A robust membrane with dual superlyophobicity for solving water-caused lubricant deterioration and water contamination

Lubricants are often contaminated by water in different ways. Water-polluted lubricants extremely accelerate wear corrosion, leading to the deterioration of lubricity performance. Recently, multiphase media superwettability has been developed to endow one surface with compatible functions, such as on-demand separation of oily wastewater. However, realizing the robustness of the dual superlyophobic surface to solve water-caused lubricant deterioration and water contamination as needed remains challenges. Herein, a robust dual superlyophobic membrane is presented to realize on-demand separation for various lubricant—water emulsions. Compared to pure lubricants, the purified lubricants have equivalent tribology performance, which are much better than that of water-polluted lubricants. The as-prepared membrane maintains dual superlyophobicity, high-efficient for water or lubricant purification, and excellent tribology performance of the purified lubricant, even after immersion in hot liquids for 24 h, multicycle separation, and sandpaper abrasion for 50 cycles. Water-polluted lubricant extremely accelerates wear corrosion to promote catalytic dehydrogenation of lubricants, generating too much harmful carbon-based debris. This work shows great guiding significance for recovering the tribology performance of water-polluted lubricants and purifying water by the dual superlyophobic membrane.


Introduction
Lubricants play a critical role to minimize friction and wear in all types of mechanical systems around people's lives. However, during the using and transportation process, lubricants were polluted by water via absorption and condensation in humid surroundings, leakage of components and parts, chemical reactions, etc. [1][2][3][4]. Especially, additivecontaining lubricants can easily form stable emulsions with free water, leading to the deterioration of tribology performance. Furthermore, oily wastewater will pose a serious threat to ecological environment. It is an urgent demand to purify water and lubricants in consideration of operation safety, resource recovery, and environmental protection.
In the past few years, by utilizing their completely opposite repellency toward oil and water, superwetting surfaces have captured tremendous exploration for oil/water separation to realize high-efficiency, convenient, and energy-saving purification [5][6][7][8][9][10][11][12][13][14][15]. Conventionally, superhydrophobic surface has underwater oleophilicity, www.Springer.com/journal/40544 | Friction which is a contradictory wettability compared with underwater superoleophobicity. Taking advantage of their opposite wettability, realizing superhydrophobicity and underwater superoleophobicity on the same surface provides a novel inspiration for oil/water separation as needed. Up to now, such a dual superlyophobic surface have not been reported for on-demand lubricant/water separation, recovering the tribology performance of water-polluted lubricants and purifying water. In addition, the robustness of dual superlyophobic surface is extremely intractable to achieve. It depends on the surface energy in a narrow range and high surface roughness [16,17].
In this work, a robust membrane with dual superlyophobicity was developed to solve water-caused lubricant deterioration and water contamination. By spraying suspension containing aluminium phosphate (AP) binder and SiO 2 nanoparticles, heat treatment, and modification with F-rich materials, the membranes acquire superhydrophobicity and underwater superoleophobicity to realize high-efficient separation for various lubricant-water emulsions as needed. Compared to pure lubricants, the purified lubricants have equivalent tribology performance, which is much better than that of water-polluted lubricants. Importantly, the as-prepared membranes maintain dual superlyophobicity, high-efficient water purification, and excellent tribology performance of the purified lubricants, even after immersion in hot liquids for 24 h, multicycle separation, and sandpaper abrasion for 50 cycles. It is found that water-polluted lubricants extremely accelerate wear corrosion while generating iron oxides as active sites to promote catalytic dehydrogenation of lubricants, generating too much harmful carbon-based debris. The results demonstrate the significance of removal of water in lubricants, offering a promising method to practically purify water and recover the tribology performance of water-polluted lubricants by the dual superlyophobic membrane.
Stainless steel meshes (SSMs) (AISI 316L, 2,300 mesh size), glass, copper foam, aluminium (Al), and ball-on-disk friction test pair (AISI E52100 steel) with a roughness of 0.02 μm were obtained from the local market. The radiuses of the disk and ball were 12 and 5 mm, respectively. Poly-alpha olefin (PAO) 2 (Chevron) was purchased from Shanghai Qixi International Trade Co., Ltd, China. 150N (China National Offshore Oil Corporation, CNOOC) and 4cst 100N (Abu Dhabi National Oil Company, ADNOC) were obtained from Shenzhen Longda Energy Co. Ltd, China. Rocket propellant (RP)-3 jet fuel and multi-alkyl cyclopentanes (MACs) were obtained from State Key Laboratory of Solid Lubrication, Lanzhou Institute of Chemical Physics, China. Other chemical reagents were directly used.

