Achieving macroscale superlubricity with ultra-short running-in period by using polyethylene glycol-tannic acid complex green lubricant

Superlubricating materials can greatly reduce the energy consumed and economic losses by unnecessary friction. However, a long pre-running-in period is indispensable for achieving superlubricity; this leads to severe wear on the surface of friction pairs and has become one of the important factors in the wear of superlubricating materials. In this study, a polyethylene glycol-tannic acid complex green liquid lubricant (PEG10000-TA) was designed to achieve macroscale superlubricity with an ultrashort running-in period of 9 s under a contact pressure of up to 410 MPa, and the wear rate was only 1.19 × 10−8 mm3·N−1·m−1. This is the shortest running-in time required to achieve superlubricity in Si3N4/glass (SiO2). The results show that the strong hydrogen bonds between PEG and TA molecules can significantly reduce the time required for the tribochemical reaction, allowing the lubricating material to reach the state of superlubrication rapidly. Furthermore, the strong hydrogen bond can share a large load while fixing free water molecules in the contact zone to reduce shear interaction. These findings will help advance the use of liquid superlubricity technology in industrial and biomedical.


Introduction
Friction is very common in our daily life and industrial processes; however, unnecessary friction and wear cause massive energy consumption and economic losses. For example, the loss caused by friction accounts for approximately 5%-7% of the gross national product in many countries [1]. Many researchers have focused on developing and designing lubricating materials with ultralow friction and anti-wear properties [2][3][4]. Shinjo and Hirano et al. [5,6] first proposed superlubricity as a lubrication state in which the friction force disappears entirely in theory. Due to the interference of the external environment and the limitations of testing technology, the lubrication state when the friction coefficient below 0.01 is called superlubricity [7]. Increasing efforts have been devoted to superlubricity because it provides a feasible solution to save energy and reduce economic losses and is one of the essential technical means to help achieve carbon neutrality [8].
According to the physical form of lubricating materials, superlubricity can be classified into solid superlubricity and liquid superlubricity [9,10]. A series of two-dimensional (2D) lamellar materials (e.g., MoS 2 [11], graphene [12], graphite [13], hexagonal boron nitride [14], and black phosphorus [15]) can achieve solid superlubricity due to a weak interlaminar force or incommensurate contact [16]. However, research on the superlubricity of 2D materials is currently limited chiefly to the nano and micro scales [17]. The state of superlubrication can be achieved only when the atomic force microscope (AFM) probe or the micromanipulator slides on the 2D materials down a specific direction [13,18]. Moreover, sample preparation and experimental operation for achieving superlubrication at the micro and nano scales are very complicated [19,20]. In addition, achieving solid superlubricity at the macroscale remains a challenge as it requires unique friction pair materials [12,21] or structural design [22][23][24] and special environmental conditions (vacuum, inert gas) [12,22]. Some solid superlubricating materials pose the problems of long running-in periods and short working life. In summary, the aforementioned problems limit the practical applications of solid superlubricity technology at the macroscale.
Compared with solid superlubricity, macroscale liquid superlubricity is easier to achieve. Many liquid lubricants, such as water [25,26], hydrated ions [27][28][29], polymer brushes [30,31], ionic liquids [32,33], aqueous solutions of alcohols containing acids [34,35] or 2D materials [36][37][38], and biological mucus [39,40], have been proposed to achieve macroscale superlubricity in ambient atmosphere. The superlubricity mechanism of the abovementioned lubricating materials mainly includes hydration, electric double layer effect, hydrodynamic lubrication, hydrogen-bonded network, and the formation of molecular brushes [10]. Tomizawa and Fischer [25] found for the first time that superlubricity can be achieved on the surface of Si 3 N 4 /Si 3 N 4 by using water as a lubricant; the friction coefficient gradually decreases to 0.002 after a running-in period. Luo's group [41][42][43] has conducted extensive research in the field of liquid superlubricity and achieved many representative results. They demonstrated that phosphoric acid solution can achieve superlubricity between different friction pairs (Si 3 N 4 /glass, Si 3 N 4 /SiO 2 , Si 3 N 4 /sapphire, and sapphire/ruby). In addition, a new type of liquid superlubricity system has been established using a mixed solution of polyhydroxy alcohol and acid [34,35,44]. Moreover, to reduce the wear of friction pairs and improve the load capacity of the liquid superlubricity system, various relatively high-cost 2D nanomaterials have been adopted as additives in liquid lubrication [8,36,37]. However, the presence of a strong acid will result in severe corrosion of friction pairs, pollution of the environment, and certain risks to users. The long running-in period for liquid superlubrication may be the primary cause of friction pair wear. For example, the wear rate of SiO 2 disk is as high as (4.4 ± 1.7) × 10 −6 mm 3 ·N −1 ·m −1 and a running-in period is 300 s with 1-ethyl-3-methylimidazolium trifluoromethanesulfonate ([EMIM]TFS) lubrication [32]. Therefore, designing green liquid lubricant materials with superlubricity properties and ultra-short running-in periods is a great challenge and of great significance [45].
