Effect of pH on the liquid superlubricity between Si₃N₄ and glass achieved with phosphoric acid

Abstract

In the present study, the pH dependence of liquid superlubricity between Si₃N₄ and glass achieved with phosphoric acid solution was investigated. It is seen that the superlubricity can be achieved only when the pH value is in the range of 0.75–2. To reveal the mechanism, the evolutions of confined solutions with different pH values between two friction surfaces were investigated by an online observation. It was seen that the superlubricity appeared when the confined solution between the two friction surfaces forms a starvation state. When the pH is in the range of 0.75–1.75, the starvation state can be formed as long as the running-in period is end. When the pH is in the range of 1.75–2, the starvation state can be formed by adding a certain amount of phosphoric acid molecules in the contact region, which leads to the transformation of an unstable friction state to a superlubricity state. When the pH is in the range of 0–0.75, the superlubricity cannot be obtained, no matter how the test conditions are changed because of the high contact pressure and lack of time for the tribochemical reaction to take place between the friction surfaces and hydrogen ions. When pH is greater than 2, the value of friction stays high because the amount of hydrogen ions adsorbed on the friction surfaces are not sufficient to make the surfaces positively charged.

---

Introduction

The concept of superlubricity, proposed by Hirano and Shinjo in the 1900s, is used to describe the physical phenomenon of complete vanishing of the frictional forces between two surfaces.^1,2 However, in actual experimental conditions, superlubricity can also be defined when the ratio of friction force to applied load is less than 0.01 (μ < 0.01).³ Because the superlubricity can effectively reduce energy dissipation in a mechanical system, investigations on it have become a hot research area in recent years.^4–7 At present, several solid lubricants are known to display superlubricity, such as DLC,^8,9 MoS₂,¹⁰ and graphite.¹¹ In addition, some liquid lubricants also exhibit superlubricity, for example, polymer brushes with water,^12,13 ceramic materials with water,^14,15 glycerol solution with acid or polyhydric alcohol,^16–18 and some types of polysaccharide mucilage from plants.^19,20 Although all of them can produce an ultralow friction coefficient less than 0.01, the mechanism of superlubricity in each of them is totally different. For solid lubricants, the superlubricity is usually attributed to incommensurate surface lattice structures, weak dispersive interlayer interactions or coulombic repulsions at the contact.^21–23 For liquid lubricants, there are at least three lubrication models, namely, tribochemistry reaction between ceramic and water, formation of hydrogen bond networks with water molecules, and formation of a hydration layer.^14,18,24 Therefore discovering new materials and establishing a new superlubricity theory is the most important task for superlubricity researchers.

Recently, the liquid superlubricity with the lubrication of phosphoric acid solution was achieved by our group.²⁵ It was found that the superlubricity is closely related to the hydrogen ions and the hydrogen bond network between phosphoric acid and water molecules.^26,27 However, there is still a problem regarding the relationship between hydrogen ions and superlubricity that needs to be solved because it was found that the superlubricity is closely dependent on the pH value of phosphoric acid solution. If this problem is solved, it would not only lead to a better understanding of the superlubricity mechanism of phosphoric acid, but can also help in the designing of new liquid superlubricity materials. Therefore, in the present work, the relationship between the frictional behavior and pH value of phosphoric acid was investigated and the mechanism of the influence of pH value on superlubricity was discussed according to the online observation.

Materials and methods

The phosphoric acid (H₃PO₄) used in the test was a commercial product (Beijing Chemical works) with a mass fraction of 85%. Before test, it was diluted using deionized water to different pH values (pH = 0, 0.5, 0.75, 1, 1.25, 1.5, 1.75, 2, and 2.5, measured by a standard pH meter). The friction pairs are a Si₃N₄ ball with a diameter of 4 mm (obtained from Shanghai Research Institute of Materials) and a glass slide with a surface roughness (Ra) of about 5 nm (part of the common glass slide for microscopy). Both the ball and the glass were sequentially cleaned in an ultrasonic bath with acetone and ethanol each for 15 min, and then they were washed with deionized water and dried with compressed air.

The friction coefficient was measured on a Universal Micro-Tribotester (UMT-3, Bruker, USA) with a rotating mode of ball-on-disk (Fig. 1). Before test, the H₃PO₄ solution was introduced between the ball and the disk by dropping few droplets (10 μL). The applied load on the disk was 3 N, corresponding to the maximal contact pressure of 700 MPa. The rotation speed of the disk was 180 rpm and the radius of the sliding track on the disk was 3 mm, corresponding to a linear sliding speed of 0.056 m s⁻¹. The ambient temperature was about 25 °C and the relative humidity was about 25%. To ensure the measuring accuracy of friction coefficient, the measurement errors in the test were also eliminated by the method described in ref. 28.