Preparation of the robust surface with dual superlyophobicity
Mixing aluminium hydroxide with orthophosphoric acid (60%) at a molar ratio of 1:3 and stirring at 100 °C for 3 h, the AP binder was obtained. AP (4 g) was dissolved in 10 mL deionized water. SiO 2 (0.7 g) nanoparticles were dispersed in the AP solution.
Anhydrous ethanol (30 mL) was added to the above mixture under stirring for 10 min and sonicating for 20 min to obtain the stable spraying suspension. Several pieces of substrates (35 mm × 35 mm) were cleaned with acetone in ultrasonic bath, and then dried in open air. Subsequently, spray the suspension uniformly on three pieces of substrates under 0.3 MPa air driving and 20 cm distance. In order to realize the robustness and crosslinking of the coatings, the spray-coated substrates were cured at 120 °C for 2 h, and the temperature was increased to 240 °C for 1 h, and then to 350 °C for 1 h to build the AP-SiO 2coated surfaces. The solidified surface was modified by 10 μL 1H,1H,2H,2H-perfluorooctyltrichlorosilane (FOTS, 78560-45-9, 97%), which was dissolved in 40 mL hexane to obtain AP-SiO 2 @FOTS-coated surface with dual superlyophobicity.

Emulsion separation
Lubricant-in-water emulsions were prepared by mixing lubricant (PAO, RP-3 jet fuel, and MACs) and water at a volume ratio of 1:100 under vigorous shaking

Tribological test
The tribological performance of lubricants was carried out by oscillating reciprocating friction and wear tester (Optimol, SRV-IV) with a ball-on-disk test pair. Lubricant (50 μL) was added on the surface of the disk. The reciprocating ball (frequency = 25 Hz, amplitude = 1 mm) was pressed against the stationary disk under 50 N preload for 30 s, and then 100 or 150 N load for 30 min. All tribological tests were performed at room temperature and in ambient air with about 40% relative humidity. For each lubricant sample, five tribological tests were performed to make the data analysis.

Hot liquid immersion test
The AP-SiO 2 @FOTS-coated SSM was immersed in PAO, RP-3 jet fuel, water, 1,2-dichloroethane (DCE), and hexane at 60 °C for 24 h, and then washed with ethanol. After drying, the treated membrane was used for contact angle (CA) observation and emulsion separation.

Sandpaper abrasion test
The AP-SiO 2 @FOTS-coated SSM was faced down the rough surface of a sandpaper (grit No. 800). Under a 50 g weight, the membrane was moved 10 cm by external drawing force, which was defined as one abrasion cycle. After every five cycles, the CA and the mass of the prepared coating were measured. Furthermore, the treated membrane was used for emulsion separation after every ten cycles.