In this study, by combining a natural and environmentally friendly weak acid, tannic acid (TA), with biocompatible polyethylene glycol (PEG), a new green liquid lubricant (PEG10000-TA) was developed. PEG10000-TA contains a large number of hydrogen bond network structures and can quickly form a lubricating layer on the surface of the friction pair. Thus, the macroscale superlubricity of PEG10000-TA lubricant was successfully achieved within an ultrashort running-in period (~9 s) by using a Si 3 N 4 ball and glass substrate as friction pairs. In addition, the anti-wear properties of friction pairs can be improved using these green lubricants. Furthermore, the experimental factors and possible mechanisms of superlubricity were systematically studied in detail. This work provides a new strategy for developing superlubricating materials with an ultra-short running-in period and will promote the practical application of superlubricating materials. materials. PEG10000 and TA were added to deionized water (resistivity > 18 MΩ·cm) in a particular proportion. As shown in Fig. 1, PEG10000-TA green liquid lubricant was prepared by mixing PEG10000 and TA solution by using the one-pot method. TA is a natural and environmentally friendly weak acid widely found in immature fruits, such as persimmons, grapes, and apples ( Fig. 1(a)). Furthermore, because its molecular structure is rich in hydroxyl, TA can be quickly adsorbed to the surface of materials [46]. PEG is a green lubricant widely used in the biological and medical fields ( Fig. 1(b)); it can form a hydrogen bond network structure with TA molecules [47]. The abovementioned composition of the lubricant fully ensures the green environmental protection and use safety of PEG10000-TA, as shown in Fig. 1(c).

Tribological experiments
Tribological experiments were performed on a ball-on-disk tribometer (TRB, Anton Paar, Austria) in a linear reciprocating mode. The upper friction pair is a Si 3 N 4 ball with a diameter of 6 mm (R a = 12 nm), and the lower friction pair is an ordinary glass sheet (R a = 5 nm, thickness = 1.2 mm). Before the test, the two friction pairs were ultrasonically treated in ethanol and acetone solution for 15 min, repeatedly washed with deionized water, and then dried in an inert atmosphere. In each experiment, 30 μL of lubricant was dripped onto the contact area of the upper and lower friction pairs. In the friction test, the range of load applied was 2-6 N. The amplitude of linear slip was 2 mm, and the range of linear speed was 0.010-0.030 m/s (corresponding frequency was 1.59-4.77 Hz). All tribological experiments were performed three times to ensure the repeatability of the experimental results. The relative humidity was 20%-50% during the entire test period, and the ambient temperature was approximately 25 °C.

