Effects of functional groups on the tribological properties of hairy silica nanoparticles as an additive to polyalphaolefin

Abstract

Hairy nanoparticles, which combine inorganic nanoparticles and organic polymers, have led to a variety of applications due to their special properties. In this paper, hairy silica nanoparticles (HSNs) with different functional groups and tribological properties were prepared and tested as additives to polyalphaolefin (PAO). Unmodified silica nanoparticles (USNs) were synthesized using the Stöber method and then modified by silanes with amino functional groups. The end groups of the silica nanoparticles were further functionalized by tethering different organic chains to the amino functionalized HSNs. HSNs-PAO lubricants were prepared by a four-step process and their tribological performances were tested using a four-ball tribometer. The coefficient of friction was recorded and the wear scar surfaces were examined by SEM and EDS. It was found that HSNs could form a stable homogenous solution when dispersed in PAO and exhibited good anti-wear and friction reduction properties. The NH₂ terminated HSNs exhibited the best tribological performance but were more concentration sensitive than other types of HSNs due to the hydrogen bonds between amino groups. The long wear test results suggested that HSNs with polar functional groups accelerated the running-in process, which could be attributed to the adsorption of HSNs.

---

1. Introduction

Hairy nanoparticles, where the nanoparticles are covered with organic chains, have been widely researched during the last two decades. With organic chains on the surface of the nanoparticles, the dispersion of the nanoparticles in organic solvents is greatly improved. Additionally, functional groups on the organic chains of hairy nanoparticles changes the properties of the nanoparticles and provides various possibilities for applications of the nanoparticles in various fields such as the electrical,¹ medical,² optical³ and mechanical fields.⁴ As the organic chains could enhance the dispersity and stability of nanoparticles in lubricants, the applications of hairy nanoparticles in tribological fields is of special interest. Different kinds of nanoparticles, such as ZrO₂,⁵ CaCO₃,⁶ TiO₂,⁷ Al₂O₃⁸ and carbon,^9,10 have been investigated, along with their tribological performances. Among those nanoparticles, hairy silica nanoparticles (HSNs) are one of the best nanoparticles for tribological applications as they are environmentally friendly and have economically efficient properties. Hairy silica nanoparticles have been prepared and their tribological properties have been investigated.^11–16 Although HSNs exhibited good anti-wear and friction reduction performances, HSNs with different functional groups showed different tribological properties. Most researchers focus on one type of functional group when studying the tribological properties of HSNs. Thus, the effects of different functional groups on the tribological properties of HSNs still need to be further investigated.

Silica nanoparticles with amino functional groups were investigated, along with their tribological performances, as additives in lubricants^11,12 and composites.¹³ Those amino functionalized nanoparticles exhibited good dispersity and stability when added to lubricants or composites, improved the tribological performance of the lubricants and enhanced the mechanical properties. An amino functionalized thin film was also reported to have good friction reduction performance, and amino ligands with longer organic chains showed a lower friction coefficient due to the densely packed organic chains.¹⁷ Despite the fact that amino functionalized nanoparticles show good tribological properties, the interparticle aggregation limits their applications. It was reported by R. Bagwe et al. that amino functionalized nanoparticles aggregated due to nonspecific bonds between amino groups. Comparing with amino groups, carboxyl functional groups could prevent the nanoparticle agglomeration and enhance the stability of the nanoparticles.¹⁸ Besides good dispersity and stability, carboxyl groups have also been reported for good corrosion resistance and frictional durability on metal evaporated tape.¹⁹ Non-polar functional groups such as phenyl groups and long chain alkyl groups were also investigated, along with their tribological performances. R. Jones et al. reported that the phenyl groups could enhance the life time of a self-assembled monolayer and reduce the friction by forming a multicomponent film on the surface.²⁰ Y. Chen et al. studied the tribological performance of a phenyl modified PTFE/silica composite and found that the dielectric properties, wear absorptions and tensile strength were enhanced.²¹ For alkyl groups, it was reported that nanoparticles with long alkyl chains could form a homogeneous solution with base oil and exhibited good load carrying, anti-wear and friction reduction properties,^22,23 while long alkyl chains showed a lower frictional force and better anti-wear properties than short chains.²⁴

Keeping in view of the previous work reviewed above, different functional groups (amino,^11–13 carboxyl,^18,19 phenyl,^20,21 and alkyl^22–24 groups) were investigated, along with their tribological properties, the functional groups directly affected the tribological properties of the HSNs. However, most researchers studied only one type of functionalized nanoparticle and only a few of them have focused on silica nanoparticles. The influence of functional groups on the tribological properties of hairy silica nanoparticles is still unclear. Thus, an investigation on the effect of functional groups on the tribological properties of HSNs is needed to develop the understanding of the tribological properties of HSNs and to extend the applications of them as potential additives in lubricants.

