samples were emery polished using silicon carbide papers. After emery polish the samples were rinsed in acetone (AR Grade, Sd Fine Chem, India) and then disc polished in BAINPOL (Chennai Metco Pvt. Limited) double disc machine using diamond paste of grade 1–3 μm. The polished samples were cleaned in acetone to remove all polishing debris.

2.1.2 Preparation of functionalized nanolubricants. GO was synthesized from natural graphite flakes by following a modified Hummers method.^37–40 In brief, 3 g of natural graphite flake was dispersed in 69 ml of concentrated sulfuric acid in a round bottom flask (RB) placed in an ice bath at 0–5 °C. Then potassium permanganate about 9 g was slowly added for 15–20 min at 0–5 °C and stirring was continued for 2 h. The reaction mixture was subsequently transferred to a pre-heated oil bath at 35 °C and stirred over night. De-ionized water (HPLC grade, Merck) about 138 ml was carefully added to the reaction mixture and stirring was continued for another 2 h. 35% hydrogen peroxide solution was added drop wise to the reaction mixture until the color of the solution became bright yellow. The excess of manganese salt was removed by dilute hydrochloric acid solution (5% by volume) by adding carefully to the resultant solution. The change in the color of the suspension from black to brown corroborated the formation of graphite oxide. The resulting suspension was filtered and rinsed several times with DI water to completely remove the residual salts and acids. Then the resulting solution was dried in a vacuum oven for 5 days at 45 °C. The purified powder was finally dispersed in water (1 mg ml⁻¹) and ultrasonically exfoliated in an ultrasonic bath. The resultant graphene oxide (GO) dispersion was found to be stable for a very long time.

The preparation of ODA functionalized GO was performed by the following described methods.^29,30,41 The dispersion and exfoliation of GO (0.9 g) in 450 ml de-ionized water was carried out ultrasonically for about half an hour. The resulting GO suspension was mixed with a solution of ODA (1.3 g) in 130 ml ethanol in a three-neck flask. The refluxing process was carried out for 20 h at 90 °C, with continuous mechanical stirring. The resulting black suspension was filtered using nylon membrane filter paper (Riviera, pore size 0.02 μm) and rinsed with ethanol several times to remove the physically adsorbed ODA. The resultant black powder was dried under vacuum for 2 days at 60 °C and stored in a dessicator for further use.

An improved method for the preparation of graphene from GO is described.^42–45 Briefly, GO was dispersed in DI water via ultrasonication to create a 1 mg ml⁻¹ GO dispersion. The resulting suspension (90 ml) was mixed with 85% hydrazine monohydrate (60 ml) in a round bottom flask. The mixture was refluxed at 100 °C for 12 h with continuous stirring. The resultant mixture was filtered and rinsed several times with water and the consequent black powder was dried at 50 °C in a vacuum oven.

The particle suspensions in heavy paraffin oil (∼C₂₀–C₂₅, 64 cSt at 37 °C, AR grade, Merck India) medium ware prepared by suspending 10 mg of nanoparticles in 10 ml of oil (0.1% w/v) using an ultrasonic probe (VCX-500, 20 kHz frequency, Sonics and Materials Inc. USA) for the particle size analysis. For the tribological analysis the nanosuspensions were prepared by suspending 100 mg of nanoparticles in 10 ml of oil (1% w/v).

2.2 Characterization techniques

The FTIR spectra of the functionalized graphitic particles were acquired by a Perkin-Elmer Spectrum 100 spectrometer. All IR spectra reported here were referenced to a bare rectangular KBr pellet and acquired over 100 scans at 4 cm⁻¹ resolution. The spectral analyses were carried out with Spectrum v10.00 software (Perkin-Elmer). UV-Vis absorption spectra of Go and graphene in water and GO-ODA in chloroform were recorded using a 1 cm path length quartz cuvette with water/chloroform as the reference on a SEC2000, ALS spectrophotometer. Resonance Raman spectra were recorded at room temperature using a standard backscattering geometry applied with Renishaw inVia reflex micro Raman spectrometer. Excitation wavelength of 785 nm was produced by a near IR diode laser source capable of supplying 300 mW of power. Spectrometer was calibrated by Si line at 520 cm⁻¹.

Morphology was determined on a JEM-3010 Transmission Electron Microscope (TEM). For TEM examination, the dried samples were ultrasonically dispersed in water/ethanol and cast onto TEM grids. Water contact angle measurements were performed using the sessile drop method for the motion of the three-phase contact line between water and particle surface. The measuring device was an OCA 15pro commercial goniometer (Dataphysics, Germany) equipped with a stepper motor for controlling the volume of the liquid supplied from a microsyringe and a CCD camera for image capture. The particle size distribution in oil suspensions were measured by Zetasizer nano ZS nanoparticle size analyzer (Malvern Instruments, UK) using Dynamic Light Scattering. The particle analyzer consists of a 4 mW He Ne laser for the scattering of the particles and the APD detector to measure the scattered light at a 90° scattering angle and gives the characteristic particle size distribution and charge on the particles. A cylindrical cell of 1 cm diameter is used for the purpose and 2 ml of suspension is required to make the measurements.