Characterization methods
The viscosity and pH value of each liquid lubricant at 25 °C were measured using a rotary rheometer (MCR302, Anton Paar) and a pH meter (PH100A, LiChen Tech), respectively. To determine the interaction between PEG10000 and TA molecules in lubricating materials, infrared spectra were obtained using a Fourier transform infrared spectrometer (FT-IR, PerkinElmer, Frontier). After the friction test, the glass substrate was cleaned with ethanol and deionized water for 20 min. The morphology of the worn area on the glass sheet was observed by optical microscopy (Axioscope 5, ZEISS) and field emission scanning electron microscopy (FE-SEM, JSM-7610F, JEOL). The three-dimensional (3D) morphology, depth, and wear volume of the wear scar on the surface of the glass sheet were studied using a noncontact optical 3D surface profilometer (MicroXAM-800, KLA-Tencor). In addition, the corresponding element distributions in the wear tracks were studied using an energy dispersion spectrometer (EDS, NORAN System 7, Thermo Scientific) equipped with FE-SEM. Moreover, the wear scars of the glass sheet were analyzed by X-ray photoelectron spectroscopy (XPS, Nexsa, Thermo Scientific) to study the chemical properties of the friction film. The glass sheet surface's contact angle (CA) was measured using a DSA100 CA meter (Kruss www.Springer.com/journal/40544 | Friction Company). The adsorption behavior of lubricant on the surface of gold chips (QSX-301, Q-sense AB) was studied by using a quartz crystal microbalance (QCM, Biolin Scientific, Sweden). The QCM experimental temperature was 25 °C, and the flow rate of the peristaltic pump was 100 μL/s.

Macroscale superlubricity behavior of PEG10000-TA
A schematic of the test equipment is illustrated in Fig. 2(a). The friction coefficients of pure water, TA solution, and PEG solution are shown in Fig. 2(b); the friction coefficient of pure water is as high as 0.781 after the 1,200 s friction test, which indicates that no lubricating films can be formed for pure water under high contact pressures in a short time [48]. The friction coefficient of TA solution (2 wt%) is 0.463, which is much lower than that of pure water, indicating that TA has potential in the field of water lubrication. However, the reduction in the friction coefficient is still not satisfactory. In contrast, PEG10000 aqueous solution (45 wt%), a well-known lubricant, can achieve a lower friction coefficient (0.012) after a long running-in | https://mc03.manuscriptcentral.com/friction period due to hydrodynamic lubrication [49]; however, it still cannot achieve macroscale liquid superlubricity in a short time. Compared with the three aforementioned lubricants, the friction coefficient of PEG10000 (45 wt%) and TA (2 wt%) mixed aqueous solution (PEG10000-TA) is less than 0.01 after only 9 s of running-in period and is stable at 0.005 after 1,200 s, as shown in Fig. 2(c). More importantly, by continuously supplementing PEG10000-TA lubricants (30 μL/h), the superlubricity state can be maintained for at least 12 h, indicating the robustness of the macroscale superlubricity ( Fig. 2(e)). Figure 2(d) shows a comparison of the superlubricity running-in period reported in this and previous works.
[EMIM]TFS ionic liquids [32] achieve superlubricity after a running-in period of 300 s for Si 3 N 4 /SiO 2 friction pairs. Moreover, when graphene-oxide nanoflakes and ethanediol mixed aqueous solution (GONFS-EDO) [36] are used as the lubricant, it takes a longer time (600 s) to achieve a superlubrication state. This means that the friction pairs may suffer from severe wear in case of a long running-in period. In the present study, the running-in period of liquid superlubricity was dramatically shortened to 9 s by using PEG10000-TA, which is a decrease of 30 times compared to that achieved in previous studies. This is by far the shortest running-in time (9 s) for Si 3 N 4 /glass (SiO 2 ) to achieve superlubrication ( Fig. 2(d)) [32,36,38,44,50,51]. In addition, Fig. S1 in the Electronic Supplementary Material (ESM) shows that the PEG10000-TA can also realize fast superlubrication between pure SiO 2 disc and Si 3 N 4 ball and the impurities in the glass do not affect the realization of above superlubrication. Combining the characteristics of an ultralow friction coefficient, ultrashort running-in period, and weak acidity (pH = 5.58), PEG10000-TA can greatly reduce the surface wear of friction pairs.