In this paper, a four-step method for preparing HSNs-PAO lubricants, including USN synthesis, HSN synthesis, solvent transfer and HSN dispersion in the lubricant, was developed. The prepared HSNs were characterized by several methods including scanning electron microscopy (SEM), X-ray diffraction (XRD), zeta potential measurements, dynamic light scattering studies (DLS), Fourier transform infrared spectroscopy (FTIR), X-ray photoelectron spectroscopy (XPS) and thermogravimetric analysis (TGA). The viscosity–temperature dependence of the HSNs-PAO lubricants were investigated by a Brookfield viscometer. The tribological performances of the HSNs with different functional groups were tested using a four-ball tribometer. The coefficient of friction (COF) and the wear scar diameter (WSD) for short and long time wear were measured. Based on the test results, the optimal functional groups for HSNs were investigated. Different running-in performances of the HSNs-PAO lubricants were found in the tribological test. It is suggested that adding HSNs into lubricant could reduce the duration of the running-in process and give a stable COF. Meanwhile, HSNs with polar functional groups would accelerate the running-in period. A three period running-in process was developed based on the test results, the mechanism of the anti-wear and friction reduction performance of the HSNs is discussed.

2. Materials and methods

2.1 Materials

The following materials listed were used as received: tetraethylorthosilicate (TEOS, 99% purity, Sinopharm), ammonium hydroxide (28% purity, Sinopharm), ethanol (99% purity, Sinopharm), toluene (99% purity, Sinopharm), N1-(3-trimethoxysilylpropyl)diethylenetriamine (DETAS, 95% purity, Sigma Aldrich), stearic acid (STA, 90% purity, Sinopharm), benzoic acid (BA, 99% purity, Sinopharm), succinic acid (SA, 99%, Sinopharm) and polyalphaolefin (PAO, 99% purity, fox chemical technology).

2.2 Preparation of HSNs-PAO lubricants

Unmodified silica nanoparticles were prepared before modification. 100 nm silica nanoparticles were prepared by the Stöber method. TEOS was first diluted with ethanol and then added into a mixture of ethanol, water and aqueous ammonia with gentle stirring at a temperature of 40 °C for 3 hours. As shown in Fig. 1(a), a transparent solution of light blue colour was obtained after the reaction, which suggested that the silica nanoparticles were monodispersed with a narrow size distribution.

Fig. 1 The schematic diagram of the HSNs-PAO lubricant preparation process: (a) unmodified silica nanoparticles dispersed in ethanol, (b) HSNs dispersed in ethanol, (c) HSNs dispersed in toluene and (d) HSNs dispersed in PAO.

The silica nanoparticles prepared above were modified with DETAS. In order to densely graft the organic chains onto the silica nanoparticles, an excess of silane was added into the silica nanoparticle solution drop by drop with rapid stirring. The solution was continually heated at 50 °C for 12 hours to ensure that the linking reaction had finished. After the reaction, unlinked silane and impurities were removed by dialysis using a 10 000 MWCO snakeskin dialysis tube obtained from Themoscientific. The NH₂ terminated HSNs were obtained after the dialysis process had finished. Those HSNs were covered with amino functional groups and could be further modified by acid-terminated organic chains.

In order to modify the end functional groups of the HSNs, carboxyl-terminated organic chains were tethered to the amino functional groups of the NH₂ terminated HSNs synthesized above. STA, BA and SA were first dissolved in ethanol respectively, and then added into the NH₂ terminated HSNs solution drop by drop with rapid stirring. DCCI was used as a dehydration agent for the formation of peptide bonds between the carboxyl groups and the amino groups. The solution was kept stirring at room temperature for 12 hours to ensure that the linking reaction had finished. The HSNs prepared above were purified again by filtration and dialysis to remove the impurities and unlinked organic chains. HSNs with carboxyl, phenyl and long-chain alkyl end groups were obtained after the purification process had finished. As shown in Fig. 1(b), the transparent solution turned into a white turbid solution after the surface modification reaction, which was due to the formation of micelles in the ethanol solution. The molecular structures of the different HSNs are shown in Fig. 2. The covalent siloxane bridges were created after the linking reaction between the hydroxyl groups and siloxane. The primary amino functional groups were further modified by tethering acid terminated organic chains to them.

Fig. 2 Different types of silica nanoparticles: (a) unmodified silica nanoparticles (USNs), (b) hairy silica nanoparticles (HSNs) and (c) HSNs which were further functionalized by organic chains. The surface structure of (d) USNs, (e) NH₂ terminated HSNs, (f) COOH terminated HSNs, (g) C₆H₅ terminated HSNs and (h) CH₃ terminated HSNs.

In order to disperse HSNs into PAO lubricant, the HSNs were first transferred from ethanol to toluene by azeotropic distillation. Toluene was added into the HSNs–ethanol solution obtained above to form a 3/1 mixture of toluene and ethanol. The solution was kept at 70 °C for 3 hours to remove the ethanol. After the azeotropic distillation process, the HSNs were transferred to toluene. PAO was then added into the solution with rapid stirring. The solution was then transferred to a convection oven to remove the toluene from PAO.