2.3 Tribological characterization

Macrotribological tests were carried out under two different pressure regime using a pin/ball-on disc machine, procured from DUCOM (Bangalore, India). The flat face of a high-speed steel pin (diameter 4 mm, RMS roughness ∼ 100 nm) was loaded normally and pressed against the flat surface of a rotating disc (RMS roughness ∼ 400 nm) to generate mean Hertzian pressures (P_m) of 0.9–3 MPa. Prior to the actual experiment a full pin-on-disc contact was established by running in for several hours at 5 N load. For ball on disc experiments a 10 mm diameter chromo steel ball (rms roughness ∼ 5 nm) is attached to the ball holder and pressed against the rotating disc to generate mean Hertzian pressures of 1.2–2 GPa. The disc was slid against the pin and ball at 0.4 m s⁻¹ surface speed. The friction force was measured by a load cell attached to the pin holder (resolution 0.1 N) and the displacement of the pin was measured using a Syscon (Bangalore, India) displacement sensor (LVDT, resolution 1 μm, range = ±500 μm). For the lubricated test, oil dispersed with the particles (1 wt%) was added at the sliding interface. The constant loading tests were conducted for 1 h. For continuous loading tribological tests, the load was gradually increased by 10 N and tests were carried out for 30 min at each load. Further to explore the status of the particle in the slid track of pin-on-disc tests, optical images of the slid tracks were obtained using Nikon optical microscope equipped with a Micropublisher 3.3 RTV CCD camera (Q Imaging, Canada). Microscopic features of the tribo tracks were obtained using a Sigma (Zeiss) field emission scanning electron microscope (FE-SEM) equipped with an EDS detector (Oxford Instruments).

3. Results

The structural modification of pristine graphite particles into functionalized nano-lubricants were carried out with an aim to alter its mechanical, dispersion and wetting behaviour in a fluid medium. At first oxidation of graphite lead to formation of hydrophilic GO, which further used as a precursor to prepare the exfoliated reduced counterparts, graphene and GO-ODA. GO largely consists of two types of chemical functionalities: (a) hydroxyl and epoxide groups, which are placed in the basal planes gallery, and (b) carboxylic groups, which are largely located on the edges of the nanosheets⁴⁶ therefore making it hydrophilic in nature. The presence of oxygen functionalities in GO make that highly susceptible to further structural modifications and are usually targeted for reduction and chemical functionalization. Chemical and structural changes that occurred during the conversion of GO into their reduced counterparts were examined by spectroscopic and microscopic techniques.

The transformation of GO to graphene using hydrazine monohydrate and RGO with ODA were investigated with FTIR, UV-Vis and Raman spectroscopy. Fig. 1a shows the FTIR spectra of pristine GO and its reduced counterparts GO-ODA and graphene. The spectrum of graphite oxide show O–H (carboxyl) at 1337 cm⁻¹ and O–H at 3412 cm⁻¹ originated from carboxylic acid and the O–H stretching mode of intercalated water. The spectra also confirms various functionalities through C=O in carboxylic acid and carbonyl moieties at 1720 cm⁻¹, C–O (epoxy or alkoxy) at 1092 cm⁻¹ and C=C at 1627 cm⁻¹ assigned to skeletal vibrations of unoxidized graphitic domains or contribution from the stretching deformation vibration of the added functional groups. The intensities of the IR peaks corresponding to the oxygen functionalities such as the C=O stretching vibration peak at 1720 cm⁻¹ and O–H deformation peak at 1337 cm⁻¹ decreases significantly after reduction. The FT-IR spectra of GO-ODA show two new peaks at 2918 and 2848 cm⁻¹ resulting from the –CH₂ stretching of the octadecyl chain together with the peak at 720 cm⁻¹ confirms the presence of octadecyl chains on GO-ODA. Furthermore, a new peak at 1564 cm⁻¹ (N–H stretching vibration) indicating the formation of –C–NH–C– bands due to the reaction between the epoxide group and the amine group.⁴¹ The peaks at ∼3412 and ∼1627 cm⁻¹ are related to the –OH and C=C bonds, respectively. As these bonds remained intact after ODA functionalization the corresponding peaks remain in the same position. In case of graphene, the intensity of the –OH stretching vibration peak at ∼3458 cm⁻¹ decreases significantly. The peak at 1720 cm⁻¹ of GO is completely disappeared in the reduced GO and ODA functionalized GO suggesting the removal of –C=O functional groups by the hydrazine reduction process as well as the functionalized by ODA. The absence of epoxy and alkoxy peaks of GO in the reduced particles suggesting successful reduction of GO with hydrazine and also ODA. All these data suggest that most of the abundant oxygen containing functional groups of GO have been successfully removed in the reduced GO.

Fig. 1 (a) FT-IR transmittance spectra of GO, GO-ODA, graphene, (b) UV-Visible absorption spectra of GO and graphene in water and GO-ODA chloroform medium, (c) Raman spectra of GO, GO-ODA and graphene.

Fig. 1b shows the UV-Vis absorption spectra of GO and graphene dispersed in water and GO-ODA dispersed in chloroform as GO-ODA being hydrophobic, is not well dispersed in water. The absorbance peak attributed to π–π* transitions of graphitic C=C in as-synthesized GO was observed at 227 nm and a shoulder peak at 310 nm is ascribed to n–π* transitions of C–O bonds due to the presence of oxygenated functionalities.^42,46 The π–π* peak red-shifted to 267 nm upon reduction to graphene with hydrazine and the absence of n–π* transitions at 310 nm indicates the removal of the oxygen functionality from GO after reduction by hydrazine. In GO-ODA, the π–π* transitions of graphitic C=C bonds shifted to 266 nm, and n–π* transitions of C–O bonds completely disappeared similar to graphene indicates the simultaneous formation of RGO and functionalization with ODA molecules.