Characterization
The scanning electron microscope (SEM; FEI, Quanta 650 FEG) equipped with an energy-dispersive X-ray spectroscopy (EDS) attachment (Element, EDAX) was used to obtain the surface morphology and element percentage and distribution. The surface roughness, stereoscopic topography, and wear volume were analyzed by two types of three-dimensional (3D) optical profilometers (Rtec-Instruments, UP-Lambda and KLA-Tencor, MicroXAM-800). The X-ray photoelectron spectrometer (Thermo Scientific, ESCALAB 250 Xi) was performed to study the chemical compositions of the samples. The crystal structure was investigated by the X-ray diffractometer (Rigaku, Smartlab-SE).
The static CA and sliding angle (SA) were measured on a CA system (Shanghai Zhongchen Digital Technic Apparatus Co., Ltd., JC2000D1). Select DCE as oil to measure the oil contact angle (OCA) in water. A digital camera (Sony, DSC-HX200) was used to take all optical images. The optical microscopy photographs of the feed emulsion and the corresponding filtrate were checked by an optical microscope (OLYMPUS, BX51). By measuring the chemical oxygen demand (COD) according to U.S. Environmental Protection Agency method 8000 (HACH, DRB 200), the organic content in the filtrate collected by lubricant-in-water emulsion separation was calculated. The oil purity of the filtrate collected by water-in-lubricant emulsion separation was obtained by a Karl Fischer titrator (Metrohm, 831 KF). The Raman spectra were studied by a confocal Raman spectrometer (HORIBA, LabRAM HR Evolution).

Results and discussion
High surface roughness and surface energy in a narrow range are two obstacles to maintain the robustness of the superhydrophobic and underwater superoleophobic surface for lubricant/water separation. The AP binder was selected to construct the coating according to two superiorities. Foremost, inorganic www.Springer.com/journal/40544 | Friction AP binder was exceptionally resistant to corrode by lubricants during lubricant/water separation. Secondly, condensation-polymerization reaction of the AP dramatically strengthened the interaction between the substrate and nanoparticles. As shown in Fig. 1, By uniformly spraying the suspension consisted of AP binder and SiO 2 nanoparticles on the SSM and then curing, the robust micro-and nanostructure was prepared. Modified by FOTS in exact concentration to make the surface energy in a narrow range, the membrane obtained dual superlyophobicity to realize on-demand lubricant/water separation and recover tribology performance of lubricants. Strong mechanical strength and small pore size of SSM are prerequisites of using as substrate for stable emulsion separation. For spray coating method, the mesh opening size of substrate greatly affects the pore size of the membrane ( Fig. S2(a) in the Electronic Supplementary Material (ESM)). Here, the SSM with 2,300 mesh size was selected to prepare the membrane for efficient emulsion separation. By spraying and curing, the crosslink structure of AP-SiO 2 -coated SSM minishes the pore size smaller than 4 μm (Fig. S2(b) in the ESM) to satisfy the needs of emulsion separation. After modifying by FOTS, the structure of the coating has barely changed. and roughness can be further quantified using the 3D optical profilometers. In Figs. 2(d)-2(f), the height of AP-SiO 2 @FOTS-coated SSM undulates drastically compared with that of pristine SSM. The root-meansquare roughness of the coating is 18.9 μm, while that of pristine SSM is 6.60 μm, which is attributed to micro/nano porous structure of AP-SiO 2 @FOTS coating.