2.3 Characterization of hairy silica nanoparticles

The morphology of the HSNs was characterized by a Hitachi SU8010 scanning electron microscope. The silica nanoparticles were treated with sputter-gold to enhance the conductivity before the experiment. The X-ray diffraction (XRD) patterns of the silica nanoparticles (2θ = 5–90°) were recorded using a Panalytical Empyrean X-ray diffractometer, with Cu Kα radiation at 25 °C, a 2θ step size of 0.02 and a scanning speed of 2° min⁻¹. The Fourier transform infrared spectroscopy (FTIR) of the HSNs was carried out using a Nicolet IS10 with a spectral range of 400–4000 cm⁻¹. The XPS experiments were performed on a Thermo Fisher Scientific X-ray photoelectron spectrometer 250Xi (Al Kα line of 1486.6 eV, energy source 120 W), the ultra-high vacuum analysis chamber was maintained at a typical base pressure of 5 × 10⁻⁹ Torr during sample analysis. The surface area of the unmodified silica nanoparticles was characterized by the Brunauer–Emmett–Teller (BET) method and the nanoparticles were pre-treated under vacuum conditions at 150 °C for 2 h before the BET test. Thermogravimetric analysis (TGA) was employed using a TA Q500 and a heating rate of 10 °C min⁻¹ to 800 °C under flowing nitrogen gas. The zeta potential and dynamic light scattering of the HSNs were characterized by a Zetasizer Nano ZS from Malvern Instruments.

2.4 Tribological properties of hairy silica nanoparticles

The viscosity–temperature dependence for the different types of HSNs-PAO lubricants was investigated by a Brookfield viscometer using a cone/plate tester, with the temperature range of 25–100 °C. The tribological properties of the HSNs-PAO lubricants were tested using a MRS-10 four-ball tribometer. The test apparatus and configuration is shown in Fig. S6.† All the tests were performed at a rotating speed of 1450 rpm, under a load of 390 N at room temperature. 12.7 mm diameter GCr15 steel bearing balls (grade 4) were cleaned ultrasonically with ethanol before the test. The COF and WSD of the HSNs-PAO lubricants with different concentrations and functional groups were tested three times for 30 min and the average COF and WSD values were calculated. Based on the results, the HSNs with different functional groups were tested for 3 h at their optimal concentration to investigate the running-in period and anti-wear performance under long-wear conditions. The wear surfaces of the steel bearing balls after the tribology tests were examined by SEM and EDS. Before the SEM and EDS analysis, the steel bearing balls were rinsed in ethanol for 5 minutes using ultrasonication and were dried in a vacuum drying oven.

3. Results and discussion

3.1 Hairy silica nanoparticle characterization

The morphology of the silica nanoparticles was characterized by SEM. As shown in Fig. 3(a), 100 nm silica nanoparticles were prepared successfully by the Stöber method, with excellent roundness and a narrow size distribution. The NH₂ terminated HSNs were synthesized by modifying USNs with DETAS. The SEM image of the NH₂ terminated HSNs is shown in Fig. 3(b), large clusters of HSNs were found at high concentrations (10 wt%) of the NH₂ terminated HSNs, which could be due to the hydrogen bonding between the amino functional groups.¹⁸ Comparing with the NH₂ terminated HSNs, the functionalized HSNs which were further modified by organic chains show better dispersity at both high and low concentrations. Fig. 3(c) and (d) show the CH₃ terminated HSNs at high concentration (10 wt%) and low concentration (0.5 wt%) respectively. The CH₃ terminated HSNs did not aggregate or form large clusters, but kept well dispersed at both low (0.5 wt%) and high (10 wt%) concentrations. The silica nanoparticles were further characterized by X-ray diffraction (XRD). The XRD pattern (Fig. S1†) of both the USNs and the HSNs exhibited broad peaks at 22°, 31.3° and 36° 2θ, which are the characteristic peaks of silicon. It can be seen from the XRD patterns and the SEM micrographs that the silica nanoparticles retain their structural and size integrity after modification.

Fig. 3 SEM micrographs of (a) 100 nm USNs, (b) NH₂ terminated HSNs, and (c and d) CH₃ terminated HSNs at high concentration (10 wt%) and low concentration (0.5 wt%).

The HSNs were characterized by FTIR and the spectra are shown in Fig. 4. All kinds of the HSNs show a strong band at 1010–1100 cm⁻¹, which is the Si–O–Si stretching vibration. A broad band at 3300 cm⁻¹ was found for the USNs, corresponding to O–H stretching. The peak at 1620 cm⁻¹ for the spectrum of NH₂ terminated HSNs, 1550 cm⁻¹ for CH₃, 1555 cm⁻¹ for COOH and 1552 cm⁻¹ for C₆H₅ corresponds to N–H bending. The broad band at 3400 cm⁻¹ for NH₂ could be assigned to H-bonded N–H stretching. A peak at 1723 cm⁻¹ and a broad band at 3300 cm⁻¹ were found in the spectrum of COOH, which correspond to the C=O stretching and O–H stretching respectively. The peaks which appear at 1454 cm⁻¹ and 1600 cm⁻¹ of the C₆H₅ spectrum are characteristic of the phenyl group while the peaks at 2858 cm⁻¹ and 2917 cm⁻¹ of the CH₃ spectrum correspond to the C–H stretching of the alkane group.

Fig. 4 The FTIR spectra of USNs and HSNs with different functional groups.

