Quantification/mechanism of interfacial interaction modulated by electric potential in aqueous salt solution

With the development of surface and interface science and technology, methods for the online modulation of interfacial performance by external stimuli are in high demand. Switching between ultra-low and high friction states is a particular goal owing to its applicability to the development of precision machines and nano/micro-electromechanical systems. In this study, reversible switching between superlubricity and high friction is realized by controlling the electric potential of a gold surface in aqueous salt solution sliding against a SiO2 microsphere. Applying positive potential results creates an ice-like water layer with high hydrogen bonding and adhesion at the interface, leading to nonlinear high friction. However, applying negative potential results in free water on the gold surface and negligible adhesion at the interface, causing linear ultra-low friction (friction coefficient of about 0.004, superlubricity state). A quantitative description of how the external load and interfacial adhesion affected friction force was developed, which agrees well with the experimental results. Thus, this work quantitatively reveals the mechanism of potential-controlled switching between superlubricity and high-friction states. Controlling the interfacial behavior via the electric potential could inspire novel design strategies for nano/micro-electromechanical and nano/micro-fluidic systems.


Introduction
Researchers have been studying friction for centuries owing to the universality of frictional forces and the massive energy wastage they cause. Friction is generally a comprehensive result of numerous surface asperity collisions at the micro/nanoscale, making it difficult to be accurately predicted and controlled [1][2][3]. Atomic force microscopy (AFM) is a versatile tool for studying friction mechanisms at the nanoscale, which significantly reduces the complexity of the problem [4][5][6]. The friction between two solid dry surfaces involves both mechanical resistance [7,8] and interfacial interaction [9,10]. Corrugation of the potential energy surface is an essential friction mechanism [11,12]. It is helpful to reduce the friction between interfaces by maintaining a flat potential energy surface [11,12]. When a liquid medium is introduced between two solid surfaces, it works as a lubricant to reduce the friction force according to the hydrodynamic effect and/or adsorbed boundary film [13][14][15][16][17][18]. Water-based lubricants have been widely studied owing to their low viscosity and environmentally friendly properties [19][20][21]. The application of external fields is also an effective strategy for controlling and regulating interfacial friction [22][23][24]; the electric potential of a surface can be especially useful for controlling its interfacial friction behavior in solution [4,25,26]. Thus, studying the effects of electric potential on friction can aid our understanding of friction mechanisms.
In aqueous solutions, water molecules rearrange their conformations on charged surfaces [27][28][29][30], resulting in viscosities near the surface that are 10 5 -10 7 orders of magnitude higher than that of bulk water [31][32][33][34][35][36][37]. Previous studies have shown that such confined water layers have a significant impact on the frictional properties of the surface. For example, Dhopatkar et al. [38] reported that, when water is confined between two charged solid surfaces modified with cationic surfactants, it becomes "ice-like" and greatly reduces the interfacial friction between the two surfaces. In contrast, Valtiner et al. [39] studied the friction between mica and gold surfaces in an electrochemical environment, and reported that the highly viscous water layer that formed on the positively charged gold surface increased the interfacial friction. Clearly, the orientation of water molecules under an electric field profoundly affects the friction between surfaces. However, the reasons for this are not yet fully understood. In an attempt to elucidate this phenomenon, Pashazanusi et al. [40] studied the friction between a SiO 2 probe and a gold surface using electrochemical AFM. They noted that the friction force under a positive potential (+0.6 V vs. Ag) was 35-times higher than that under a negative potential (−0.6 V vs. Ag). These results were attributed to the formation of hydrogen bonds between the SiO 2 probe and the ice-like water layer on the positively charged gold surface. Such hydrogen bond formation was further confirmed by the results of adhesion tests performed by Li et al. [41], who suggested that hydrogen bonds form between hydrophilic SiO 2 microspheres and the ice-like water layer; conversely, hydrophobic polystyrene microspheres form no such bonds owing to the lack of hydroxyl groups on their surfaces.
Previous studies on electrically controlled friction have successfully converted the frictional states of surfaces from low to high (and vice versa) by controlling the adsorption of surfactants [42,43], the adsorption structure of ionic liquids [26,44], and the structure of adsorbed water molecules [38][39][40][41]. However, changing the surface frictional state between superlubricity (coefficient of friction (CoF): <0.01) and high friction in real-time remains a challenge. With the development of industrial applications that require improved frictional control, strategies for reversibly switching between superlubricity and highly frictional states are in increasing demand [45][46][47].