Besides micro/nano rough structure, narrow-range surface energy is another decisive factor to obtain dual superlyophobicity. The surface chemical compositions were detected by the EDS and X-ray photoelectron spectroscopy (XPS) measurement. According to element distribution maps of AP-SiO 2 -coated SSM ( Fig. S2(d) in the ESM), O, P, Si, and Al disperse extensively compared with those of pristine SSM (Fig. S1(c) in the ESM), indicating that the hydrophilic AP-SiO 2 coating covers evenly on the substrate, which is also proved by the XPS spectrum (Figs. S1(b) and S2(c) in the ESM). The X-ray diffraction (XRD) pattern of AP-SiO 2 @FOTS-coated SSM shows the characteristic peaks of the SiO 2 and pristine SSM. The appearance of the peaks at about 21.5° is attributed to the transformation of AP from amorphous to crosslink crystal structures (Fig. S5 in the ESM). After modifying, element distribution maps had no significant change, but F (1.5%) is widely distributed on the surface of AP-SiO 2 @FOTS-coated SSM (Fig. S3 in the ESM). In the XPS spectra, there are five main peaks belonging to C, O, F, P, and Si elements (Figs. 2(g) and 2(h) and   In order to obtain dual superlyophobicity, the narrowrange surface energy is another exacting precondition, which needs co-existence of water-loving and waterrepelling substances in coating. From the C 1s XPS spectra in Fig. 2(h), three peaks at 293.2, 290.8, and 284.4 eV are ascribed to CF 3 -, CF 2 -, and C-C bonds of FOTS with low surface energy, respectively. But it is worth noting that the C-O peak appearing at 285.4 eV are ascribed to the unmodified AP-SiO 2 component with high surface energy. It is feasible for modifying by FOTS with accurate content to obtain dual superlyophobicity. Besides SSM, other substrates (glass, copper foam, and Al) are also appropriate for preparing AP-SiO 2 @FOTS-coated surfaces according to Figs. S6-S10 in the ESM.
High roughness and surface energy in a narrow range of AP-SiO 2 @FOTS coatings make the as-prepared surfaces acquire dual superlyophobicity, as shown in Fig. 3(b) and Fig. S11 in the ESM. Water droplets in air and oil droplets in water are all spherical and move easily without any residue. The wetting properties of AP-SiO 2 @FOTS-coated SSM, glass, copper foam, and Al were further observed by a CA analyzer. WCAs in air and OCAs in water are all above 150°, as shown in Fig. 2(i). AP-SiO 2 @FOTS coatings can be applied on various substrates to achieve dual superlyophobicity.
By utilizing completely opposite repellency toward oil and water of superhydrophobicity/underwater superoleophobicity, AP-SiO 2 @FOTS-coated SSM was prewetted by water/lubricant to form a wetting layer, and could extremely repel the lubricant/water and allow water/lubricant to penetrate the membrane, realizing lubricant-water emulsion separation without any external stimulate. Taking various lubricant-inwater (PAO/water, RP-3/water, and MACs/water = 1/100; v/v) and water-in-lubricant (water/PAO = 1/500,