The functionalized HSNs were further characterized by XPS, the XPS survey spectrum for the USNs, NH₂ terminated HSNs and CH₃ terminated HSNs are shown in Fig. 5(a–c). It can be seen from the XPS survey spectrum that Si2p at 101.1 eV and O1s at 543.7 eV were found in the spectra of the USNs. After modification with amino functionalized silanes, an N1s peak at 397.5 eV emerged, which indicated that the amino functional groups were successfully introduced on the silica nanoparticle surface. With further functionalization using long alkyl chains, the peak area of C1s increases by 30% while the peak area of N1s remains the same. The deconvoluted C1s subregions for the NH₂ terminated, COOH terminated, C₆H₅ terminated and CH₃ terminated HSNs are shown in Fig. 5(d–g). As shown in Fig. 5(d), the XPS spectrum of C1s could be deconvoluted into three Gaussian peaks, which are typical peaks of C–Si at 282.7 eV, C–H at 283.9 eV and C–C at 284.6 eV. After functionalization with the carboxyl groups, three new peaks emerged, which are the C–O peak at 285.8 eV, the C=O peak at 287.1 eV and the O–C=O peak at 288.9 eV. The C1s subregion for the C₆H₅ terminated HSNs was deconvoluted into four peaks, corresponding to the C–Si peak at 282.8 eV, the C–H peak at 283.8 eV, the C–C peak at 284.6 eV, and the C₆H₅ peak at 286.2. The spectra for the CH₃ terminated HSNs was deconvoluted into three peaks, which are the C–Si peak at 282.9 eV, the C–H peak at 283.8 eV and the C–C peak at 284.6 eV. It can be concluded from the experiment results that the functional groups were introduced on the surface of the silica nanoparticles successfully.

Fig. 5 The XPS survey spectra of (a) USNs, (b) NH₂ terminated HSNs and (c) CH₃ terminated HSNs. Deconvoluted C1s subregions of (d) NH₂ terminated, (e) COOH terminated, (f) C₆H₅ terminated and (g) CH₃ terminated HSNs.

The organic content of the HSNs was quantified by TGA. HSNs with different organic chains were heated from room temperature to 800 °C at a heating rate of 10 °C min⁻¹, the result is shown in Fig. 6. It can be seen from the graph that all four TGA curves show the same trend, in which the organic chains start to degrade around 300 °C, and are fully degraded around 600 °C while the inorganic silica core remains at the end of the test. The surface area of the unmodified silica nanoparticles was tested with the BET method and the surface area was found to be 271 m² g⁻¹. Based on the BET and TGA test results, the grafting density of the HSNs was 1.1–1.3 ligands per nm².

Fig. 6 Thermogravimetric analysis of HSNs with different functional groups.

The zeta potential of the silica nanoparticles dispersed in ethanol was characterized by a Malvern Zetasizer Nano ZS. The zeta potential value for the USNs was measured at −42 mV, which was due to the abundant hydroxyl groups on the surface of the silica nanoparticles. After modification with amino functionalized silanes, the zeta potential increased to +27 mV, which can be attributed to the primary amino groups on the surface of the HSNs. The change in the zeta potential value confirmed that the amino functional groups were successfully introduced onto the surface of the silica nanoparticles. The zeta potential value decreased to about +21–+24 mV after further functionalization of the silica nanoparticles (Fig. S3†), which could be due to the transfer from primary amino groups to secondary amino groups.

The long-term stability of the HSNs-PAO lubricants was studied by keeping the lubricants standing for 2 months after the preparation process had finished. A photograph of the different kinds of lubricants is shown in Fig. 7. The four kinds of HSNs-PAO lubricants prepared by the four-step process introduced in this paper are shown in Fig. 7(a–d) and two kinds of HSNs-PAO lubricants prepared by a common method are shown in Fig. 7(e and f). It can be seen that the HSNs form homogenous solutions with PAO lubricants and remain stable after 2 months standing. The CH₃ (shown in Fig. 7(a)) and C₆H₅ (shown in Fig. 7(b)) terminated HSNs show better optical transparency than the COOH and NH₂ terminated HSNs, which can be attributed to excellent compatibility of the non-polar functional groups in the organic solvent. Fig. 7(e) shows the USNs dispersed in PAO and Fig. 7(f) shows NH₂ terminated HSNs dispersed in PAO using a common method which dries the HSNs to powder after the preparation, then redisperses them into the lubricant by stirring or ultrasonic dispersion. Comparing with Fig. 7(d) in which the same kind of HSNs are dispersed in PAO by the four step process, the HSNs dispersed by common method have partly sedimented on the bottom after 2 months standing, which suggests that the HSNs dispersed by common method have worse stability than the HSNs dispersed by the four-step process. Meanwhile, as shown in Fig. 7(e), the USNs dispersed in PAO have completely sedimented on the bottom of the bottle after 2 months standing, which shows the worst stability among all the lubricants shown in Fig. 7. It can be found from the long-term stability test that the HSNs dispersed by the four-step process formed homogeneous solutions with PAO and exhibited excellent stability. This point is confirmed by the DLS test (Fig. 8) as the average particle size of the HSNs in PAO only increases slightly after 2 months standing. Keeping in view of the four-step process from the synthesis of the HSNs to the dispersion in PAO, the nanoparticles were kept dispersed in solutions (ethanol, toluene and PAO) during the preparation of the HSNs-PAO lubricant. Comparing with the commonly used process which dries the nanoparticles before dispersing them into lubricants, the new process increases the stability of the HSNs while avoiding the oxidation of the HSNs during the drying process.