Fig. 1 Schematic illustration of the apparatus for online measuring friction coefficient and observing confined solution in the contact region.

The apparatus for online observing the confined solution between the two friction surfaces is schematically shown in Fig. 1 (the radius of the sliding track on the disk is the same as that in UMT-3). An optical microscope (×20) was located above the contact region between the two friction surfaces. A white light from a halogen lamp was focused through the microscope on to the contact region, and the views of the contact region were captured using a digital CCD camera at the same location on the same scale. Although the friction force cannot be measured by the in situ observation apparatus, the test conditions (speed, load and environment), the friction pairs and lubricant in the online observation were all the same as those in the friction tests to ensure that the evolution of friction coefficient with time is in accordance with that of the contact region with time.

Results and discussion

Fig. 2 shows the evolution of friction coefficient with time under the lubrication of H₃PO₄ solution with different pH values. It is found that when the pH values of H₃PO₄ solution are 0 and 0.5, the friction coefficients can reduce to 0.021 and 0.012, respectively, after a running-in period, and then have no further reduction. This indicates that the superlubricity cannot be obtained in both the cases. However, when the pH values of H₃PO₄ solution are 1 and 1.5, the friction coefficients reduce to less than 0.01 after a running-in period of 320 s and 600 s, respectively, which enters the superlubricity regime. When the pH value of H₃PO₄ solution increases to 2, the friction coefficient reduces to 0.011 after a running-in period of 600 s, but it shows a sudden increase from 0.011 to 0.5 at the end of the test. When the pH value of H₃PO₄ solution increases further to a higher value (pH = 2.5), the evolution of friction coefficient with time is totally different from that for pH < 2. It is seen that the friction coefficient does not reduce further; instead it retains a high value (greater than 0.5) all the time.

Fig. 2 Evolution of friction coefficient with time under the lubrication of H₃PO₄ solution with different pH values: (a) pH = 0, (b) pH = 0.5, (c) pH = 1, (d) pH = 1.5, (e) pH = 2, and (f) pH = 2.5.

According to the abovementioned friction results, it can be found that the pH value of H₃PO₄ solution has a significant effect on the final friction state. According to the further subdivision of pH value, the relationship between the final friction state and pH is shown in Fig. 3. It can be seen that only when the pH is in the range of 0.75 to 1.75, the superlubricity state (μ < 0.01) can be achieved. When the pH is less than 0.75, the final friction coefficient is in the range of 0.01 to 0.05 (the low friction state). When the pH is in the range of 1.75 to 2.0, the friction coefficient reduces to about 0.01–0.02, but increases to a high value at the end of test (the unstable friction state). When the pH is greater than 2, the friction coefficient keeps a high value (μ > 0.1) till the end of the test (the high friction state).

Fig. 3 Relationship between the friction state and pH values of H₃PO₄. The friction coefficient for every pH value was measured for five times under the same condition. The statistical error is the standard deviation of these five values.

In our previous work, it was found that when the pH value of acid solution (such as sulphuric acid, lactic acid and oxalic acid) was less than 2, the friction coefficient reduced to about 0.05 after the running-in period, which was attributed to the surface protonation reaction between hydrogen ions and friction surfaces (SiOH + H⁺ → SiOH₂⁺, SiO⁻ + 2H⁺ → SiOH₂⁺) during the running-in process.²⁶ This can lead to the formation of many positively charge surface sites to produce the stern layer and the electrical double layer to lower the friction force. In regards to H₃PO₄, it can be inferred that the mechanism of friction reduction is the same as the other acids when the pH < 2. When the pH > 2, the ionization of hydrogen ions from surfaces (≡SiOH → ≡SiO⁻ + H⁺) would tend to dominate and therefore, there would not be sufficient amount of hydrogen ions to be adsorbed on the two friction surfaces to make the surfaces positively charged.²⁹ This is the reason for the friction to retain a high value during the whole test. However, the above-described lubrication model of surface protonation reaction can only explain the reason for the decrease of friction coefficient in the running-in period. It cannot give an explanation regarding the superlubricity that can be obtained only in the pH range of 0.75–1.75. To investigate this phenomenon, the evolution of contact region with test time between the two friction surfaces was observed and analyzed by an online apparatus.

