Superlubricity of 1-Ethyl-3-methylimidazolium ... - ACS Publications

Apr 19, 2018 - The robust liquid superlubricity of a room-temperature ionic liquid induced by tribochemical reactions is explored in this study. Here ...
0 downloads 0 Views 4MB Size
Article Cite This: Langmuir XXXX, XXX, XXX−XXX

pubs.acs.org/Langmuir

Superlubricity of 1‑Ethyl-3-methylimidazolium trifluoromethanesulfonate Ionic Liquid Induced by Tribochemical Reactions Xiangyu Ge, Jinjin Li,* Chenhui Zhang, Zhongnan Wang, and Jianbin Luo State Key Laboratory of Tribology, Tsinghua University, Beijing 100084, China ABSTRACT: The robust liquid superlubricity of a room-temperature ionic liquid induced by tribochemical reactions is explored in this study. Here, 1ethyl-3-methylimidazolium trifluoromethanesulfonate ([EMIM]TFS) could realize stable superlubricity (μ < 0.01) with water at the interfaces of Si3N4/ SiO2. A superlow and steady friction coefficient of 0.002−0.004 could be achieved under neutral conditions (pH of 6.9 ± 0.1) after 600 s of runningin process. Various factors that could affect superlubricity were explored, including concentration of [EMIM]TFS, sliding speed, applied load, and volume of the lubricant. The results reveal that superlubricity can be achieved with [EMIM]TFS aqueous solution under a broad scope of conditions. The results of surface analysis show that a steady composite tribochemical layer comprising [EMIM]TFS, silica, ammonia-containing compounds, and sulfides was formed by tribochemical reactions between [EMIM]TFS and Si3N4 during the running-in period. The film thickness calculation reveals that the achieved superlubricity is in a mixed lubrication regime that comprises boundary lubrication and thin film lubrication. The superlubricity state is governed by a firm composite tribochemical layer, a molecular adsorption layer (electric double layer of [EMIM]TFS), and a fluid layer. The liquid superlubricity achieved by the ionic liquid is helpful for the development of new ionic liquids with superlubricity characteristics and is of great significance for scientific understanding as well as engineering applications.



INTRODUCTION Room-temperature ionic liquids (RTILs) are of great interest in both fundamental and applied research owing to a variety of excellent properties such as nonflammability, negligible volatility, wide electrochemical window, high electrical conductivity, and high thermal stability.1 These exclusive physicochemical properties of RTILs have boosted the current technologies and fostered the explosive growth in new technological applications. In the field of surface science and engineering, RTILs have shown superior performance both as lubricants2,3 and lubricant additives,4−6 including in the reduction of friction and wear of light alloys7,8 and polymers,9,10 and also as promoters of coatings,11 or as corrosion inhibitors.12,13 Because of these reasons, researchers continue to explore RTIL lubricants. When the friction coefficient (CoF) reduces to below 0.01, either by solid or liquid lubricants, a new lubrication state is introduced called superlubricity.14 At present, superlubricity is the prevalent research subject, as it is closely related to energy issues. Improvements in lubricants and their mechanism of superlubricity have been ceaselessly sought by researchers for energy efficiency and durability. Researchers have achieved superlubricity with various materials in both micro- and macroscales.15−19 However, superlubricity of RTILs has only been explored under a low applied load (20 mN) with carbon quantum dots,20 or in microscale: by lubricating contact pairs of © XXXX American Chemical Society

silica/graphite with 1-hexyl-3-methylimidazolium tris(pentafluoroethyl)trifluorophosphate, superlubricity can be in situ “switched” on and off by switching the potentials of the graphite surface. At positive potentials, friction can reach the superlubricity level when the ion layer is governed by anions.21 In macroscale, only a minimum CoF of 0.0001 with a mean friction value of 0.02 was achieved when a protic ionic liquid aqueous solution is used to lubricate sapphire/steel contact pairs.6 Therefore, robust superlubricity of RTILs in macroscale or at high applied loads needs to be explored urgently. In this study, we achieved macroscale superlubricity between Si3N4 and SiO2 interfaces by an RTIL aqueous solution (RTIL(aq)) under a wide range of conditions. Owing to the outstanding properties of RTIL(aq), including its neutrality, incredible lubricity, and stability under air atmosphere, it has the potential for engineering applications. The superlubricity mechanism achieved by this RTIL(aq) was investigated and analyzed intensively, also a possible superlubricity model is offered.