Fig. 7 Photograph of PAO with (a) CH₃ terminated HSNs, (b) C₆H₅ terminated HSNs, (c) COOH terminated HSNs and (d) NH₂ terminated HSNs prepared by the four-step process, and PAO with (e) USNs and (f) NH₂ terminated HSNs dispersed by a common method.

Fig. 8 The average particle size of the HSNs dispersed in ethanol, toluene and PAO (the orange column is the average particle size of the HSNs just after the dispersion and the green column is the particle size after HSNs-PAO is left standing for 2 months).

The dispersity and stability of the HSNs were also studied by characterizing the nanoparticle size of the HSNs dispersed in ethanol and PAO using dynamic light scattering. The result is shown in Fig. 8. The average particle size of the unmodified silica nanoparticles in ethanol was 102.1 nm. After modification with amino functionalized silane, the average nanoparticle size increased to 356.3 nm, indicating the aggregation of primary HSNs. With further functionalization using carboxyl, phenyl and alkyl chains, the nanoparticle sizes increased slightly by 20–40 nm. However, when the HSNs were transferred from ethanol to toluene, the average particle sizes of the HSNs decreased sharply to 170–270 nm, which indicated high dispersion of the HSNs in the nonpolar medium. After transfer from toluene to the PAO lubricant, the average particle sizes of the HSNs increased slightly, and after 2 months standing the average particle sizes increased by tens of nanometers. It can be seen from the results that the HSNs did not disperse well in ethanol but when transferred into a non-polar solvent, the dispersion of the HSNs improved significantly. Among the HSNs, the CH₃ terminated HSNs exhibited the smallest average particle size and the best stability when dispersed into PAO, which implied the best compatibility between the nanoparticles and the lubricant. The NH₂ and COOH terminated HSNs were found to have a larger average particle size and the particle size increased more after 2 months standing compared to the other types of HSNs, which indicates easier aggregation when dispersing in PAO.

3.2 Tribology properties of hairy silica nanoparticles

The viscosity of pure PAO and the HSNs-PAO lubricants was tested using a cone-plate rheometer tester. The viscosity as a function of HSN concentration and temperature are shown in Fig. S4 and S5† respectively. It can be seen from Fig. S4† that with an increase in the HSN concentration, the viscosity of the HSNs-PAO lubricants increased. The NH₂ and COOH terminated HSNs exhibited a higher viscosity, which increased at a higher rate, than the other HSNs (the viscosity of the NH₂ and COOH terminated HSNs-PAO lubricants had increased by 80% at a concentration of 4 wt% compared to the viscosity of PAO) while the CH₃ terminated HSNs exhibited the lowest viscosity (the viscosity of the CH₃ terminated HSNs-PAO lubricant had increased by 20% at a concentration of 4 wt% compared to the viscosity of PAO) among all the HSNs. From Fig. S5,† all kinds of the HSNs show good viscosity–temperature properties, the viscosity index of the PAO lubricant increased by 3% after adding the HSNs into it (see ESI†).

The COF and WSD values of different concentrations of the HSNs with different functional groups are shown in Fig. 9. Comparing with the PAO lubricants without HSNs, all types of the HSNs exhibited good anti-wear and friction reduction properties when added into PAO with appropriate concentrations. Comparing with the COF and WSD values of pure PAO without HSNs, which were 0.093 and 870 μm, the COF and WSD values of the HSNs-PAO were reduced by 40% under optimal conditions. It can be seen from the graph that the amino functionalized nanoparticles show the best friction reduction and anti-wear properties (when their concentration was at 1 wt%, the COF was 0.055 and the WSD was 301 μm) among all types of the HSNs. That can be attributed to the chelation effect of the amino groups and the hemilability of amine–metal bonds,²⁵ which could enhance the interaction between the HSNs and metal surfaces. With more HSNs adsorbed on a metal surface, the protecting effect (formation of a protecting film on the metal surface), rolling effect (prevention of two surfaces contacting) and polishing effect (improvement of the surface roughness by filling the nano-grooves and polishing the nano-bumps) would be improved. However the tribological performance of the amino functionalized HSNs shows stronger concentration sensitive behaviour than the other three types of HSNs, in which the COF and WSD values fluctuated drastically with the concentration. This phenomenon may be caused by the polarity of the primary amino functional groups and the hydrogen bonds between those amino functional groups. With strong polarity, the compatibility of the amino functionalized HSNs with the PAO lubricant would not be as good as the other three types of HSNs. Meanwhile, the amino functionalized HSNs would attract each other and form large clusters using hydrogen bonds, which would cause three-body abrasion wear. As shown in Fig. 3(b), the NH₂ terminated HSNs form large clusters at high concentration. The hydrogen bonds between the amino groups would give rise to aggregation of the nanoparticles, which might lead to a partly uneven distribution at low concentration (0.5%) and the formation of large clusters of nanoparticles causing three body abrasion at high concentration (2% and 4%). As a result, an insufficient or excess amount of the amino functionalized HSNs would cause wear damage and increase the COF. The COOH terminated HSNs show worse anti-wear and friction reduction performances, which might be due to the corrosive effect of the acid functional groups. With further modification using the organic chains, the COF and WSD values of the HSNs become less concentration sensitive, which may be attributed to the longer organic chains and lower polarity. With longer organic chains and fewer primary amino functional groups, the dispersity and stability of the C₆H₅ and CH₃ terminated HSNs were enhanced dramatically, meanwhile, with their nonpolar functional groups, the HSNs more easily formed homogenous stable solutions when dispersed into PAO lubricants.