Raman spectroscopy is a powerful nondestructive tool to distinguish order and disorder in the crystal structure of materials as Raman scattering is strongly dependent upon its electronic structure. The crystal structure of graphite is altered during the oxidation process of graphite to graphite oxide. However, the reduction process of GO to graphene and GO-ODA partially restored the ordered crystal structure and also repaired few of the structural defects in the process. Therefore Raman scattering can be an useful tool to characterize the functionalized graphitic materials. Fig. 1c shows the Raman spectra of GO, GO-ODA, graphene. Raman spectra of graphitic materials are generally characterized by two main features: the G mode appears due to the first-order scattering of E_2g photons by graphitic sp² carbon atoms, and the D band, arising from a breathing mode of κ-point photons of A_1g symmetry^47,48 due to defects. In the present study due to oxidation, the D band of GO is positioned at 1350 cm⁻¹ with a broader peak, indicating the destruction of in plane sp² character and the formation of defects in the sheets. The G band of GO became more prominent and shifted to 1598 cm⁻¹ due to the presence of isolated double bonds that vibrate at higher frequencies than the D band. In case of graphene the G band shifted to 1584 cm⁻¹ indicating the removal of oxygen moieties and restoration of the sp² network during the reduction process and D band remain almost same at 1346 cm⁻¹ compared to GO. The Raman spectrum of GO-ODA are in line with the graphene, indicating a successful simultaneous reduction and functionalization process. The formation of GO-ODA pushed the G band from 1598 to 1588 cm⁻¹, and getting close to that of graphene (1584 cm⁻¹) and pristine graphite (1581 cm⁻¹), indicative of a larger degree of recovery of graphitic domains.

Representative transmission electron microscopy (TEM) images of the as-synthesized graphene and GO-ODA samples are shown in Fig. 2 where the samples were prepared by drop casting of ethanol/chloroform dispersion on a microscopic grid. The graphene and GO-ODA produced herein clearly display the nanostructured form of the carbon product. The TEM micrograph in Fig. 2a shows the degree of sheet structure and the tendency for sheets to coalesce into overlapped regions for graphene samples and indicating the production of layer graphene. The thickness of graphene sheets are estimated using high resolution TEM imaging to be around 2 nm as shown in Fig. 2b, indicating the ultrathin nature of the hydrazine reduced graphene sheets. However GO-ODA is found to have thin wrinkly flakes and several folded overlapped regions as reported earlier.⁴⁹ These flakes could be a assembly of few graphene layers covered with long octadecyl chains. The ODA functionalized sample showed no distinct sheet formation as seen with reduced graphene. The wrapping of long alkoxy chains on the sheets play a crucial role in their distinct dispersion and wetting behaviour.

Fig. 2 (a) High-resolution TEM image of the graphene, (b) high magnification TEM image of graphene (c) high-resolution TEM image of the GO-ODA.

The influences of surface energy and micro scale surface roughness on the wetting properties of graphene and GO-ODA films are accessed by the contact angle measurements, which are shown in Fig. 3. After hydrazine reduction and ODA functionalization, the reduced particles were hydrolytically pressed to form pellets on which the contact angles were compared. The starting GO film (Fig. 3a) exhibited a contact angle of 50.6° due to the presence of hydrophilic oxygen functionalities in the lattice plane. Following reduction of GO to graphene, contact angle increases to 89.8° indicating partial hydrophobic nature of graphene as a result of comprehensive reduction of the hydrophilic groups (Fig. 3b). Similar contact angle has been reported earlier for the epitaxial grown graphene film.⁵⁰ After ODA functionalization the presence of long C18 hydrocarbon chains makes the GO-ODA film more hydrophobic. The contact angle recorded on the GO-ODA pellet increases to 109.8° (Fig. 3c) which is similar for a self assembled monolayer (SAM) of C18 silane (Octadecyl Trichloro Silane ODTS) on Al, therefore confirmed the attachment of hydrocarbon chains of ODA to GO surface.

Fig. 3 Water contact angle measurement on (a) GO pellet, (b) graphene pellet, (c) GO-ODA pellet, and (d) ODTS SAM on Al.

For key lubrication applications involving these particles in the field of lubricated tribology, dispersion issue of the particles remains a great challenge. The particle size, their colloidal stability and transport of particles to the contact in an oil suspension are few key issues which decide the lubricity of the particles. In an oil medium the oil is absorbed by the particles depending on their wetting behavior. The presence of oil in the slip plane weakens the particle and under modest loading or simple agitation in a ultrasonic mixer the particle fragments into smaller units allowing them a good access to the contact zone and participates in generating an effective antifriction transfer film. The particle suspensions (1 mg ml⁻¹) in oil medium were prepared ultrasonically without any added dispersants to manage their size and dispersion prior to the suspension being used as a lubricant in tribology. The particle size distribution of graphite, GO, graphene, GO-ODA measured by dynamic light scattering is shown in Fig. 4. For the prepared samples it was observed that, particles have a wide size distribution, but the majority of them were dispersed within a narrow range, as shown in Fig. 4. Fig. 4a shows that it is possible to achieve very small particles by sonicating a suspension of monolithic graphite. As the graphite particles are hydrophobic in nature, oil can be ingested by the particle and under ultrasonication the fragmentation of particles (60 mesh, ∼250 μm) leads to a broad distribution of the suspended particles with an effective particle size of 280 nm.

Fig. 4 Particle size distribution in the oil suspension mixed using an ultrasonic probe for (a) graphite, (b) GO, (c) graphene and (d) GO-ODA.