To demonstrate that the wear of friction pairs can be significantly reduced by fast superlubrication, the wear areas of Si 3 N 4 balls and glass sheets were characterized in detail by optical microscopy and a noncontact 3D surface profilometer (Fig. 3). When pure water, TA solution, and PEG10000 solution were used as lubrication materials, wear marks were evident on the surface of the glass after the friction test (Figs. 3(a)-3(c)), and the depths were 16.24, 10.73, and 9.92 μm, respectively (Figs. 3(e)-3(g)). In contrast, a negligible wear scar was noted on the surface of the glass sheet lubricated by PEG10000-TA, and the depth was only about 28.2 nm. In addition, the wear rate of the glass corresponding to PEG10000-TA was only 1.19 × 10 −8 mm 3 ·N −1 ·m −1 , which is 2-4 orders of magnitude lower than that of the other three lubricating materials (pure water: 1.12 × 10 −4 mm 3 ·N −1 ·m −1 , TA solution: 7.26 × 10 −5 mm 3 ·N −1 ·m −1 , PEG10000 solution: 5.79 × 10 −5 mm 3 ·N −1 ·m −1 ) (Fig. 3(m)). Furthermore, the wear rate in our work is only 1% of that of a reported superlubricating material, [EMIM]TFS ((4.4 ± 1.7) × 10 −6 mm 3 ·N −1 ·m −1 ), with running-in time of 300 s [32]. Moreover, the wear scar diameter of the Si 3 N 4 ball lubricated by PEG10000-TA is only about 111.42 μm, much smaller than that of other lubricants ( Fig. 3(n)). According to the diameter of the wear area of the Si 3 N 4 ball after the friction experiment, the final contact pressure corresponding to PEG10000-TA was calculated as 410 MPa, which is higher than that of the liquid lubricant used in other studies ( Fig. 2(d)). These wear data further demonstrate that PEG10000-TA has excellent anti-wear performance because it can achieve macroscale superlubrication in an ultra-short time. Moreover, it can be seen from Fig. S2 in the ESM that the diameter of the wear area of the Si 3 N 4 ball does not change much after the friction experiment at different times, which indicates almost no wear during the superlubricity period. In summary, the green liquid lubricant developed in this study, PEG10000-TA, effectively solves the problem that friction pairs undergo serious wear before realizing superlubricity [8].

Influence of chemical component content on tribological performance
To further optimize the friction performance of the lubricant, the effect of the amount of PEG and TA in PEG10000-TA on the friction coefficient was systematically studied (Fig. 4). As shown in Fig. 4(a), the friction coefficient cannot be reduced to 0.01 after 1,200 s of friction with the mass fraction of PEG10000 in the range of 10-40 wt%. When the PEG content is increased to 45 wt%, the friction coefficient of PEG10000-TA is less than 0.01 after a short running-in period of 9 s to achieve macroscale liquid superlubricity. As the basic lubricant, PEG plays a key role in the realization of superlubricity. When the content of PEG10000 is relatively low (10-30 wt%), the viscosity of the corresponding lubricating material is relatively low (Fig. 4(b)), resulting in poor film-forming properties and thus inferior lubrication performance. Owing to the relatively high viscosity (740 mPa·s) [51], the friction resistance and friction energy consumption may increase when the PEG content in the lubricant increases to 50 wt%. As a result, the average friction coefficient increases slightly to approximately 0.008. Thus, the material containing 45 wt% PEG is chosen as the best option. Furthermore, as shown in Figs. 4(c) and 4(d), corresponding friction experiments were conducted to investigate the effect of TA on superlubricity. The results show that superlubricity can be realized quickly when the content of TA in the lubricating material PEG10000-TA is 2-4 wt%. For lubricants with other mass fractions of TA (ω > 4 wt% or ω < 2 wt%), the friction coefficients are all higher than 0.01 in the friction test of approximately 1,200 s. In addition, considering that the friction pair may get corroded in an acidic environment, PEG10000-TA solutions containing 2 wt% TA (pH = 5.58) were adopted as the superior options for our next experiment.