First, the contact region between the two friction surfaces lubricated by H₃PO₄ with a pH of 1.5 was investigated, as shown in Fig. 4. In the early stages of the test (from 0 s to 100 s), there was no change with time in the contact region, except for the increasing of wear region. However, when the test time lasted for 350 s, it was seen that the volume of solution confined between the two surfaces began to reduce within the frame size of CCD camera because of the evaporation of free water in the solution. With continuously increasing the test time, a bright “tail” appeared in the outlet region at the time of 450 s (Fig. 4(d)) and its length gradually increased with time. Eventually, after 525 s, it broke the edge of surrounding solution, and expanded to the whole track, as shown in Fig. 4(e). Simultaneously, there was also a change appearing in the inlet region. A small bright meniscus appeared in the inlet region (Fig. 4(e)), which began to expand towards the contact region. Finally, the confined solution between the two surfaces presented an “H” distribution, which is the same as the distribution of lubricant in the lubricating state of starvation.³⁰ Even after testing for a long period, the distribution of lubricant remained constant due to the non-volatility of phosphoric acid.

Fig. 4 Optical images of contact region between two surfaces lubricated by H₃PO₄ with a pH of 1.5. (a), (b), (c), (d), (e), and (f) correspond to the six moments on the friction coefficient curve. All of these images were captured at the same location on the same scale.

Second, the contact region between the two friction surfaces lubricated by H₃PO₄ with a pH of 0 was investigated, as shown in Fig. 5. It was found that the wear region initially increased with time (from 0 s to 100 s). At 100 s (Fig. 5(c)), there appeared a small bright “tail” in the outlet region, which began to gradually expand with time. At 400 s (Fig. 5(e)), the length of the “tail” reached the maximum (about 50 μm). Subsequently, the shape of the “tail”, and the friction coefficient remained constant. In addition, it was found that there was no change in the inlet region, and the confined solution between the two friction surfaces suffused the whole screen of CCD camera during the entire test, which was totally different from the final starvation state for pH = 1.5.

Fig. 5 Optical images of contact region between two surfaces lubricated by H₃PO₄ with a pH of 0. (a), (b), (c), (d), (e), and (f) correspond to the six moments on the friction coefficient curve. All of these images were captured at the same location on the same scale.

Third, the contact region between the two friction surfaces lubricated by H₃PO₄ with a pH of 2 was investigated, as shown in Fig. 6. It was found that the wear region increased with time during the early stages (from 0 s to 300 s), which was the same as that for pH = 1.5 and pH = 0. After that, it could be seen that the volume of solution confined between the two friction surfaces decreased with time. When the friction coefficient was reduced to 0.012 (t = 550 s), a considerable amount of excess solution was not observed around the contact region (Fig. 6(e)). After a short period (t = 610 s), a severe wear was seen in the contact region and in the whole track (Fig. 6(f)), which indicated that there was no liquid film to separate the two friction surfaces (dry friction). At this moment, the friction coefficient suddenly increased to 0.5.

Fig. 6 Optical images of contact region between two surfaces lubricated by H₃PO₄ with a pH of 2. (a), (b), (c), (d), (e), and (f) correspond to the six moments on the friction coefficient curve. All of these images were captured at the same location on the same scale.

Comparing the behaviour of the contact region for the three different pH values, it was found that the evolutions of the contact region with different pH values are the same at the early stages of test. However after the early stages, the contact regions exhibit different behaviours at different pH values. At pH = 1.5, the superlubricity can be achieved after a running-in period. From Fig. 4, it is seen that the solution confined between the two friction surfaces presents the “H” distribution (forming the starvation state) when superlubricity appears. At this moment, there is a thin film with a hydrogen bond network between H₃PO₄ and H₂O molecules adsorbed on the whole track.²⁰ At pH = 0, the distribution of solution confined between the two friction surfaces is different from that at pH = 1.5. There is a bright “tail” in the outlet region, but there is no bright meniscus in the inlet region. In addition, the solution suffuses the whole screen of CCD camera, corresponding to a friction coefficient of 0.021. When pH = 2, in the final stages, the excess solution is not left in the contact region, leading to a very high friction. According to the relationship between the friction coefficient and distribution of solution, it can be inferred that the requirement for superlubricity is the formation of a starvation state by the solution confined between the two surfaces.