The increase in hydrophilicity due to emergence of oxygen functionalities in GO might reduce oil ingestion by the particles, and may thus lead to poor particle fragmentation. The effective particle size increased to 580 nm after oxidation of graphite to GO with a broader bimodal size distribution (Fig. 4b). However the removal of polar functional groups post reduction process allowed the graphene particles to retain its hydrophobicity. The average particle size of 175 nm for graphene and broad range distribution after reduction indicates better oil ingestion and comprehensive fragmentation of the particles (Fig. 4c). Further functionalization with ODA molecules makes GO-ODA more hydrophobic. The size distribution for GO-ODA suspension (Fig. 4d) show an unimodal distribution and average particle size of 275 nm which is less than GO but more than graphene. The hydrophobic nature of GO-ODA allows the particles to suspend in oil, therefore provides a better size distribution compared to GO. However the addition of long hydrocarbon chains to the reduced GO-ODA surface might be responsible for a larger particle size and distribution in compared to graphene.

3.1 Lubricated tribology

The lubricant was prepared by suspending the graphitic particles in paraffin oil. The particle suspensions was added on the steel substrate and macrotribometry tests were done using that suspension as a lubricant. The macrotribometric experiments where a steel pin/ball makes contact with a number of particles simultaneously gives in general idea on how the lubricant particle will behave under real condition and how the trapped particles between the contact transform themselves into a effective anti-friction film to support load with respect to the contact pressure.

Variation in coefficient of friction (COF) with time for the nanolubricants as oil suspensions was measured using POD tribometer with a 4 mm pin loading (contact pressure ∼ 0.9 MPa) and the results are displayed in Fig. 5. The graphite particle, which is the starting material for the functionalized lubricant additives, showed a COF of ∼0.06 but with sliding time this value gradually increased to 0.08. On the other hand GO and graphene started with a high COF (∼0.15) and gradually decreases to a COF of 0.09 for GO and 0.08 for graphene. However GO-ODA act in a different way between the contacts and offer the lowest COF (0.015) among the functionalized particles. The graphite particles due to the weak inter-lamellar bonding facilitates easy shear as intracrystalline slip occurs in the weak interplanar regions. This leads to the formation of smooth transfer films by wear and withstand the load for a quite bit of time before it starts to wear out a little leading to a relatively high COF. The introduction of polar oxygen functionalities in GO provide a poor dispersion of particles in oil medium and a broader size distribution and bigger particle size compared to graphite (Fig. 4b). The larger particles may have difficulty in penetrating into the contact initially which resulted in the enhancement in the initial friction for GO. Once the particles are trapped in the contact, the fragmentation of the particle into smaller ones takes place under load,⁵¹ which initiates a transfer film between the contact and COF remain constant and independent of time.

Fig. 5 Variation in coefficient of friction (COF) with time data from the pin loading tribometric studies (P_m = 0.9 MPa) at normal load of 30 N with sliding speed of 0.4 m s⁻¹.

However the removal of polar functional groups in graphene post reduction facilitates the dispersibility of the particles in oil. Since, the prepared graphene consists of 1–2 nm layers as single flakes, the small enhancement in COF of graphene-oil nanolubricants compared to GO can be attributed to the graphite layered structure and built up of an effective transfer film with time. The COF for graphene is relatively higher compared to that of graphite and can be ascribed to its lamellar structure and self-lubrication. Therefore the accumulation of the transfer film might be more favourable for multi layered graphite rather than single graphene sheet and with sliding time when the transfer film wears out a little the COF value of graphite matches with that of graphene. The functionalization with ODA molecules makes GO-ODA more hydrophobic and that results a controlled particle distribution in the oil medium and provides the least COF among the functionalized particles. This reduction in friction is attributed to the low resistance to shear between layers of GO-ODA stacked by van der Waals interactions. This binding of long chain molecules was further confirmed by the frictional data of pristine ODA molecules suspended in paraffin oil, where it shows a comparatively higher friction (∼0.1). This concludes that ODA molecules alone are unable to provide a low friction barrier. However when bound to the graphitic plane, the presence of long hydrophobic chains on the particles might likely to bind with the viscous oil medium and act like a soft cushion therefore prohibiting the two body contact.

The lubricant molecules when trapped in contact, undergo slip due to traction and transfers part of the particles to the counterface which when further sheared on the track and pressed along with the lubricant molecules already present on the track and form the transfer film.¹ We have examined the load bearing ability of these nanolubricants and the transfer film by continuous pin loading at a step of 10 N after 30 min of sliding without adding any further particle suspensions. The difference lies here is in individual loading a fresh transfer film is formed every time whereas in continuous loading the transfer film that is formed in initial loading has to support the extra loads that has to be added in stepwise. The friction profile shown in Fig. 6 is mostly dependent on the quality of the film formed on the track which appears to be largely influenced by the normal load. The tribological behavior of all the functionalized nanolubricants under pin loading at 30 N load is already discussed. As the load increased gradually from 30 N to 100 N, the COF for graphite and GO suspension also increased a bit almost linearly with load. However the COF for graphene was found to be almost identical irrespective of the load. Although initially the COF for graphene was higher than graphite as discussed above, but at higher loads the COF remains lower compared to graphite. This confirms a better load bearing ability of graphene compared to graphite as former has an advantage of better mechanical and thermal properties which reflects in tribological performance at higher loads. GO-ODA shows a friction behavior similar to graphite and GO. Although the COF increased linearly with load, it remains the lowest amongst all the particles. The increase might be due to the lowering of the cushion effect as under high load and pressure the long chain hydrocarbons will be squeezed out such that the graphitic sheets might be exposed in the contact, therefore increasing the friction values a little. For comparison the experimental COF values for the individual loading tests at 30 N, 60 N and 90 N are shown in Fig. 6 and those values are found to be similar to those observed COF in continuous loading tests at parallel loads. This indicates that the tribo films that formed initially is quite stable over the time and didn't wear out with the additional loads that was added further.