Previously [40], the reversible switching of friction forces between an AFM tip and a gold surface was observed upon changing the electric potential. However, since the applied load was relatively large (micronewton magnitude) and the tip was very sharp (radius of curvature: <10 nm), the lowest CoF that could be obtained was still above 0.1. In this study, we used a relatively low load (nanonewton magnitude) and a SiO 2 microsphere tip (radius: 10 μm), which allowed reversible switching between superlubricity and high nonlinear friction by controlling the electric potential of the gold surface in aqueous salt solution electrolyte. In addition, we established a quantitative relationship between adhesion and friction for the boundary lubrication state. Thus, the current work provides new insight into the origins of frictional forces and the physical properties of confined liquid layers on solid surfaces. The findings will aid the design and control of micro/nano-electromechanical systems (M/NEMSs).

Experimental section
Preparation of gold film: A grade-V1 mica plate (S & J Trading, Inc.) was cut into 10 mm × 10 mm sheets and cleaved using scotch tape, then immediately placed in a sputtering vacuum chamber (ATC2200-V sputtering system, AJA International, Inc.) to avoid contamination by pollutants in the air. A gold film with a thickness of approximately 100 nm was then deposited on the mica surface. The deposition rate was ~1 Å/s and the vacuum was maintained at 10 −6 Torr. The gold-coated mica was removed from the chamber and immediately placed into a closed sample box filled with dry nitrogen gas. Prior to conducting the electrochemical and friction experiments, the gold-coated mica was washed with deionized water and dried with nitrogen gas. Electrochemistry and friction experiments: A SiO 2 microsphere with a diameter of 20 μm was glued onto a tipless AFM cantilever (FORT-TL), which was used to measure the interfacial forces on the charged goldcoated substrates in combination with a Multimode 8 AFM system (Bruker Corporation, Santa Barbara, CA). An AFM tip was used to image the topography of the gold surface. A universal biopotential (Veeco Instruments) was used to control the electric potential of the gold surface, and a 0.1 M NaCl aqueous solution was used as the electrolyte. Ag and Pt wires with diameters of 0.5 mm were used as the quasi-reference electrode and counter electrode, respectively. The gold surface was used as the working electrode. The potential was limited to between −0.6 and +0.6 V, because the adsorption of ions (e.g., chloride ions) is minimal in this range, and there are no faradaic processes such as gold oxidation [40,41]. All the electric potentials in this work are given as vs. Ag electrode. When the potential was higher/lower than the potential of zero charge (U pzc = 100 mV), the gold was considered to be positively/negatively charged [41].
The spring constant of the AFM cantilever was calculated by the Sader method [48][49][50]. The deflection sensitivity was calibrated based on the slope of a retraction force curve obtained by pressing a SiO 2 microsphere glued to a cantilever onto a sapphire wafer. The normal force between the SiO 2 microsphere and the gold-coated substrate was measured in ramp mode as a function of the z-sensor position ( f-z curve). The adhesive force was determined from the maximum deflection force of the cantilever in the retraction process just before the microsphere left the surface. The f-z curves were converted to normal force vs.
surface separation ( f-s) curves using a conversation method described elsewhere [51]. The lateral stiffness of the cantilever was calibrated by an improved wedge calibration method before each measurement [52,53]. The friction force was measured by sliding the SiO 2 microsphere perpendicularly to the long axis of the cantilever.

Results and discussion
The AFM images in Figs. 1(a)-1(c) show the topography of the gold surface when switching between potentials of −0.6 and +0.6 V. The AFM images were prepared using a sharp tip in contact mode with an applied force of 320 nN. When a negative potential of −0.6 V was applied, the surface was rough with some irregular bulges. However, when the potential was increased to +0.6 V, the AFM image portrayed a smooth surface. Then, when the potential was returned to −0.6 V, the surface then returned to its original rough topography. This indicates that the gold surface was not damaged by the sharp tip; rather, at +0.6 V, there was some distance between the gold and the tip that created a flat image. Considering that the double-layer electrostatic force acting on the tip can be neglected (as discussed in our previous study [40,41]), and that there is a strong hydration effect at +0.6 V (see Fig. 2), we can conclude that a hydration layer exists on the gold surface at +0.6 V that separates the tip from the surface.