The starvation usually occurs in the cases of high speed, high viscosity and limited lubricant supply.³¹ In these tests, the speed and the viscosity of lubricant (after the running-in period) are almost the same for the three different pH values. Therefore, they are not the reasons for the starvation. Thus, the only factor causing starvation is the limited amount of solution available to the contact, leading to the deficiency of the solution fed in the outlet region and inlet region, which does not provide a fully flooded condition. From Fig. 4(f), it can be seen that there remains a small volume of solution confined between the two friction surfaces to form the starvation state when pH = 1.5, which is in accordance with the factor for the abovementioned starvation. As for pH = 0 (Fig. 5(f)), it can be seen that the volume of solution confined between the two friction surfaces is considerably larger than that for pH = 1.5 because of which the starvation state cannot be formed. As for pH = 2, the result is just contrary, and it is found that the excess solution is not left between the two friction surfaces to form the starvation state. Analyzing these results, it can be inferred that if we reduce the initial volume of the solution added in the contact region when pH = 0, or increase the initial volume of the solution added in the contact region when pH = 2, the final volume of the solution confined between the two surfaces would become limited (similar to Fig. 4(f)), leading to the formation of the starvation state, and thus possibly achieving superlubricity by meeting the requirement for superlubricity.

To further confirm this inference, we reduced the initial volume of solution from 10 μL to 0.5 μL when pH = 0, and increased the initial volume of solution from 10 μL to 30 μL when pH = 2. The friction coefficient results and the final lubrication state are shown in Fig. 7(a) and (b). It was found that when pH = 0, the friction coefficient reduces to 0.02 after a short period of 35 s, and then remains constant. This suggests that the superlubricity cannot be achieved even by reducing the initial volume of solution to a very small value to form the starvation state when pH = 0. However, when pH = 2, after a long period of 500 s, the starvation state can be formed and the friction coefficient reduces to 0.006, which indicates that the superlubricity can be achieved if the initial volume of solution increases to a higher value to form the starvation state when pH = 2. According to this result, it can be inferred that the reason for the unstable friction state when pH = 2 is that the initial volume of H₃PO₄ solution is too small. To further confirm this inference, we added some H₃PO₄ solution (20 μL) in the contact region immediately after testing H₃PO₄ solution (pH = 2) for 550 s (Fig. 6(e)). It can be found that the stable superlubricity state can be achieved after a running-in period. Therefore, it can be concluded that the stable superlubricity state can be achieved as long as there are enough amounts of H₃PO₄ molecules left in the contact region to form the starvation state when pH = 2. However, for pH = 0, it can be confirmed that the initial volume of H₃PO₄ solution being too large to form the starvation state is not the only reason because of which the superlubricity cannot be achieved. In other words, there exist other factors that probably have influence on the superlubricity state when pH = 0.

Fig. 7 (a) Evolution of friction coefficient with time under the lubrication of H₃PO₄ solution (pH = 0) with an initial volume of 0.5 μL. (b) Evolution of friction coefficient with time under the lubrication of H₃PO₄ solution (pH = 2) with an initial volume of 30 μL. The inserts are the optical images of contact regions at the moment of 600 s.

Comparing Fig. 7(a) and (b), it can be seen that the running-in period for pH = 0 is considerably shorter than that for pH = 2. In our previous work, it was found that the running-in process is very important for superlubricity.²⁵ If there is no running-in process, the superlubricity cannot be obtained even with the lubrication of H₃PO₄ solution with pH = 1.5. Therefore, it can be inferred that the running-in period for pH = 0 is too short to obtain superlubricity. To confirm whether it is the real reason why the superlubricity cannot be obtained when pH = 0, an experiment was designed as follows. First, we used the H₃PO₄ solution with pH = 1.5 as lubricant. Second, we measured the friction coefficient for a short period (T_s), and then stopped the measurement. Third, we waited for measurement until there was no liquid left on the substrate (instead a transparent solid-like film is formed on the substrate) due to the evaporation of water in the solution. The friction coefficient was again measured after that under the same test condition (the load is 3 N and the rotation speed is 180 rpm). In this case, the running-in period can be controlled by adjusting the length of T_s. Here, we chose four different T_s; 15 s, 50 s, 100 s, and 200 s. The friction coefficients with the four different T_s are shown in Fig. 8. It is found that when T_s = 15 s, the final friction coefficient after the waiting period is 0.035. With the increase of T_s, the final friction coefficient after the waiting period decreases. When T_s increases to 100 s, the final friction coefficient becomes 0.011; when T_s further increases to 200 s, the superlubricity (μ = 0.004) can be achieved. These results indicate that the final friction coefficient reduces with increasing running-in period. Superlubricity can be obtained only when the running-in period is sufficiently long. Therefore, it can be concluded that the superlubricity cannot be obtained when pH = 0 is because the running-in period is too short.