Fig. 6 Variation in coefficient of friction (COF) with time data for continuous pin loading (30–100 N, P_m ∼ 0.9 MPa to 3 MPa) tribometric studies with sliding speed of 0.4 m s⁻¹. The average COF from the individual pin loading at different loads (30 N, 60 N, 90 N) are shown as symbols.

To validate the effect of contact pressure on the frictional performance of the nanolubricants and the transfer film, we carried out the tribological tests at very high contact pressure (∼1.2–1.8 GPa) with a 10 mm ball loading in the macrotribometer. Fig. 7 shows the COF profiles for the ball loading tribological tests at different loads. The frictional trend was found to be nearly opposite (Fig. 8) to what established at low pressure (pin loading, 0.9 MPa). The COF reported for graphite at ball loading (∼0.08) is relatively higher than the pin loading (∼0.06) at the corresponding load. However this value found to be similar to the values what reported for pin loading at higher loads. Graphene on the other hand displayed similar COF at two different pressure regimes. This authenticated the stability and durability of graphene transfer film. But the genuine contradiction in both the pressure regime was observed for both GO and GO-ODA. The COF evaluated in ball loading tests for GO was decreased drastically whereas for GO-ODA, the value enhanced significantly might be due to the breakdown of the carbon chain shield over the particles, thereby exposing the particles at contact at very high pressures.

Fig. 7 Variation in coefficient of friction (COF) with time data from the ball loading tribometric studies (P_m ∼ 1.2–1.8 GPa) at normal load of (a) 20 N, (b) 50 N, (c) 80 N with sliding speed of 0.4 m s⁻¹.

Fig. 8 Comparison in coefficient of friction (COF) from the pin (P_m ∼ 0.9 MPa) and ball loading (P_m ∼ 1.32 GPa) tribometric studies at normal load of 30 N with sliding speed of 0.4 m s⁻¹.

Further we have studied the load bearing ability of these nanolubricants and the stability of transfer film at high pressure regime by continuous ball loading at a step of 30 N after 30 min of sliding without adding any further particle suspensions and compare the COF with the individual loading experiments at similar load. The COF deduced from individual loading was compared against their counterpart continuous loading and shown in Fig. 9. Apart from GO, the COF recorded for other graphitic particles were found to be nearly identical in both individual and continuous loading. The hydrophobicity of all the three particles allow that oil to be ingested throughout the particles and under pressure the particles further rupture to build an effective thick transfer film between the contact. Thus it can be believed that once the transfer film is formed in the initial test, that film has enough strength to deal with any extra load that is added further. However the tribological behavior of GO is quite contrary compared to the other particles. The observed COF was the lowest among all the tested particles. The COF observed for both individual and continuous loading at 30 N load is quite similar. But at higher loads these values differ significantly. The COF recorded in individual loadings are reasonably less than the continuous loading. This signifies that the tribo film that formed at first, starts to deplete as the load added in stepwise whereas in individual loading the film that forms between the contact initially gradually accumulated to assemble a thicker film that hold well for the entire tested duration at each applied load.

Fig. 9 Variation in coefficient of friction (COF) with time data from continuous ball loading (20 N to 50 N to 80 N, P_m ∼ 1.2–1.8 GPa) with sliding speed of 0.4 m s⁻¹. The average COF from the individual ball loading at different loads are shown as symbols.

4. Discussion

Taking into account of the above set of tribological data, it is somehow interesting to examine the lubrication mechanism and tribological behaviour of the functionalized particles dispersed in oil medium. The above tribological experimental results suggest that the frictional condition can be tuned as per the contact requirement by the functionalization of the particles.

Fig. 10 shows the COF of the lubricant particles with respect to the contact pressure. Graphite and single layer graphene being similar in structure, the experimental COF was found to be nearly independent of the contact pressure and show similar frictional behaviour. However GO and GO-ODA both displays quite diverse frictional behaviour and appears to be contrary to each other. As we discussed earlier, the presence of long carbon chains in GO-ODA induce a cushioning effect between the contact at low contact pressure. As the pressure builds up, the said effect deteriorates, thereby exposing the particles between the contact and provides a frictional behaviour similar to graphene. However GO being hydrophilic in nature, disperse poorly in oil. At low pressure the transport of these particles to the contact may not match to the more dispersed hydrophobic particles. That's why the experimental COF at low pressures are moderately high compared to other particles. Increasing the normal pressure pushed the particles close to the contact and appears to enhance the ability of the particle to stay on the substrate. Once it is very near to the metallic substrate, they are likely to adhere to substrate more strongly due to its hydrophilic nature. We suggest that particle fragmentation increases with normal pressure and also increasing number of contact. Therefore we propose that smaller particles post fragmentation have greater access to contact in reducing friction. Fig. 7 showed that the experimental COF reduced gradually with time as the tribo film built up very effectively between the contact. Similar conclusions have been reported previously by several researchers^1,12,13,52 and are well known mechanisms for the excellent lubricity of solid lubricants.

Fig. 10 Variation in coefficient of friction (COF) with contact pressure (P_m ∼ 0.9 MPa to 2 GPa) from pin/ball on disc tribometric studies of the prepared lubricant particles.

The ability of a particle to migrate to and stay at contact appears to be the principal requirements for good tribological performance of the functionalized particles tested here.^1,12,13 These particles are easily dispersed in paraffin oil without any added surfactant and found to be stable for months. One of the most important finding of the present study is the ability of the dispersed particles to form the transfer film in contact dictated by the tribological parameter and particle functionalization to promote wear resistance to the substrate. We will discuss here this mechanism in two different load regime separately. The lubricant particles when added to the base oil are migrated to the contact and it fills up the micro- and nanoasperities of the mating surfaces and prevent direct metal contact and reduces the friction as shown in Fig. 11.