Effect of applied load and linear sliding speed on superlubricity
To study the effect of applied load and linear sliding speed on superlubricity, parallel friction experiments  were conducted under different loads and sliding speeds, as shown in Fig. 5. Figures 5(a) and 5(b) show that the range of the applied load is 2-4 N (the corresponding initial Hertzian contact pressure is 559-704 MPa), and the friction coefficient of PEG10000-TA is lower than 0.01 after a short running-in period. Moreover, the corresponding average friction coefficient increases with increasing applied load. The superlubrication state of PEG10000-TA can only be maintained for a short time (~680 s) at an applied  www.Springer.com/journal/40544 | Friction load of 5 N; the friction coefficient subsequently increases rapidly to 0.02. This may be since the lubricating film formed on the surface of the friction pair fails to bear the excessive load, which leads to the abrupt change of the friction coefficient. When the load increases to 6 N, the friction coefficient reaches a higher value of 0.02-0.03. Therefore, PEG10000-TA can achieve liquid superlubricity only when the initial contact pressure is less than 704 MPa. Figure 5(c) presents the typical friction coefficient curve at different linear sliding speeds, and Fig. 5(d) displays the changing trend of the average friction coefficient of PEG10000-TA with increasing sliding velocity. The average friction coefficients of PEG10000-TA are only 0.031 and 0.026 under sliding speeds of 0.010 and 0.015 m/s, respectively, indicating that superlubricity cannot be achieved at low speeds. When the linear speed of the friction test is increased to 0.025 m/s, the calculated average friction coefficient is approximately 0.005 because the increase in sliding speed accelerates the formation of lubricating film and is more conducive to hydrodynamic lubrication, thereby improving the lubrication performance of PEG10000-TA [48]. When the sliding speed is further increased to 0.030 m/s, the corresponding average friction coefficient slightly increases to 0.008. These experimental results reveal that 0.020-0.030 m/s is a suitable sliding speed range for the superlubricity of PEG10000-TA. Therefore, a load of 4 N and a linear sliding speed of 0.025 m/s were selected as the best conditions for other tribological tests.

Specificity of PEG10000 and TA for superlubricity
In previous studies, composite lubricants based on PEG with lower molecular weights (200, 300, and 400) have achieved superlubricity, and the lubrication effect of PEG with lower molecular weight is better [44,48]. However, the molecular weight of PEG used in this study was approximately 10,000; thus, the superlubricity achieved by PEG10000-TA is different from that achieved by materials containing lowmolecular-weight PEG. To study the effect of highmolecular-weight PEG on superlubricity, the tribological properties of lubricating materials were tested using different molecular weights of PEG, as shown in Figs. 6(a) and 6(b). Interestingly, the friction coefficients of lubricants containing PEG with low molecular weight (400, 600, 800, 1000, and 2000) are all higher  | https://mc03.manuscriptcentral.com/friction than 0.01 in the 1,200 s running-in process. Moreover, with the increase in the molecular weight of PEG, the average friction coefficient of the mixed solution of PEG and TA decreases continuously. The PEG10000-TA solution prepared using PEG with a molecular weight of 10,000 can quickly achieve superlubricity, and the average friction coefficient is approximately 0.005. Therefore, when the average molecular weight of PEG is less than 10,000, the higher the average molecular weight of PEG, the better the tribological properties of the corresponding composite lubricants. This may be due to the stronger interaction between high-molecular-weight PEG and TA molecules, which guarantees the realization of superlubricity. Unfortunately, with the increase in the molecular weight of PEG, the viscosity of the mixed solution increases greatly. The exceptionally high viscosity of PEG20000-TA (5,800 mPa·s) increases the friction resistance and the corresponding average friction coefficient increases to approximately 0.009.
To clarify the role of TA in superlubricity, PEG10000 solution blended with TA and other acids (i.e., boric acid, phosphoric acid, oxalic acid, and citric acid) as additives were prepared. The lubrication behavior and pH value of these lubricants were investigated (Figs. 6(c) and 6(d)). During the 1,200 s friction process, the friction coefficient of TA-based composite lubricant (PEG10000-TA) is considerably lower than that of other acid-based lubricants. Even boric acid and phosphoric acid [41,44] cannot achieve superlubricity when combined with PEG10000. These results show that liquid superlubricity can be achieved only with the specific synergistic action of TA and PEG with a molecular weight of 10,000. In addition, compared with other strong acids, the weak acidity of PEG10000-TA (pH = 5.58) does not cause severe corrosion of the friction pairs (Fig. S3 in the ESM) and is relatively safe for users.