EXPERIMENTAL SECTION

Materials. Four kinds of RTILs were introduced in this study and their chemical structures are shown in Figure 1. 1-Ethyl-3Received: March 16, 2018 Revised: April 19, 2018 Published: April 19, 2018 A

DOI: 10.1021/acs.langmuir.8b00867 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

between the ball and disk varied from 0.05 to 0.25 m/s with an interval of 0.05 m/s. To ensure the accuracy of the testing results, the levelness of the tribotester was adjusted to achieve a similar CoF in both clockwise and counterclockwise rotating directions. Each sample was tested three times, and the accuracy of CoF was ±0.001. All tests were carried out at a temperature of 24−27 °C and 10−20% relative humidity. A standard rheometer (Physica MCR301, Anton Paar) and a pH probe (InLab Routine Pro, Mettler Toledo) were employed to measure the viscosity and pH value of each lubricant at 25 °C, respectively. The worn depth and the roughness of the worn zone were obtained with a three-dimensional white light interferometry microscope (ZYGO, Nexview). A scanning electron microscope (SEM, Quanta 200 FEG) was employed to observe the worn surface topographies under a high vacuum. Owing to the low electrical conductivity of the Si3N4 balls and SiO2 disks, they were platinumcoated to enhance their electrical conductivity and image contrast before SEM observation. Furthermore, X-ray photoelectron spectroscope (XPS) was employed to measure the elements present on the worn zone lubricated with [EMIM]TFS(aq) to analyze the chemical characteristics of the tribofilm.

Figure 1. Chemical structure of RTILs, [EMIM]TFS, [EMIM]BF4, [EMIM]DCA, and [EMIM]SCN. methylimidazolium trifluoromethanesulfonate ([EMIM]TFS, C7 H11 F3N 2O3 S), 1-ethyl-3-methylimidazolium tetrafluoroborate ([EMIM]BF4, C6H11BF4N2), and 1-ethyl-3-methylimidazolium dicyanamide ([EMIM]DCA, C8H11N5) with 98% purity were purchased from Aladdin Industrial Corporation, and 1-ethyl-3-methylimidazolium thiocyanate ([EMIM]SCN, C7H11N3S) with over 98% purity was purchased from Tokyo Chemical Industry Co., Ltd (TCI). No further treatment was applied to these RTILs. The RTILs(aq) were formulated via mixing RTILs with pure water, whose electrical resistivity is larger than 18 MΩ·cm, at a concentration of 40 wt % and abbreviated as 40% [EMIM]TFS(aq), 40% [EMIM]BF4(aq), 40% [EMIM]DCA(aq), and 40% [EMIM]SCN(aq). Tribology Tests. A Universal Micro-Tribotester (UMT-5, Bruker, Billerica, MA) with a rotating mode of ball-on-disk was used to perform the friction test. The upper Si3N4 ball with 4 mm diameter and 10 nm roughness (Ra) was obtained from Shanghai Research Institute of Materials, and the lower SiO2 disk has a roughness (Ra) of 5 nm. All of the balls and disks were sonicated in ethanol and acetone for 10 min, followed by douching with pure water and exsiccated in an air oven. For a single test, the lubricant was injected to the contact interface at a volume of 50 μL. The load applied to the contact pairs ranged from 2 to 4 N (600−750 MPa), and the relative sliding speed