Fig. 8 Evolution of friction coefficient with time under the lubrication of H₃PO₄ solution (pH = 1.5) for four different T_s (15 s, 50 s, 100 s, and 200 s).

Next, we attempted to find the reasons for the length of the running-in period to influence the superlubricity. During the running-in period, the contact area between the two friction surfaces would increase with time due to the original surface layer being ground off by mechanical action. Fig. 9(a) gives the relationship between the diameter of contact region and test time with the lubrication of H₃PO₄ solution (pH = 1.5). It is found that the diameter of contact region increases from 160 μm to 258 μm with the test time increasing from 4 s to 160 s, corresponding to an average pressure decrease from 149 MPa to 57 MPa. When the running-in period is very short, the contact pressure would become sufficiently high to exceed the maximal bearing pressure of the hydrogen bond network composed of H₃PO₄ and H₂O molecules. In this case, the hydrogen bond network would be destroyed, and instead there would appear direct asperity contact (dry friction), which would cause a high friction due to the high shearing strength of solid–solid contact. It is consistent with the phenomenon that the friction coefficient becomes high with increased applied load. For comparison, the diameter of wear region after the lubrication of H₃PO₄ solution with different pH values was measured, as shown in Fig. 9(b). It is found that the diameter of contact region increases from 170 μm to 268 μm with pH value increasing from 0 to 2, which would lead to the average contact pressure decrease from 132 MPa to 53 MPa. For pH = 0 and 0.5, the contact pressure is about 132 MPa and 92 MPa, which is equal to the contact pressure when the running-in period is 10 s and 24 s for pH = 1.5, respectively. Therefore, it can be concluded that the high contact pressure for pH = 0 and pH = 0.5 is one of the main reason because of which the superlubricity cannot be achieved.

Fig. 9 (a) Relationship between the diameter of wear region (contact pressure) and test time with the lubrication of H₃PO₄ solution (pH = 1.5). (b) Relationship between the diameter of wear region (contact pressure) and pH value of H₃PO₄ solution.

In addition, our previous work indicated that the running-in process of superlubricity includes two parts; they are tribochemical reaction between hydrogen ions and friction surfaces and the formation of the hydrogen bond network between H₃PO₄ and H₂O molecules.²⁷ In the early stages of the test, with the effect of rubbing, the hydrogen ions can be adsorbed on the friction surfaces by tribochemical reaction (SiOH + H⁺ → SiOH₂⁺, SiO⁻ + 2H⁺ → SiOH₂⁺) to form the stern layer and the electrical double layer. After that, with the evaporation of free water in the solution, a thin film with hydrogen bond network would be finally formed between the two surfaces, which would lead to the further reduction in friction coefficient to 0.004. If there is no running-in period, there would be no rubbing action. In this case, the tribochemical reaction between hydrogen ions and friction surfaces would not occur. Thus, there is only a thin film with hydrogen bond network formed between two friction surfaces due to which the superlubricity cannot be achieved. Similarly, when the running-in period is very short for pH = 1.5, there is not enough time for the tribochemical reaction between hydrogen ions and friction surfaces due to the short period of rubbing action, which also leads to high friction (Fig. 8). For pH = 0, the friction coefficient very quickly reduces to 0.02 (Fig. 7), which indicates that the thin film with the hydrogen bond network is rapidly formed. In this case, the two surfaces are separated by the hydrogen bond network in a very short period (only 35 s), and therefore, the time required for the tribochemical reaction between hydrogen ions and friction surfaces is not sufficient. As a result, there is only a thin film with hydrogen bond network formed between the two surfaces, which is in accordance with the final friction state, when the running-in period is very short for pH = 1.5. Therefore, it can be concluded that the lack of sufficient time for the tribochemical reaction between hydrogen ions and friction surfaces is also one of the main reason due to which the superlubricity cannot be achieved at pH = 0.