Superlubrication mechanism of PEG10000-TA mixed solution
In order to reveal the mechanism of rapid realization of superlubricity by PEG10000-TA, the morphology of the wear area on the glass surface was characterized by optical microscopy and FE-SEM, and the corresponding element distribution of the wear track was determined by EDS (Fig. 7). Moreover, to enhance the electrical conductivity and make the characterization image clearer, a 6-nm-thick layer of gold was plated on the surface of the glass before the test. As shown in Fig. 7(a), after approximately 20 min of the friction test, the wear tracks on the surface of the friction pair www.Springer.com/journal/40544 | Friction lubricated with pure water are pronounced. Upon further magnifying the dotted frame position, many deep grooves are found (Fig. 7(e)), indicating that water as a lubricant causes severe wear. Although the wear marks become shallower after lubrication with TA solution and PEG solution (Figs. 7(b) and 7(c)), grooves and abrasion can still be observed upon magnifying the wear marks (Figs. 7(f) and 7(g)).
Contrary to the above lubricants, the wear area on the surface of the glass sheet when using PEG10000-TA is relatively smooth. No obvious wear marks can be found (Fig. 7(d)), except for slight wear and pits (Figs. 7(h) and 7(i)). The results show that PEG10000-TA dramatically reduces the wear of friction pairs caused by the long running-in period of current superlubricity materials. The EDS mapping results of the wear position shown in Fig. 7(j) demonstrate that the concentration of PEG10000-TA lubricant is higher in the wear position on the glass sheet surface than in other regions. Therefore, PEG10000-TA is easily adsorbed at the contact position of the glass surface due to the octopus-like polyphenolic claws of TA (Fig. S4 in the ESM). Furthermore, the grooves caused by friction can also serve as a repository of lubricants to prevent further wear in the contact area [38,52].
To further investigate the chemical composition of the friction film, XPS spectra on the wear scars of the glass surface were obtained. All the high-resolution XPS photoelectron spectra were calibrated by C 1s of 284.8 eV, and the peaks were separated by Gaussian-Lorentzian fitting [52]. As shown in Fig. 8(b), Si 2p peaks located at 102.4 and 101.8 eV are attributed to Si-O bonds and Si-N bonds, respectively [50,53]. The Si-N bond may be due to the formation of a transfer film on the glass surface by Si 3 N 4 during friction. In the spectrum of C 1s, three independent peaks are observed at 284.8, 286.6, and 288.6 eV, proving the existence of C-C or C-H bonds, C-O bonds, and C=O bonds, respectively [44,54,55]. These carbon-containing chemical groups in the lubricant could be derived from PEG or TA (Fig. 8(c)). The N 1s spectrum has two peaks: 399.7 eV (N-H bond) and 397.2 eV (Si-N bond) (Fig. 8(d)) [32,44,56]. The existence of the N-H bond is attributed to the chemical reaction between Si 3 N 4 and H 2 O during friction which can be expressed as Eq. (1) [25,57]. The peak of O 1s at 532.5 eV is attributed to Si-O bonds in the glass itself or the C=O bonds in the lubricant on the glass surface [15]. Moreover, the appearance of C-O (532.9 eV) and O-H (536.0 eV) chemical bonds in the O 1s spectrum further proves the presence of PEG10000-TA in the wear area ( Fig. 8(e)) [50].
Based on the above analysis, a possible superlubricity model for PEG10000-TA mixed solution is proposed, as shown in Fig. 9(a). As previously reported [34,35,41,58], the superlubricity model in the current study also has a three-layer structure. First, a layer of negatively charged silica sol [25,57], namely Stern layer, forms on the surface of friction pair materials due to tribochemical reaction [32] (as shown in Eqs. (2) and (3)). More importantly, the actual thickness of the liquid film between the contact surfaces during the friction process is calculated to be about 84 nm by the Hamrock-Dowson formula (Eq. (S1) in the ESM).
| https://mc03.manuscriptcentral.com/friction Moreover, according to the thickness-roughness ratio (λ = 5.0 > 3), the lubrication state of the above superlubrication stage belongs to full-film lubrication (hydrodynamic lubrication), which effectively avoids the peak contact between the two contact surfaces ( Fig. 9(b)). It is precisely because of the strong hydrogen bonds that the prepared PEG10000-TA lubricant has relatively high viscosity (500 mPa·s) compared with other superlubricating materials, which is more conducive to the realization of hydrodynamic lubrication ( Fig. 9(c)). Thus, the full-film lubrication (hydrodynamic lubrication) due to the relatively high viscosity of prepared lubricant PEG10000-TA may be an important reason for the short running-in time and low wear rate in this paper [48,59,60]. In addition, due to larger load sharing, this strong hydrogen bond network structure effectively ensures that the PEG10000-TA lubricant can still achieve superlubrication even under contact pressure up to 410 MPa. Furthermore, due to the hydration effect, the free water molecules existing as a shear layer between the two hydrogen bond networks can reduce shear strength, greatly reduce energy dissipation in the friction process, and provide a good guarantee for the realization of superlubrication.