RESULTS AND DISCUSSION Superlubricity of [EMIM]TFS(aq). First, the lubricity of pure water and pure RTILs ([EMIM]TFS, [EMIM]BF4, [EMIM]DCA, and [EMIM]SCN) were tested for comparison. Because of the inability of forming an effective lubricating film in only 600 s of the running-in process, CoF of pure water stayed larger than 0.45 throughout the test, whereas all four RTILs exhibited steady CoFs of 0.04−0.07, as shown in Figure 2a. The four kinds of 40% RTILs(aq) were also tested under the same conditions, as shown in Figure 2b. At the beginning of the test, all four RTILs(aq) exhibited very high friction (>0.1), as it is difficult for them to develop a hydrodynamic film at a relative low sliding speed and a high contact pressure based on Hamrock−Dowson (H−D) theory22 because of their low

Figure 2. CoFs of contact pairs of Si3N4/SiO2 lubricated (a) with pure water, [EMIM]DCA, [EMIM]SCN, [EMIM]BF4, and [EMIM]TFS; (b) with [EMIM]DCA(aq), [EMIM]SCN(aq), [EMIM]BF4(aq), and [EMIM]TFS(aq) (the concentration of RTILs is 40 wt %, “η” is the viscosity of RTILs(aq) in mPa·s), the inset is the partly detailed CoF from 600 to 1200 s; (c) with [EMIM]TFS(aq), the inset is the detailed CoF from 360 to 1800 s; and (d) with [EMIM]TFS(aq) with 12 h of suspension. The applied load was 3 N, and the sliding speed was 0.1 m/s. B

DOI: 10.1021/acs.langmuir.8b00867 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Figure 3. CoFs of Si3N4/SiO2 contact pairs lubricated (a) with [EMIM]TFS(aq) at various sliding speeds under an applied load of 3 N, the inset shows the relationship between sliding speeds and average CoFs during superlubricity stage; (b) with [EMIM]TFS(aq) under various applied loads at a sliding speed of 0.1 m/s, the inset shows the ball worn zone diameter; (c) with [EMIM]TFS(aq) at various concentrations ranging from 10 to 80% by weight, the inset shows the relationship between average CoF during superlubricity stage and concentration of [EMIM]TFS; and (d) with [EMIM]TFS(aq) at various volumes ranging from 40 to 60 μL with an interval of 5 μL, TSL in the inset represents the time required to reach superlubricity from the beginning of the test. For (c,d), the applied load was 3 N and sliding speed was 0.1 m/s.

when the sliding speeds were 0.1, 0.15, and 0.2 m/s, the CoFs all reduced to less than 0.005, and the CoFs (0.002) were obtained at a speed of 0.1 m/s, as shown in the inset of Figure 3a. Thus, it is concluded that in the range of 0.1−0.2 m/s, superlubricity could be realized with [EMIM]TFS(aq). As for the effect of applied load on superlubricity, it could be seen from Figure 3b that under 2, 3, and 4 N (contact pressures of 600, 700, and 750 MPa), [EMIM]TFS(aq) could achieve superlubricity after 300−500 s of running-in period with average CoFs of 0.0018, 0.002, and 0.0025 during superlubricity stage, respectively. Additionally, the diameters of the worn zone on these balls under 2, 3, and 4 N were 302, 321, and 367 μm, measured by a 3D white light interferometry microscope, corresponding to the contact pressures of 28, 37, and 38 MPa, respectively. On the basis of these results, it is clear that the final CoF becomes lower with the decrease in contact pressure. Therefore, it can be concluded that the reduction in contact pressure is advantageous to superlubricity, which agrees with our previous work reasonably well.23 The superlubricity behavior of the concentration of [EMIM]TFS was studied to further understand the superlubricity mechanism of [EMIM]TFS. As shown in Figure 3c, when the concentration of [EMIM]TFS increased from 10 to 60%, superlubricity could be achieved for all concentrations, with a running-in period decreasing from 900 to 300 s correspondingly, and the minimum CoF (0.002) was obtained at a concentration of 40%. However, when the concentration of [EMIM]TFS exceeded 80%, its CoF reduced to 0.08 after a short running-in period of 150 s and then further reduced to 0.03 gradually, indicating that superlubricity cannot be achieved. In addition, the initial viscosities of the lubricants at various concentrations showed an increasing trend, ranging