3(d) and 3(e) and
Figs. S12 and S13 in the ESM, different types of emulsions are separated by the sieving effect and exhibit significant differences from milky white to the transparent state. The countless micro/nano oil and water droplets in emulsions are presented in the views by using the optical microscope. In contrast, no droplet can be found from all the optical microscopy images of the filtrates, which indicates that all the emulsions have been successfully purified. To further evaluate the separation performance, fluxes and CODs of filtrate in lubricant-in-water separation, and fluxes and oil purity of filtrate in water-inlubricant separation were obtained according to Figs. 3(a) and 3(c), respectively. The separation fluxes are higher than 500 L/(m 2 ·h), and the COD values are lower than 135 mg/L for various lubricant-in-water emulsions. For water-in-lubricant emulsions, the separation fluxes and oil purity of filtrates are higher than 300 L/(m 2 ·h) and 99.995%, respectively. Therefore, the dual superlyophobic membrane exhibits highefficiency separation performance for different kinds of lubricant-water emulsions as needed.
So far, many extensive works have paid close attention to the separation efficiency of the membrane, but the usability and operational performance of the filtrate have received little focus. The tribological performance of lubricants was explored with a ballon-disk test pair. As shown in Fig. 1, a reciprocating ball was pressed against a stationary disk, and both were added with lubricant. Figure 4(a) shows the friction coefficients (COF) of the test pair lubricated by PAO with different water contents (1/100 and 1/500), pure PAO, and filtrate after separating PAO with 1/500 water by AP-SiO 2 @FOTS-coated SSM. The COF of the test pair lubricated by pure PAO is 0.11-0.15 after a temporary increase. When trace amounts of water (1/500) sneak into PAO, the COF of the test pair sharply increases from 0.30 to 0.63, and then lasts 70 s with drastic fluctuation. Whereafter, the COF dramatically declines to 0.25 at 150 s and slowly dips down to 0.185. Especially, more water in PAO (1/100) seriously damages its lubricating performance, resulting a rapid increase of COF approaching 0.70 at 300 s followed by violent ups and downs. These results show that the increasement of water content in the lubricants leads to more rise and fluctuation of COF. After lubricate/water emulsion separation by the as-prepared membrane, the COF-time curve of the purified PAO is nearly identical with that of pure PAO.
To comprehensively analyze the influence of water content in lubricants and tribology performance of filtrates, wear scars obtained by tribological test were  | https://mc03.manuscriptcentral.com/friction studied by a 3D optical profilometer. As shown in Fig. 4(b), the wear volumes of the test pair lubricated by pure PAO, PAO with 1/500 water, and PAO with 1/100 water are 0.101×10 6 , 5.260×10 6 , and 7.817×10 6 μm 3 , respectively, indicating that the increasement of wear volume is caused by more water contents in PAO. After the high-efficient emulsion separation, the wear volume of the test pair lubricated by PAO filtrate oil is 0.098×10 6 μm 3 , showing the equivalent lubricating performance comparing with pure PAO.
The 3D images of wear scars visibly manifest the fatigue of lubricating capability for water-polluted PAO and the recovery of tribology performance for filtrate, comparing with those of wear scar lubricated by pure PAO (Figs. 4(c)-4(e)). The result is consistent with the SEM images (Figs. 4(f)-4(h) and Figs. S14(a) and S15(a) in the ESM). In Fig. 4(f) and Fig. S14(a) in the ESM, the wear scar lubricated by pure PAO is narrow, and the furrows on the worn surface are shallow. In contrast, the wear scar lubricated by 1/500 water-polluted PAO becomes wider and deeper, with some eroded pits observed on the worn surfaces ( Fig. 4(g) and Fig. S15(a) in the ESM). The EDS analysis further verifies in Figs. S14(b) and S15(b) in the ESM. The oxygen percentage and distribution maps of the wear scar lubricated by water-containing PAO are obviously higher and larger than those lubricated by pure PAO, respectively. It is mainly ascribed to the wear corrosion in the friction process. The iron of the test pair would be oxidized in ambient air under the intense friction. However, the presence of water in lubricant can accelerate the wear corrosion of the metallic substrate. Especially, shallow grooves are observed on the worn surface lubricated by PAO after separation, analogous to pure PAO.
Apart from PAO, other lubricants such as 150N www.Springer.com/journal/40544 | Friction and 4 cst 100N were also selected to investigate the influence of water in lubricants on their tribology performance. Similar to those of PAO, with the increasement of water in 150N or 4cst 100N, the COFs and wear volumes become higher, indicating the fatigue of lubricating capability caused by water. After lubricant-water emulsion separation by AP-SiO 2 @FOTS-coated SSM, all the filtrates recover their tribology performance comparing with pure lubricants (Fig. S16-S19 in the ESM).