To further explore this interfacial hydration layer, normal force curves were obtained at different potentials as the microsphere approached the gold surface (Fig. 2). According to the Derjaguin-Landau-Verwey-Overbeek (DLVO) theory, the microsphere should encounter double-layer electrostatic forces and van der Waals forces as it approaches the gold surface [19,28]. However, the ionic strength used in our experiments was sufficiently large (0.1 M; corresponding to a Debye length of less than 1 nm) that the electrical double-516 Friction 9(3): 513-523 (2021) | https://mc03.manuscriptcentral.com/friction layer force may be ignored [28,40]. In addition, the van der Waals attraction force is not significant for any of the approach curves; thus, the microsphere encounters strong repulsive forces at all surface distances. According to previous studies [19,27,28,32,40], these repulsive forces can be ascribed to the steric effect of surface-confined water molecules, which is known as hydration force [19]. The effective surface viscosity is relatively low at negative potential because the surface water molecules can move freely [27,40]; however, as the applied potential increases, the surface water molecules align at the interface to form an ice-like surface layer, increasing the viscosity of the water layer and therefore the repulsion forces [27,31,32,34,40,41]. Classical hydrodynamic drainage (F = 6πηR 2 V/D) can be used to calculate the viscosity of the surface layer [32,40,54,55]. In the present case, F is the repulsive force acting upon the SiO 2 microsphere; η is the effective viscosity of the surface water layer; R is the radius of the microsphere; V is the approach velocity; and D is the distance between the microsphere and the gold surface. Thus, the effective viscosity of the surface water layer is about 0.05 Pa·s at −0.6 V and about 80 Pa·s at +0.6 V-the latter being five orders of magnitude higher than that of bulk water. These results are in accordance with those of previous studies [32,39,54].  Figure 3 shows the typical friction behavior between the SiO 2 microsphere and the gold surface at positive (+0.6 V) and negative (−0.6 V) potentials. The apparent CoF (defined as the ratio of friction force to normal load) was much higher at positive potential than that at negative potential, as shown in Fig. 3(c). At −0.6 V, the apparent CoF was below 0.01, which is in the range of superlubricity [56,57]; whereas at +0.6 V, the apparent CoF was approximately 0.035. This change in friction behavior is due to the differences in the confined water structure [38][39][40][41]. The ice-like water layer that exists at positive potential is capable of forming hydrogen bonds with the SiO 2 microsphere, resulting in a high CoF [39][40][41]. However, the ice-like structure is disrupted at negative potential [27,39,40,58], and the aqueous salt solution acts as a boundary lubricant, reducing the attractive forces between the SiO 2 microsphere and gold surface [40,41] and reducing the CoF [39][40][41]. Furthermore, as the gold surface is rough at negative potential and smooth at positive potential (see Fig. 1), the actual contact area between the gold surface and the microsphere would be smaller at negative potential, which may also contribute to lower friction.
It is worth noting that a superlubricious state is observed, which is different from the low-friction state that was observed previously at negative potential for the same electrochemical system [40,41], because the hydration force (shown in Fig. 2) arising from the aqueous salt solution plays a more important role in reducing friction at negative potential. Previously [40], friction switching was observed with a relatively large load on the gold surface (micronewton magnitude) and a sharp AFM tip (radius of curvature: ~10 nm). The CoF was low at negative potential due to the lubrication of the salt solution [39][40][41] and the disappearance of the structured water. However, the salt solution at the interface was easily removed upon contact because of the high hertzian contact pressure (>21 GPa); therefore, the lubrication effect could not be fully exerted. In contrast, the SiO 2 microsphere used in this study had a larger size (radius: ~10 μm) and was applied at a smaller load (nanonewton magnitude; the maximum Hertzian contact pressure was about 110 MPa), causing the salt solution to create a better hydration lubrication effect at −0.6 V. In addition, the adhesion between the microsphere and the gold surface is very small at negative potential (less than 1 nN, see Figs. S1 and S2 in the Electronic Supplementary Material (ESM)), which indicates that the interfacial reaction between the microspheres and the gold substrate is extremely weak [59,60]. These factors promote the occurrence of superlubricity. Figure 3(d) shows the load dependence of the friction force at −0.6 and +0.6 V. A linear relationship was observed at −0.6 V, while a nonlinear relationship was observed at +0.6 V. Notably, this transition was fully reversible upon alternating between −0.6 and +0.6 V several times, with excellent repeatability.