viscosities (close to water in Figure 2b). It could be assumed that the boundary lubrication was the main mechanism at the initial stage of the friction test. During the running-in period, the CoFs with the lubrication of [EMIM]BF4(aq), [EMIM]DCA(aq), and [EMIM]SCN(aq) quickly reduced to a very low level (0.015−0.02) and then remained stable. The superlubricity could not be achieved by these RTILs(aq) because their final CoFs were always above 0.01. However, the CoF with the lubrication of [EMIM]TFS(aq) gradually reduced to 0.002 after 300 s of running-in process, and the CoF remained stable with only a slight increase (0.002−0.004) for more than 1 h, as shown in Figure 2c. Because these RTILs have the same cations, it can be inferred that the superlubricity of [EMIM]TFS(aq) is mainly owing to its anions, which may lead to the formation of a firm and effective lubricating film via a possible tribochemical reaction between its anions and the friction surfaces during the running-in period. Additionally, the superlubricity was still realized after the experiment was halted for 12 h in the absence of shear and pressure, and the CoFs before and after suspension were nearly the same, as shown in Figure 2d, indicative of the robustness of the lubricating film in its natural state. To study the superlubricity behavior of [EMIM]TFS at various sliding speeds, experiments were performed at a range of sliding speeds from 0.05 to 0.25 m/s with an interval of 0.05 m/s (Figure 3a). At a sliding speed of 0.05 m/s, the CoF of 40% [EMIM]TFS(aq) could only reduce to around 0.013 after 600 s of running-in. Analogously, at a sliding speed of 0.25 m/s, the lubrication state of 40% [EMIM]TFS(aq) was at the edge of superlubricity regime (0.01) after 200 s of the running-in period. These results indicate that too high or too low sliding speeds could lead to the failure of superlubricity. However, C

DOI: 10.1021/acs.langmuir.8b00867 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

μm), and the worn zone of the disk was rough with many grooves (5−14 μm) and wear pits (5−10 μm). To study the chemical characteristics of the tribochemical layer and prove our hypothesis, a XPS spectroscope was employed to probe the chemical elements on the worn zone of Si3N4 ball, as shown in Figure 5. The spectrum of C 1s has three individual peaks (Figure 5a), indicating the presence of both cations and anions of [EMIM]TFS. According to Figure 1, the carbon atoms in the EMIM cations form four different chemical bonds: aliphatic C−C chain (C1, at 284.8 eV),25 aliphatic C−N (C2, the carbon atom bonded to only one nitrogen atom), and aromatic C−N in the imidazolium ring (C3, the carbon atom bonded only to one nitrogen atom and C4, the carbon atom bonded to two nitrogen atoms). The peak located at 292.7 eV is assigned to the carbon atom in C−F of the TFS anions (C5).26−28 The aliphatic C−C chain (C1) and the C−F bond (C5) are clearly observed in the spectrum, while the binding energies of the three C−N bonds are so close (285−288 eV) that they form one peak shoulder containing three unresolved overlapping components.26 Nitrogen atoms are only present in the imidazolium ring of [EMIM]TFS. If the cations are presented in the tribofilm, then it is expected that at least two N 1s peaks should be observed, one from the Si3N4 ball substrate and the other from the imidazolium ring of the deposited compound. As shown in Figure 5b, the N 1s spectrum contains three peaks at 397.6 eV (Si3N4 the ball), 400.3 eV (N−H bond), and 402 eV (N1 in the imidazolium ring).26−29 In terms of the N−H bond, the Si3N4 ball and water could react on the ball surface within the running-in process, as described in eq 1.30,31 In addition, the deposition of ammonia-containing compounds, which may be formed by the rubbing motion of surfaces,32 and the degradation of imidazolium moiety could led to the generation of a N−H bond peak in the spectrum.