Fig. 9 The COF and WSD values of HSNs with different functional groups.

The COF values versus time for the HSNs with different functional groups are shown in Fig. 10. It can be seen from the figure that all the COF values show a similar trend: a remarkable increase in the COF is found followed by a period of smooth running, the increase is the start of the running-in period. When new equipment is running for the first time, the moving parts will wear against each other severely until they settle into a stable condition. The HSNs with different functional groups show different running-in performances: the COF of pure PAO rose up dramatically at 1900 s and became stable at 0.11 around 6000 s. Comparing with pure PAO without HSNs, all types of the HSNs-PAO lubricants show a shorter duration of the running-in stage and a lower steady state COF. The COOH terminated HSNs show the highest peak value and longest duration time among the four kinds of HSNs (the COF of the COOH terminated HSNs increased to 0.095 around 3600 s and became stable at 5400 s). With an acidic end group, the COOH terminated HSNs might react with the metal surface and cause corrosion during the formation of the transfer film. That would lead to an unstable development of the transfer film and result in a longer and unstable duration time for the running-in period. The NH₂ terminated HSNs exhibited the lowest stable COF, which could be attributed to the higher adsorption of the HSNs on the metal surface. Under the effects of physical adsorption (the chelation effect and polar adsorption of the amino groups) and chemical adsorption (the hemilability of the amine–metal bond), more HSNs were adsorbed on the metal surface so the anti-wear and friction reduction properties of the HSNs were enhanced. The C₆H₅ and CH₃ terminated HSNs show short running-in durations and smooth stable periods, which could be due to their better dispersity in PAO. With their non-polar groups, the C₆H₅ and CH₃ terminated HSNs would disperse better than the polar group terminated HSNs in PAO, which could prevent the aggregation more effectively and reduce the three body abrasion caused by large clusters of aggregated HSNs.

Fig. 10 COF versus time for different HSNs.

It can be seen from the test results that the COF values of the HSNs-PAO lubricants show a similar trend in which the COF values decrease slightly at first and are then interrupted with a dramatic increase, followed by a quick decrease. After that the COF would remain stable. The schematic diagram of the running-in process is shown in Fig. 11. When the counterparts are sliding against each other under the lubrication of HSNs-PAO, the HSNs would first polish the micro-bumps and fill the micro-grooves on the metal surface. This will improve the surface roughness of the two counterparts, and as a result, the COF goes down slightly during the first period. This point is confirmed by the SEM micrograph shown in Fig. 12(c). This period is the pre-running-in period. With the increasing numbers of the HSNs adsorbed on the metal surface, some of the HSNs would interact with the surface, leading to surface oxidation, passivation and formation of the transfer film. This process will cause wear and generate heat in the contact area. The COF would increase dramatically in this period. After the formation process of the protecting film, the metal surface was covered with the protecting film including an adsorbed layer and a chemically reacted layer. The wear rate and COF will decrease under the protection of the transfer film. Although the protecting film will be removed and reformed on a small scale on the metal surface when debris and micro bumps are passed by, the surface will be generally covered with the protecting film and the COF will be kept stable in the following period. This is confirmed in Fig. 12(e and f), in which both the cracking and the newly formed protecting layer are found on the protecting film.

Fig. 11 Schematic diagram of the running-in process including: (a) the pre-running-in period, (b) the running-in period and (c) the stable period.

Fig. 12 The SEM micrographs of a steel ball wear surface for the pre-running-in period (a, b and c) and the stable period (d, e and f). (b), (c), (e) and (f) are the high-magnification of marked regions in (a), (b), (d) and (e) respectively.

It should be noted here that an interesting running-in period shift was found in which the polar group (NH₂ and COOH) terminated HSNs would transfer from the pre-running-in period to the running-in period earlier than the non-polar group (CH₃ and C₆H₅) terminated HSNs. As reported by Chang et al.²⁶ solid lubricants with PTFE could form thicker transfer films, which resulted in a shorter running-in stage. Bosman et al.²⁷ reported that the mechanically altered layer at the top of the bulk material has an important influence on the running-in period. In our research, different running-in period durations and period transfer times were found. With the polar groups, the NH₂ and COOH terminated HSNs have more opportunities to adsorb on the metal surface due to the chemical and physical adsorption of the end groups, which would accelerate the period transfer. The NH₂ terminated HSNs show the fastest period transfer among all types of the HSNs. This might be attributed to the chelation effect of the amino functional groups and the hemilability of the amine–metal bonds that enhance the adsorption of the HSNs on the metal surface.