Owing to the highly viscous water layer that forms on the positively charged gold surface, it is possible that the nonlinear behavior observed at +0.6 V is caused by hydrodynamic drag acting on the microsphere in the lateral direction. To investigate this further, the load dependence of the friction force between a polystyrene microsphere and a gold-coated substrate was examined at −0.6 and +0.6 V (Fig. S3 in the ESM). Using the polystyrene microsphere, the friction force was almost the same at +0.6 V as at −0.6 V, and a linear relationship was observed between the friction force and load, irrespective of the potential. This indicates that drag force is unlikely to be the reason for the nonlinear friction behavior observed for the SiO 2 microsphere at positive potential. In addition, such drag forces tend to increase as the sliding speed increases [61,62]; however, the friction force was observed to decrease with sliding speed (see Fig. S4 in the ESM), further implying that drag contributes little to the friction between the SiO 2 microsphere and positively charged gold surface. A detailed explanation of this decrease in friction with sliding speed is given in the ESM.
The oriented water molecules on the positively charged gold surface create a contact plane with a large number of exposed dangling hydroxyl groups. These hydroxyl groups partake in hydrogen bonding with the hydroxyl groups on the SiO 2 microsphere, leading to high adhesion [27,40,41,63]. Adhesion plays an important role in nanoscale friction, with friction generally increasing with adhesion forces [64,65]. Nanoscale friction is linearly correlated with load for weak adhesion between two surfaces [66,67].  However, the correlation changes from linear to nonlinear as the adhesion forces increase [66,67]. Thus, it is necessary to explore the effect of adhesion on friction. Figure 4(b) shows that at positive potential, the relationship between friction and adhesion is linear as the load increases. At negative potential, the adhesion and friction forces were small, as shown in Figs. S1(a), S2(a), and S2(b) in the ESM. The results indicate that adhesion plays an important role in nonlinear friction at positive potential. Accordingly, we analyzed the friction at positive potential using the classical binomial law of friction (Eq. (1)), which describes friction in terms of the mechanical resistance and molecular attraction [66,67]. Considering the load dependence, the friction force F f takes the form where F L is the normal load, A is the actual contact area, and μ is a constant representing load dependence. The physical meaning of α was clarified by Israelachvili [28]; it represents the contribution of the adhesive hysteresis ∆γ and its linearly with α. In this work, the adhesion hysteresis is high (see the approach and retraction curves in Fig. S1(b) in the ESM); hence, the adhesive hysteresis is approximately equal to the adhesive interaction strength of the retraction curve (∆γ ≈ γ R ). Therefore, the friction force takes the form For random rough surface contact, the measured adhesive force F adh is considered to be linear with the actual contact area A and the adhesive interaction strength γ R . It follows that As a result, the friction force is dependent on both the normal load F L and the adhesive force F adh . The relationship between the adhesive force and load at positive potential is shown in Fig. S1(a) in the ESM, and the relational expression is given in the supplementary information (Eq. (S1) in the ESM). Combining Eqs. (1) and (S1) yields where b is the attenuation factor of the nonlinear adhesion behavior. By substituting a load of F L = 0 into Eq. (3), the friction force F f equals to K. Therefore, K represents the strength of adhesion-dominated friction caused by hydrogen bonding [41], which is related with the electric field-induced water structure. The friction data taken at positive potentials in Fig. 4(a) were fitted using Eq. (3), and the fitting parameters are given in Table S1 in the ESM. The values of R 2 are close to 1, indicating that Eq. (3) agrees well with the experimental data. The extremely low adhesive forces and friction forces at negative potentials (see Fig. S2 in the ESM) indicate that adhesion contributes little to friction in this state. The high adhesion and friction forces at positive potentials are coupled with extremely low μ values (below 0.01, as shown in Table S1 in the ESM), indicating that the high friction at positive potential originates almost completely from the high adhesive force. The values of the dynamic parameters K and b decrease with decreasing potential, indicating a decrease in hydrogen bonding strength, which is consistent with previous studies [40,41].