Fig. 11 Optical images of the tribometer tracks generated with pin on disc tests (30 N) for (a) graphite, (b) GO, (c) graphene, (d) GO-ODA suspended in oil medium.

When these particles are squeezed in the contact, particle slabs are sheared between asperities of the mating surface to commence generation and regeneration of transfer films with sliding time in order to prevent metal to metal contact. As the process of squeezing and shearing of the particles is driven by contact pressure, this mechanism may not be valid for all the particle and depend on particle properties. For pin loading tests, where the contact pressure is low, graphite and graphene being layered pristine particles follow the above mechanism. Even if at high pressure regime for ball loading tests, they still follow the same mechanism. Thus their frictional characteristics are almost independent of contact pressure. Earlier we have seen the contrast tribological behavior of GO and GO-ODA with respect to the contact pressure. At low pressures the coefficient of friction of GO is quite high as bigger suspended particles might not get access to the contact zone. The corresponding disc track showed a thin and discontinued particle film on the surface as shown in Fig. 11b. Alternatively the tribo track for GO-ODA (Fig. 11d) showed the presence of a thick and continuous film without any polishing marks or any wear damage on the surface, thus validating the cushioning effect of the long carbon chain wrapped graphitic particles as we discussed before.

The comparison of the wear scar diameter of the balls used in ball loading tests for all the four particles is shown in Fig. 12. All the tested particles except GO showed similar degree of wear scar diameter on the balls. The recorded coefficient of friction for these particles at 30 N ball loading were found to be similar as shown in Fig. 10. Additionally for GO, at high contact pressures bigger particle slabs slip further to adhere to the substrate to enhance film coverage and stability, which in turn reduced the recorded coefficient of friction. The obvious decrease in wear scar diameter of steel ball used for GO can be due to the proficient access of hydrophilic GO sheets into the contacts easily under high pressure.

Fig. 12 Variation in wear scar diameter for the balls used in ball on disc tests (30 N) for the lubricant particles. The optical images of the ball used in GO and GO-ODA tests are also shown.

The composition and structure of the film present in the slid tracks were investigated by detailed surface analysis of the counter pin and ball surfaces with FE-SEM coupled with EDS. Fig. 13 shows the SEM images of the rubbing pin and ball surfaces with GO and GO-ODA as well as the corresponding EDS analysis. For GO, the pin loading surface shows the presence of a very thin film on the pin surface and the corresponding EDS analysis shows a dominant Fe signature from the pin surface. However the ball loading data shows a significant presence of broken particles and film on the surface and an increase in carbon content in the EDS spectra. In contrast, for GO-ODA, the pin surface for pin loading test was found to be covered with a thick film and large number of broken particles. The carbon content was predominantly present in the EDS spectra with a minimal signature of Fe. For ball loading test, a discontinuous film was present on the ball surface with a increased Fe content in the EDS spectra indicating a thinner particle film on the ball surface. This result clearly proved that GO at high pressure and GO-ODA at low pressure in oil easily form protective coatings. The presence of oxygen in GO was reflected by its signature peak in the EDS spectra where as for GO-ODA the oxygen content was absent indicating its removal post reduction.

Fig. 13 High resolution FE-SEM and corresponding EDS images of the tribometer tracks generated for GO with (a) pin loading (P_m ∼ 0.9 MPa) and (b) ball loading (P_m ∼ 1.32 GPa), and GO-ODA with (c) pin loading (P_m ∼ 0.9 MPa) and (d) ball loading (P_m ∼ 1.32 GPa).

Further we carry out Raman spectroscopy of the deposited particles and tribo films (Fig. 14) on the track to identify the nature of the particles present on the track. Earlier it was reported that^4,5 the Raman intensity can be a proportion to identify the thickness of the particle film present on the track. When a film present on the substrate the reflection from the substrate changes the spectrum of the film with reduced intensity compared to that of the bulk particles. The intensity reduces with decreasing film thickness. In our present experiments, a weak Raman signal was observed for the film present on the track tested with GO (Fig. 13a) in pin on disc arrangement. However when the same particles were tested under a ball on disc configuration, a higher intensity signals with a G band of 1598 cm⁻¹ was obtained (Fig. 14b) which resembles the bulk spectrum of GO. This indicates the presence of a thicker film in a ball on disc configuration which is responsible for a low coefficient of friction at higher contact pressure as discussed earlier. However the tracks corresponding to GO-ODA in pin on disc configuration (Fig. 14c) in general reflect the bulk spectrum with higher intensity peaks with a G band of 1588 cm⁻¹. This once again confirms the presence of a thicker film in the contact interface at low pressure regime which supports the load and leads to a low friction condition. At higher pressure regime in a ball on disc configuration the Raman spectrum was observed with the same shifts (Fig. 14d) as that of greater film thickness but with lower intensities implying the presence of a thin smeared film at the interface.

Fig. 14 Raman spectra of the generated tribo films from the tracks generated from pin/ball on disc tests formed by GO and GO-ODA particles under various conditions.

The above data concludes that the ability of these functionalized particles to migrate and adhere at the contact interface would to be the prime requirements for their good tribological performance. In this present study we believe the above proposition decides the lubrication behavior of the particles at various lubrication conditions. The dispersion of polar GO particles in oil medium might be poor (Fig. 4), but once these particles are allowed to migrate and stay at contact, they likely to give rise to very thin sheared slabs and form a film to provide lubrication. At higher pressures these slabs having active oxygen functionalities were further pushed down and may adhere with the active metallic counterface to yield a thicker particle film which may be tribologically useful. On the other hand the hydrophobic GO-ODA particles likely to assemble together in a non polar medium at the metallic contact and build a bulky film to minimize the friction substantially. This film likely to rupture at a higher pressures, thereby exposing the particles in the interface, where the friction values resemble the pristine graphene and graphite. These observations are corroborated by the Raman spectroscopy and post imaging results of the tribo tracks. Present work thus specify the prospect of obtaining good tribological property with functionalized nanoparticles where such particles may be made useful by suitable functionalization as per the tribological demands.