from 1.1 mPa·s (10%) to 6.3 mPa·s (80%), and there was an obvious reduction in the running-in period (from 900 s (10%) to 150 s (80%)) with the increasing concentration. Moreover, the superlubricity could not be achieved when the running-in period was too short (e.g., 150 s). Compared with these results, it can be concluded that the concentration of [EMIM]TFS in its aqueous solution could dominate the duration of the running-in process by changing the viscosity of the lubricant and thereafter affected the superlubricity momentously. To investigate the role of volume of water in superlubricity, experiments were performed with various volumes of [EMIM]TFS(aq). As shown in Figure 3d, superlubricity could be achieved with all five volumes of [EMIM]TFS(aq) after a running-in period. In addition, it could be seen that the duration of the running-in process was extended with the increase in volume, for example, 340 s for 40 μL and 490 s for 60 μL (Figure 3d inset). Because the evaporation rate of water is constant approximately,24 it could be regarded that the extended running-in process is caused by the evaporation of extra water to reach a certain water concentration on the friction surface. It therefore can be inferred that there is a key concentration of water for superlow friction, which is consistent with our previous work.24 Discussion. To probe the specificity of [EMIM]TFS, the topography details of the worn zone lubricated with [EMIM]TFS(aq) was obtained by a 3D white light interferometry microscope and SEM, as shown in Figure 4. The diameter and

Si3N4 + 6H 2O = 3SiO2 + 4NH3

(1)

Oxygen is present in the TFS anions in only one chemical bond. However, the O 1s spectrum shown in Figure 5c exhibits only one peak that can be divided into two peaks, located at 532.4 and 532.9 eV, respectively. The former binding energy is attributed to the oxygen in the TFS anions,26−28 whereas the latter is assigned to SiO2 formed from the reaction between the Si3N4 ball and water. In Figure 5d, the peak of Si 2p located at 102.3 eV also confirmed the presence of SiO2.33 In the F 1s spectrum shown in Figure 5e, a peak located at a wide range of around 688.9 eV indicates the presence of C−F because of the deposition of TFS anions. In the S 2p spectrum, there is one peak at approximately 169 eV because of the TFS anions,28 which also indicates that anions are deposited on the worn surface. In addition, the spectrum reveals that there is in fact an additional S 2p peak at a lower binding energy of approximately 161 eV, which indicates the presence of sulfides on the worn surface.3,34−37 To confirm that the sulfides were productions of tribochemical reactions, one simple test was conducted. Two precleaned Si3N4 balls were dipped into [EMIM]TFS(aq) for 30 min, and one of them was washed with pure water and dried, whereas the other one was dried without washing. Then, the XPS spectra of these two Si3N4 balls were measured and noted as “dip no wash” and “dip & wash,” respectively, as shown in Figure 5f. The S 2p spectrum of the dip no wash ball has just one peak at approximately 169 eV, indicating the presence of SO3 derived from the TFS anions, as mentioned before, and no peak at approximately 161 eV was

Figure 4. (a,b) 3D white light interferometry microscopy images of the worn zone lubricated with 40% [EMIM]TFS(aq), the inset shows the worn surface profile of the ball and disk; and (c,d) SEM images of worn surfaces lubricated with 40% [EMIM]TFS(aq). The applied load was 3 N and sliding speed was 0.1 m/s.

roughness of the worn zone of the Si3N4 ball were 321 μm (wear rate is (4.8 ± 1.2) × 10−7 mm3/N·m) and 14.4 nm, respectively, as shown in Figure 4a. The depth and roughness of the worn zone of the SiO2 disk were 0.47 μm (wear rate is (4.4 ± 1.7) × 10−6 mm3/N·m) and 16.2 nm, as shown in Figure 4b. The details of the worn zone under high magnification are shown in Figure 4c,d. The worn zone of the ball was smooth with a few cracks and small wear pits (