Previous studies have revealed that the adhesion between a SiO 2 microsphere and a positively charged gold surface in an aqueous solution originates from hydrogen bonding between the SiO 2 microsphere and the interfacial confined ice-like water layer [27,40,41,63]. By analyzing the adhesion results at positive potential (Figs. 4(b) and S1(a) in the ESM), it is reasonable to infer that hydrogen bonding becomes exponentially weaker with distance from the gold-coated substrate. The high confinement of the interfacial water layer originates from the alignment of water molecules [27,68]. The population of oriented water molecules is highest adjacent to the positively charged gold surface, and decreases with distance from the surface into the bulk solution [40,41]. When the SiO 2 microsphere penetrates into the viscous water layer, a high number of oriented water molecules form strong hydrogen bonds with the microsphere. In contrast, when the SiO 2 microsphere is suspended close to (but not in) the viscous water layer, the number of oriented water molecules it encounters is so low that little hydrogen bonding occurs. Based on analysis of the adhesion results (Figs. 4(b) and S1(a) in the ESM), a reasonable inference is that the hydrogen bonding becomes exponentially weaker with distance from the goldcoated substrate. This nonlinear decay of hydrogen bonding would become more pronounced with increasing potential, because the surface water molecules align to a greater extent at higher potentials [27,40,68], resulting in a larger difference in the number of oriented water molecules between the surface and bottom of the confined water layer.
Thus, the nonlinearity of the friction behavior with increasing load on the positively charged gold surface can be attributed to the strength of hydrogen bonding at the contact interface. When the loading force is low (10 nN), hydrogen bonding at the contact interface is so weak that the attraction between the SiO 2 microsphere and the gold-coated substrate contributes little to interfacial friction, as shown in Fig. 4(c). Upon increasing the loading force, the contact interface is forced closer to the gold-coated substrate, where the hydrogen bonding becomes exponentially stronger ( Fig. 4(d)). This strong hydrogen bonding causes the friction force to increase rapidly, resulting in nonlinear friction behavior with load. Conversely, under negative potential, the oriented water structure is disrupted, preventing hydrogen bonding between the SiO 2 and the surface water layer. The free water adsorbed at the Au/SiO 2 interface bears the nanonewton scale normal load, and acts as an effective boundary lubricant, resulting in a superlubricious state that has a linear correlation with load, as shown in Figs. 4(e) and 4(f). Since the transition of the water structure under positive and negative potential is reversible [27,40,41], the transition of friction is also reversible.
Owing to the lubrication mechanism, the reversible switching of the lubrication state and nonlinear friction behavior is not limited to systems using aqueous NaCl solution. Similar friction and lubrication behaviors were observed with NaOH, KCl, and NaNO 3 aqueous solutions (see Fig. S5 in the ESM). Thus, this study may have a far-reaching impact on many important fields of science and technology related to friction, adhesion, and wear. For example, it is difficult to precisely control the lifetime of M/NEMSs, because control of the friction state with very small dimensions and loads and high surface area-to-volume ratios is required [26]. The system explored in the present study demonstrates that appropriate electrical stimulation can precisely alter the interfacial friction behavior. Furthermore, this approach has good durability (Fig. S6 in the ESM). In addition, aqueous salt solutions are good conductive lubricants for contacting electrified surfaces [69,70]. Traditional oil lubricants or molecular lubricants such polytetrafluoroethylene are not suitable for such situations because of their electrical insulation properties.

Conclusions
We demonstrated the reversible switching between superlubricity and nonlinear high friction of a gold surface sliding against a SiO 2 microsphere in aqueous solution by tuning the electrochemical potential. At a negative potential, the apparent CoF was lower than 0.01, which is in the range of superlubricity; the friction force changed linearly with load, in accordance with Amonton's law. Conversely, at positive potential, the apparent CoF was higher, and the friction force increased nonlinearly with load. A relationship between external load, interfacial adhesion, and friction at positive potentials was developed, which indicates that the nonlinear high friction is caused by the strength of hydrogen bonding at the shearing plane. Thus, our findings constitute an advancement in the theory of potential-controlled friction between surfaces. This may have far-reaching implications for the development of nanotechnologies for use in M/NEMSs.

Electronic Supplementary Material:
Supplementary materials (load-dependent adhesive force at different potentials; friction and adhesive force at negative potentials; friction behavior between a polystyrene (PS) microsphere and the charged gold surface; friction force vs. sliding velocity at −0.6 and +0.6 V; lubrication state and friction behavior in NaOH, KCl, and NaNO 3 aqueous solutions; apparent CoF versus the repeat cycle) are available in the online version of this article at https://doi.org/10.1007/s40544-019-0354-7.
Open Access: This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.