5. Conclusions

A versatile approach was addressed here to prepare functionalized graphitic lubricant particles. Relatively hydrophobic graphite was first oxidized to form hydrophilic GO and then it was used as precursor to prepared exfoliated hydrophobic graphene and functionalized reduced GO-ODA. Chemical and structural characterizations based on FTIR, UV-Vis, TEM and Raman spectroscopy reveal that the oxygen functionalities along the basal plane gallery are reduced significantly. The functionalization with ODA molecules results an extended π conjugated layered network and the presence of alkyl chains impart better hydrophobicity in GO-ODA and facilitates its dispersion in non-polar paraffin oil.

In lubricated tribological tests the failure of transport and adhesion of lubricant particles to the area of contact and their instability under shear results high friction. Functionalization with suitable organic functionalities facilitates the structural and dispersion stability of the particles and assist the particles to travel and adhere to the contact helping in controlling the friction. The lubricant particles form a transfer film on the substrate under shear to provide lubricity and the ability to build this film is dependent on the functionalization and tribological conditions. The recorded coefficient of friction of pristine layered graphite at low contact pressures is quite low compared to those at high contact pressures due to possibility of shearing out of contact. In contrast graphene provides a constant coefficient of friction irrespective of contact pressures. The excellent tribological performance of graphene is attributed to the ultimate mechanical strength and unique topological structure.

On the other hand the performance of hydrophilic GO and hydrophobic GO-ODA is quite contrary to each other with reference to contact pressure. GO being hydrophilic is poorly dispersed in oil and particle size is comparatively higher than other functionalized particles. Therefore the big particles might have difficulty to access the contact. Therefore at low contact pressure the friction coefficient is quite high. However with increasing contact pressure, the trapped particles rupture to yield smaller units which might have easy access to contact. The hydrophilic particles have better adherence to the substrate, which leads to formation of a stable thick continuous film on the interface, thereby lowering the friction. The organized interaction between the carbon chains attached to the graphene nanosheets and the paraffin oil plays a crucial role in controlling their frictional behaviour. This interaction gives rise to a assembly of lubricant and oil molecules together at contact to induce a kind of cushion effect at low contact pressure to reduce the friction to lowest. The lubricity enhancement could be attributed to the covered weakly stacked lamellar structure of graphene nanosheets with long ODA molecules which prevent direct contact between the rubbing surfaces. But at high pressures this effect is cracked down to expose the graphene sheets at contact and provide friction values similar to reduced graphene. This approach provides a new prospect to develop stable dispersion of functionalized graphene, which have enormous scope for its utilization as tunable lubricant additive in industrial applications as per demand.

Acknowledgements

The authors are grateful to CSIR, India for the grant through 12FYP project ESC-0112 and ESC-0203 in carrying out this work. Special thanks to Central Research Facility for providing the SEM-EDS data. We are also thankful to our Director for the permission to publish this work.