In order to further investigate the tribological mechanism of the HSNs, the wear surfaces were examined by SEM. Typical wear surfaces of the steel balls during the pre-running-in period and the stable period are shown in Fig. 12. Fig. 12(a–c) show a typical wear surface during the pre-running-in period (the steel balls were tested in a tribological experiment for 30 min) at different magnifications. Fig. 12(b) is the high-magnification micrograph of the marked region in Fig. 12(a) while Fig. 12(c) is the high-magnification micrograph of the marked region in Fig. 12(b). From Fig. 12(a) and (b), micro-grooves and nano-grooves are found on the wear surface. When we magnified the nano-grooves, we found a large amount of nanoparticles in the grooves. Comparing with the flat surface beside the nano-grooves which barely had any nanoparticles on it, the nanoparticles were successfully filled into the grooves. This micrograph fits well with the analysis of the pre-running-in period showing that the nanoparticles will fill the grooves, which would improve the surface roughness and reduce the COF. However, as the wear surface is still in the pre-running-in period, the groove is not fully filled with nanoparticles. With a longer time period, more nanoparticles would fill the grooves and the protecting film would then be generated on the wear surface. Fig. 12(d–f) show a typical wear surface after the running-in period, in which the steel balls were tested for 120 min. Fig. 12(e) is the high-magnification micrograph of the marked region in Fig. 12(d) while Fig. 12(f) is the high magnification micrograph of the marked region in Fig. 12(e). It can be seen from Fig. 12(d) that a large area of the protecting film was formed on the surface of the steel ball, which occupied most of the area of the micrograph. When we magnified the protecting film (shown in Fig. 12(e)), we found that micro-cracking had appeared on the protecting film, which means that the protecting film might be delaminated and removed by friction. However, when the marked region of Fig. 12(e) was further magnified, we found some silica nanoparticles on the edge of the protecting film. As shown in Fig. 12(f), some of the nanoparticles were unbroken and were just attached to the film while some of the nanoparticles were broken and were in the process of forming the protecting film. This confirmed the analysis that in the stable period, the protecting film would be removed and reformed at the same time.

The wear surface of the different kinds of HSNs were investigated by SEM and EDS. The SEM and EDS results of typical wear surfaces for the NH₂, COOH, C₆H₅ and CH₃ terminated HSNs are shown in Fig. 13(a, e), (b, f), (c, g) and (d, h) respectively. As can be seen from Fig. 13, dark areas on the wear surface were found after 30 minutes of wear, the NH₂ terminated HSNs show evenly distributed grey areas while the COOH terminated HSNs show spotted dark grey areas. Meanwhile, the C₆H₅ and CH₃ terminated HSNs show a lighter colour on the wear surfaces than the NH₂ and COOH terminated HSNs. The NH₂, C₆H₅ and CH₃ terminated HSNs show a smooth wear surface with shallow wear tracks while the COOH terminated HSNs show deep scratching lines on the wear surface, which indicate that the ploughing phenomenon has occurred seriously on the surface. The wear surfaces were then characterized by EDS. It can be seen from the figure that, besides the high peaks corresponding to Fe and Cr, the peak corresponding to Si was found at around 2 keV. The dark areas observed on the wear surfaces were found to contain higher amounts of Si than the lighter coloured areas, which suggests that the HSNs adsorbed on the dark areas had formed a chemical reaction layer. The wear surfaces of the NH₂ and COOH terminated HSNs show higher Si counts than the C₆H₅ and CH₃ terminated HSNs, which fits well with the interpretation that the polar groups would lead to more silica nanoparticle adsorption and a thicker chemical reaction layer on the metal surface.

Fig. 13 The SEM images and EDS spectra of wear scar surfaces for the different kinds of HSNs: (a and e) SEM and EDS for NH₂ terminated HSNs, (b and f) SEM and EDS for COOH terminated HSNs, (c and g) SEM and EDS for C₆H₅ terminated HSNs and (d and h) SEM and EDS for CH₃ terminated HSNs.

4. Conclusions

Hairy silica nanoparticles with four kinds of functional end groups were synthesized and dispersed into PAO lubricants via a four-step process. The HSNs were characterized by SEM, XRD, zeta potential measurements, DLS studies, XPS, TGA, BET theory and FTIR. The tribological properties of the HSNs were characterized using a four-ball tribometer. Based on the test results, the effect of the functional groups on the tribological properties of the HSNs was discussed. It was found that the NH₂ terminated HSNs exhibited the best anti-wear and friction-reduction properties due to the chemical and physical adsorption of NH₂ groups. However, the NH₂ terminated HSNs showed concentration sensitive behaviour, which could be attributed to the hydrogen bonding between the primary NH₂ groups which caused aggregation. The CH₃ and C₆H₅ terminated HSNs enhanced the stability of the tribological performance of the HSNs, which could be due to the non-polar end groups that increase the compatibility of the HSNs with the PAO lubricant. Moreover, the running-in phenomenon of the HSNs-PAO lubricants was investigated, three different periods of the running-in phenomenon (the pre-running-in period, running-in period and stable period) were proposed and discussed based on the experiment results. The nanoparticles would fill the nano-grooves on the wear surface and form protecting films during the running-in period. The polar functional group (COOH and NH₂) terminated HSNs show more adsorption of the HSNs on the metal surface, which could lead to the acceleration of the running-in process.