As shown in Fig. 9(d), the above hydrogen bond network is formed by the interaction between the ether bond (C-O-C) in PEG and the hydroxyl group (O-H) in TA. As seen from the infrared spectrum in Fig. 9(e), for PEG10000-TA, the vibration of the O-H bond in TA decreases from 3,275 to 3,246 cm −1 , and the vibration of the C=O bond increases from 1,698 to 1,717 cm −1 . The shift of these two chemical bonds proves the presence of hydrogen bonds between PEG and TA, which is consistent with previous studies [47,61]. With the increase of TA content in PEG10000-TA lubricant, the shift of O-H bond and C=O bond will become more and more obvious. And with the continuous addition of TA, the viscosity of PEG10000-TA mixed solution increased significantly due to the strong hydrogen bonding (Fig. S5 in the ESM) [47]. 4 (2) Furthermore, the mechanism of fast superlubrication can be attributed to the rapid formation of lubrication film on the surface of the friction pair. To test this hypothesis, 30 μL of PEG10000-TA was added to the surface of the precleaned glass sheet, and the surface was cleaned with deionized water after 10 s. As shown in Figs. 9(f) and 9(g), the water contact angle on the surface of the glass coated with PEG10000-TA www.Springer.com/journal/40544 | Friction is smaller than that before. The results show that PEG10000-TA quickly gets adsorbed on the surface of the friction pair because of its abundant hydroxyl groups and forms a lubricating film with excellent friction performance. Moreover, the decrease in the contact angle reflects that the hydrophilicity of the lubricating film improves, and it is easier to fix the water molecules as the shear layer between the two friction pairs. In addition, the adsorption behavior of lubricant and friction pair surface was further explored by quartz crystal microbalance (QCM). Figure 9(h) shows that PEG10000-TA lubricating materials have better adsorption properties than TA solution and PEG10000 solution, and the frequency change (Δf ) is about 97 Hz. Moreover, after washing with deionized water, the adsorption frequency of PEG10000-TA basically returned to the initial value. These results show that PEG10000-TA lubricant is easier to adsorb on the surface of friction pairs to prevent the direct contact of rough peaks, and the adsorption behavior belongs to reversible physical adsorption. Therefore, the existence of TA can not only increase the viscosity of the lubricant to enhance hydrodynamic lubrication, but also help to form an interfacial adsorption layer to reduce friction. In other words, TA may play a vital role in the rapid realization of superlubrication.

Conclusions
In this study, macroscale liquid superlubricity was achieved on the surface of Si 3 N 4 /glass within an ultrashort running-in period (9 s) by using polyethylene glycol-tannic acid complex green liquid lubricant (PEG10000-TA) mixed liquid as the lubricant. The results revealed that only lubricants containing PEG10000 and TA in a certain concentration range can achieve superlubrication, and materials containing other acids or low-molecular-weight PEG fail to achieve superlubrication. In addition, as a green liquid lubricant, PEG10000-TA does not cause corrosion to the friction pairs and is relatively safe for users. Most importantly, due to the ultrashort running-in time, the wear rate of the glass surface is only 1.19 × 10 −8 mm 3 ·N −1 ·m −1 , and the wear scar diameter of the Si 3 N 4 ball surface is small (111.42 μm). Furthermore, the lubricating material achieves fast and robust superlubrication under a relatively high contact pressure (410 MPa). The strong hydrogen bond network structure between PEG and TA molecules is an essential factor in the proposed lubrication model. Therefore, this study enriches the system of liquid superlubricity materials and provides a potential method for realizing macroscale superlubricity with ultra-short running-in period. The findings presented in this work may have significance for applying liquid superlubrication in industrial production processes.