References

1. I. L. Singer, Langmuir, 1996, 12, 4486–4491.
2. V. N. Bakunin, A. Y. Suslov, G. N. Kuzmina and O. P. Parenago, J. Nanopart. Res., 2004, 6, 273–284.
3. Y. Penga, Y. Hua and H. Wanga, Tribol. Lett., 2007, 25, 247–253.
4. R. R. Sahoo and S. K. Biswas, Tribol. Lett., 2014, 53, 157–171.
5. R. R. Sahoo and S. K. Biswas, Thin Solid Films, 2010, 518, 115–125.
6. V. Eswaraiah, V. Sankaranarayanan and S. Ramaprabhu, ACS Appl. Mater. Interfaces, 2011, 3, 4221–4227.
7. M. Chhowalla and G. A. J. Amaratunga, Nature, 2000, 407, 164–167.
8. A. P. M. Barboza, H. Chacham, C. K. Oliveira, T. F. D. Fernandes, E. H. M. Ferreira, B. S. Archanjo, R. J. C. Batista, A. B. de Oliveira and B. R. A. Neves, Nano Lett., 2012, 12, 2313–2317.
9. L. J. Pottuz, F. Dassenoy, M. Belina, B. Vache, J. M. Martin and N. Fleischer, Tribol. Lett., 2005, 18, 477–485.
10. S. Kumari, O. P. Sharma, R. Gusain, H. P. Mungse, A. Kukrety, N. Kumar, H. Sugimura and O. P. Khatri, ACS Appl. Mater. Interfaces, 2015, 7, 3708–3716.
11. R. G. Dickinson and L. Pauling, J. Am. Chem. Soc., 1923, 45, 1466–1471.
12. K. J. Wahl and I. L. Singer, Tribol. Lett., 1995, 1, 59–66.
13. K. J. Wahl, M. Belin and I. L. Singer, Wear, 1998, 214, 212–220.
14. K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Science, 2004, 306, 666–669.
15. Y. B. Zhang, Y. W. Tan, H. L. Stormer and P. Kim, Nature, 2005, 438, 201–204.
16. C. Lee, X. D. Wei, J. W. Kysar and J. Hone, Science, 2008, 321, 385–388.
17. S. Ghosh, I. Calizo, D. Teweldebrhan, E. P. Pokatilov, D. L. Nika, A. A. Balandin, W. Boa, F. Miao and C. N. Lau, Appl. Phys. Lett., 2008, 92, 1–3.
18. A. A. Balandin, S. Ghosh, W. Z. Bao, I. Calizo, D. Teweldebrhan, F. Miao and C. N. Lau, Nano Lett., 2008, 8, 902–907.
19. D. A. Areshkin and C. T. White, Nano Lett., 2007, 7, 3253–3259.
20. F. Schedin, A. K. Geim, S. V. Morozov, E. W. Hill, P. Blake, M. I. Katsnelson and K. S. Novoselov, Nat. Mater., 2007, 6, 652–655.
21. T. Hung, T. Li, Y. Xin, B. Jin, Z. Chen, C. Su, H. Chen and S. Nutt, RSC Adv., 2014, 4, 19814–19823.
22. K. Hou, P. Gong, J. Wang, Z. Yang, Z. Wang and S. Yang, RSC Adv., 2014, 5, 56543–56551.
23. K. Kim, H. J. Lee, C. Lee, S. K. Lee, H. Jang, J. H. Ahn, J. H. Kim and H. J. Lee, ACS Nano, 2011, 5, 5107–5114.
24. W. S. Hummers and R. E. Offeman, J. Am. Chem. Soc., 1958, 80, 1339.
25. S. Stankovich, D. A. Dikin, R. D. Piner, K. A. Kohlhaas, A. Kleinhammes, Y. Jia, Y. Wu, S. B. T. Nguyen and R. S. Ruoff, Carbon, 2007, 45, 1558–1565.
26. X. Z. Tang, W. Li, Z. Z. Yu, M. A. Rafiee, J. Rafiee and F. Yavari, et al., Carbon, 2011, 49, 1258–1265.
27. H. Bai, C. Li and G. Shi, Adv. Mater., 2011, 23, 1089–1115.
28. L. Q. Xu, W. J. Yang, K. G. Neoh, E. T. Kang and G. D. Fu, Macromolecules, 2010, 43, 8336–8339.
29. G. Wang, X. Shen, B. Wang, J. Yao and J. Park, Carbon, 2009, 47, 1359–1364.
30. Z. Y. Lin, Y. Liu and C. P. Wong, Langmuir, 2010, 26, 16110–16114.
31. S. Niyogi, E. Bekyarova, M. E. Itkis, J. L. McWilliams, M. A. Hamon and R. C. Haddon, J. Am. Chem. Soc., 2006, 128(24), 7720–7721.
32. J. Ou, J. Wang, S. Liu, B. Mu, J. Ren, H. Wang and S. Yang, Langmuir, 2010, 26, 15830–15836.
33. D. Berman, A. Erdemir and A. V. Sumant, Carbon, 2013, 59, 167–175.
34. J. Lin, L. Wang and G. Chen, Tribol. Lett., 2011, 41, 209–215.
35. J. Pu, S. Wan, W. Zhao, Y. Mo, X. Zhang, L. Wang and Q. J. Xue, Phys. Chem. C, 2011, 115, 13275–13284.
36. H. Kinoshita, Y. Nishina, A. A. Alias and M. Fujii, Carbon, 2014, 66, 720–723.
37. B. G. Choi, H. Park, T. J. Park, M. H. Yang, J. S. Kim, S. Y. Jang, N. S. Heo, S. Y. Lee, J. Kong and W. H. Hong, ACS Nano, 2010, 4, 2910–2918.
38. D. A. Dikin, S. Stankovich, E. J. Zimney, R. D. Piner, G. H. B. Dommett, G. Evmenenko, S. T. Nguyen and R. S. Ruoff, Nature, 2007, 448, 457–460.
39. Y. Zhu, S. Murali, W. Cai, X. Li, J. W. Suk, J. R. Potts and R. S. Ruoff, Adv. Mater., 2010, 22, 3906–3924.
40. T. Kuila, P. Khanra, S. Bose, N. M. Kim, B. C. Ku, B. Moon and J. H. Lee, Nanotechnology, 2011, 22, 305710.
41. W. Li, X. Z. Tang, H. B. Zhang, Z. G. Jiang, Z. Z. Yu, X. S. Du and Y. W. Mai, Carbon, 2011, 49, 4724–4730.
42. X. Gao, J. Jang and S. Nagase, J. Phys. Chem. C, 2010, 114, 832–842.
43. V. H. Pham, T. V. Cuong, S. H. Hur, E. Oh, E. J. Kim, E. W. Shin and J. S. Chung, J. Mater. Chem., 2011, 21, 3371–3377.
44. P. G. Ren, D. X. Yan, X. Ji, T. Ren and Z. M. Li, Nanotechnology, 2011, 22, 055705.
45. O. K. Park, M. G. Hahm, S. Lee, H. I. Joh, S. I. Na, R. Vajtai, J. H. Lee and B. C. Ku, Nano Lett., 2012, 12, 1789–1793.
46. O. C. Compton and S. T. Nguyen, Small, 2010, 6, 711–723.
47. C. Z. Zhu, S. J. Guo, Y. X. Fang and S. J. Dong, ACS Nano, 2010, 4, 2429–3247.
48. A. C. Ferrari, Solid State Commun., 2007, 143, 47–57.
49. H. P. Mungse, N. Kumar and O. M. Khatri, RSC Adv., 2015, 5, 25565–25571.
50. S. Wang, Y. Zhang, N. Abidi and L. Cabrales, Langmuir, 2009, 25, 11078–11081.
51. R. R. Sahoo, S. Math and S. K. Biswas, Tribol. Lett., 2010, 37, 239–249.
52. R. R. Sahoo and S. K. Biswas, Tribol. Lett., 2010, 37, 313–326.

---