Acknowledgements

The author would like to thank Prof. Donald Koch, Prof. Lynden Archer and Rahul Mangal, Chemical and Biological Engineering, Cornell University, for their helpful discussions and suggestions.

Notes and references

1. C. A. Grabowski, H. Koerner, S. Meth, A. Dang, C. M. Hui, K. Matyjaszewski, M. R. Bockstaller, M. F. Durstock and R. A. Vaia, ACS Appl. Mater. Interfaces, 2014, 6, 21500–21509.
2. L. Casal-Dujat, M. Rodrigues, A. Yagüe, A. C. Calpena, D. B. Amabilino, J. Gonzaílez-Linares, M. Borraís and L. Peírez-Garciía, Langmuir, 2012, 28, 2368–2381.
3. X. Lin, W.-L.-J. Hasi, X.-T. Lou, S. Lin, F. Yang, B.-S. Jia, D.-Y. Lin and Z.-W. Lu, RSC Adv., 2014, 4, 51315–51320.
4. K. Smaali, S. Desbief, G. Foti, T. Frederiksen, D. Sanchez-Portal, A. Arnau, J. P. Nys, P. Leclère, D. Vuillaume and N. Clément, Nanoscale, 2015, 7, 1809–1819.
5. S. Ma, S. Zheng, D. Cao and H. Guo, Particuology, 2010, 8, 468–472.
6. N. Xu, M. Zhang, W. Li, G. Zhao, X. Wang and W. Liu, Wear, 2013, 307, 35–43.
7. Q. Xue, W. Liu and Z. Zhang, Wear, 1997, 213, 29–32.
8. H. Etemadi, A. Shojaei and P. Jahanmard, J. Reinf. Plast. Compos., 2014, 33, 166–178.
9. M. M. Saatchi and A. Shojaei, Mater. Sci. Eng., A, 2011, 528, 7161–7172.
10. A. A. Alazemi, V. Etacheri, A. D. Dysart, L.-E. Stacke, V. G. Pol and F. Sadeghi, ACS Appl. Mater. Interfaces, 2015, 5514–5521,  DOI:10.1021/acsami.5b00099 .
11. D. Kim and L. A. Archer, Langmuir, 2011, 27, 3083–3094.
12. X. Li, Z. Cao, Z. Zhang and H. Dang, Appl. Surf. Sci., 2006, 252, 7856–7861.
13. T. Jiang, T. Kuila, N. H. Kim, B. C. Ku and J. H. Lee, Compos. Sci. Technol., 2013, 79, 115–125.
14. S. G. Vilt, N. Martin, C. McCabe and G. Kane Jennings, Tribol. Int., 2011, 44, 180–186.
15. E. Amerio, P. Fabbri, G. Malucelli, M. Messori, M. Sangermano and R. Taurino, Prog. Org. Coat., 2008, 62, 129–133.
16. Y. Kang, X. Chen, S. Song, L. Yu and P. Zhang, Appl. Surf. Sci., 2012, 258, 6384–6390.
17. S. Song, R. Chu, J. Zhou, S. Yang and J. Zhang, J. Phys. Chem. C, 2008, 112, 3805–3810.
18. R. P. Bagwe, L. R. Hilliard and W. Tan, Langmuir, 2006, 22, 4357–4362.
19. T. Iwano and K. Kobayashi, IEEE Trans. Magn., 2005, 41, 3010–3012.
20. R. L. Jones, B. L. Harrod and J. D. Batteas, Langmuir, 2010, 26, 16355–16361.
21. Y.-C. Chen, H.-C. Lin and Y.-D. Lee, J. Polym. Res., 2004, 11, 1–7.
22. Y. Gao, G. Chen, Y. Oli, Z. Zhang and Q. Xue, Wear, 2002, 252, 454–458.
23. V. Jaiswal, R. B. Rastogi, R. Kumar, L. Singh and K. D. Mandal, J. Mater. Chem. A, 2014, 2, 375–386.
24. B. Yu, L. Qian, J. Yu and Z. Zhou, Tribol. Lett., 2008, 34, 1–10.
25. A. C. Hervé, J. J. Yaouanc, J. C. Clément, H. Des Abbayes and L. Toupet, J. Organomet. Chem., 2002, 664, 214–222.
26. L. Chang, Z. Zhang, L. Ye and K. Friedrich, Wear, 2007, 262, 699–706.
27. R. Bosman and D. J. Schipper, Tribol. Lett., 2011, 41, 263–282.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra22932d

---