Realizing the robustness of dual superlyophobic surface is significant for practical application but tough to achieve. It highly depends on the surface energy in a narrow range and high surface roughness. The slight changes of chemical compositions will greatly affect the dual superlyophobicity, which poses a higher requirement to chemical durability of the surface. During the process of oil/water separation, the membrane was inevitably immersed in water or other organic surroundings, making the variation of surface energy, destructing the bonding force of the coating, and blocking the channel of membrane probably [18][19][20][21][22]. Here, we selected lubricants (PAO and RP-3), water, and polar and nonpolar organic solvents (DCE and hexane) as liquids to verify the chemical durability. As shown in Figs. 5(a) and 5(b), even AP-SiO 2 @FOTS-coated SSMs are immersed in harsh liquid environment at 60 °C for 24 h, the membranes retain dual superlyophobicity. After hot liquid immersion test, the prepared membranes were performed for PAO/water emulsion separation. In Figs. 5(c) and 5(d), the fluxes of PAO-in-water and water-in-PAO emulsion separation are about 600 and 400 L/(m 2 ·h), respectively. Furthermore, COD values and oil purity of filtrates are all around 60 mg/L and 99.9955%, respectively. Compared with the corresponding results without hot liquid immersion tests in Fig. 3, the prepared membranes manifest their chemical durability and show a steady flux and high efficiency for on-demand separation.
Mechanical durability plays another vital role for the robustness of dual superlyophobic surface, which can keep the micro-and nanostructure to retain high surface roughness [23][24][25][26][27]. As shown in the inset of Fig. 6(a), sandpaper abrasion test was performed to systematically investigate the mechanical durability of the AP-SiO 2 @FOTS-coated SSM. After sandpaper abrasion for 50 cycles, the WCAs in air and underwater  OCAs are all above 150° (Fig. 6(a)), and the WSAs in air and the OSAs in water are about 5.7° and 2.0°, respectively. The mass loss of the AP-SiO 2 @FOTS coating is about 25% (Fig. 6(a)). Although there is no significant variation observed by naked eye after abrasion test (Fig. 6(c)), some destroyed structures are found by the SEM at high magnification ( Fig. 6(b)). From the EDS analysis in Fig 6(b), the element percentages and distributions of AP-SiO 2 @FOTS coating are almost unchanged, especially F (1.4%) compared with that in Fig. S3 in the ESM. Moreover, from full and C 1s XPS spectra in Fig. S20 in the ESM, the CF 3 -, CF 2 -, C-C, and C-O peaks still prove the co-existence of water-loving and water-repelling substances, showing the outstanding mechanical and chemical durability of the membrane. After 10 abrasion cycles, the prepared membrane was used to separate PAO-in-water or water-in-PAO emulsion as needed, and then perform for another 10 abrasion cycles and separation over and over again. In Figs. 6(d) and 6(e), the membrane exhibits steady flux, high-efficient separation, and low wear volume to recover the lubricating performance after multicycle separation and sandpaper abrasion.
After tribological tests, the optical microscopy images of the wear scars lubricated by pure and 1/500 water-polluted PAO are compared in Fig. 7. The wear scar created by pure PAO is relatively neat except some brown and green narrow areas. Whereas there are many large dark areas in wear scar lubricated by PAO with 1/500 water. The chemical components of the worn area are further investigated by the Raman spectra. As shown in Fig. 7, the brown worn areas lubricated by pure PAO have the characteristic peaks www.Springer.com/journal/40544 | Friction at about 670 cm −1 , which is attributed to iron oxides. However, it is worth noting that the characteristic peaks of black worn areas created by 1/500 watercontaining PAO appear at about 1,391, 1,545, and 2,790-2,940 cm −1 , which is corresponding to the D, G, and 2D bands of diamond-like carbon (DLC) film, respectively. As is well known, water in lubricants can greatly aggravate the oxidation and corrosion of metal substrates. According to Erdemir et al. [28], Berman et al. [29], and Wu et al. [30], the products of metal ions dispersed in water-polluted lubricants can facilitate the catalytic dehydrogenation of lubricant molecules, generating too much undesired carbon-based debris during the tribological tests, which could cause great hidden risk in operation.

Conclusions
In summary, we develop a robust membrane with superhydrophobicity and underwater superoleophobicty. By spraying AP binder and SiO 2 nanoparticles on substrate and then curing, the stable micro/ nanostructure is constructed. After delicately controlling surface energy by FOTS, the dual superlyophobic membranes are prepared to realize high-efficient separation for various lubricant-water emulsions as needed. Compared to pure lubricants, the purified lubricants have equivalent low wear volumes and COFs. The as-prepared membranes maintain dual superlyophobicity, high-efficient water purification, and excellent tribology performance of the purified lubricants, even after immersion in hot liquids (lubricant, water, and polar and nonpolar organic solvents) for 24 h, multicycle separation, and sandpaper abrasion for 50 cycles. Water extremely deteriorates the tribology performance of lubricants, accelerating wear corrosion while generating iron oxides as active sites to promote catalytic dehydrogenation of lubricants, generating too much harmful carbon-based debris. These results demonstrate the significance of on-demand separation for various lubricant-water emulsions, offering a prospect to practically purify water and recover the tribology performance of water-polluted lubricants by the dual superlyophobic membrane.