According to the abovementioned analysis, it can be found that the too short running-in period can lead to two negative factors for superlubricity; high contact pressure and insufficient time for tribochemical reaction, which are the two essential reasons due to which the superlubricity cannot be obtained when pH is lower than 0.75. Therefore, to obtain superlubricity when pH is lower than 0.75, the running-in period has to be increased to reduce the final contact pressure and to simultaneously provide the conditions for the tribochemical reaction between hydrogen ions and friction surfaces. It is found that the length of running-in period is mainly dependent on the pH value and volume of H₃PO₄ solution.²⁴ When the pH is higher and the volume is larger, the running-in period becomes longer. For pH = 0, there is only one way to increase its running-in period, which is by increasing the volume of lubricant. However, it is incompatible of forming a lubricating state of starvation. Therefore, it can be concluded that the superlubricity cannot be obtained when pH = 0, irrespective of the changes made to the test conditions. For pH = 2, the superlubricity can be achieved as long as the volume of the lubricant increases to a value that ensures that there are enough H₃PO₄ molecules left in the contact region to form the starvation state.

Conclusion

In summary, we have shown that the superlubricity can be achieved when the pH value of H₃PO₄ solution is in the range of 0.75–2. When pH is less than 0.75, the running-in period is significantly short, leading to a high contact pressure as well as lack of time for the tribochemical reaction between hydrogen ions and friction surfaces to occur which is the main reason due to which the superlubricity cannot be achieved. When pH is greater than 2, the friction surfaces cannot become positively charged to form the stern layer due to the low concentration of hydrogen ions, which is the reason due to which the friction coefficient maintains a high value. When pH is in the range of 0.75–2, the superlubricity can be obtained as long as the amount of H₃PO₄ molecules in the contact region is sufficiently large to form the starvation state. These results show that the superlubricity is not only closely linked to the concentration of hydrogen ions, but it is also simultaneously related to the amount of H₃PO₄ molecules in the contact region, which is useful for the better understanding of the mechanism of liquid superlubricity.

Acknowledgements

The work is financially supported by National Key Basic Research Program of China (2013CB934200), NSFC of China (51405256, 51335005, 51321092) and China Postdoctoral Science Foundation funded project (2014M550056).

References

1. M. Hirano and K. Shinjo, Phys. Rev. B: Condens. Matter Mater. Phys., 1990, 41, 11837–11851.
2. K. Shinjo and M. Hirano, Surf. Sci., 1993, 283, 473–478.
3. A. Erdemir and J.-M. Martin, Superlubricity, Elsevier Academic Press, New York, 2007.
4. J. J. Li and J. B. Luo, Sci. China: Technol. Sci., 2013, 56, 1–11.
5. J. Klein, Friction, 2013, 1, 1–23.
6. Y. Hu, T. Ma and H. Wang, Friction, 2013, 1, 24–40.
7. Q. F. Zeng, F. Yu and G. N. Dong, Surf. Interface Anal., 2013, 45, 1283–1290.
8. A. Erdemir, O. L. Eryilmaz and G. Fenske, J. Vac. Sci. Technol., A, 2000, 18, 1987–1992.
9. J. A. Heimberg, K. J. Wahl, I. L. Singer and A. Erdemir, Appl. Phys. Lett., 2001, 78, 2449–2451.
10. J. M. Martin, C. Donnet, T. Lemogne and T. Epicier, Phys. Rev. B: Condens. Matter Mater. Phys., 1993, 48, 10583–10586.
11. M. Dienwiebel, G. S. Verhoeven, N. Pradeep, J. W. M. Frenken, J. A. Heimberg and H. W. Zandbergen, Phys. Rev. Lett., 2004, 92, 126101.
12. U. Raviv, S. Giasson, N. Kampf, J. F. Gohy, R. Jerome and J. Klein, Nature, 2003, 425, 163–165.
13. M. Chen, W. H. Briscoe, S. P. Armes and J. Klein, Science, 2009, 323, 1698–1701.
14. M. Chen, K. Kato and K. Adachi, Wear, 2001, 250, 246–255.
15. F. Zhou, X. L. Wang, K. J. Kato and Z. D. Dai, Wear, 2007, 263, 1253–1258.
16. C. Matta, L. Joly-Pottuz, M. I. D. Bouchet and J. M. Martin, Phys. Rev. B: Condens. Matter Mater. Phys., 2008, 78, 085436.
17. Z. Z. Ma, C. H. Zhang, J. B. Luo, X. C. Lu and S. Z. Wen, Chin. Phys. Lett., 2011, 28, 4.
18. J. J. Li, C. H. Zhang, L. R. Ma, Y. H. Liu and J. B. Luo, Langmuir, 2013, 29, 271–275.
19. S. Arad, L. Rapoport, A. Moshkovich, D. van Moppes, M. Karpasas, R. Golan and Y. Golan, Langmuir, 2006, 22, 7313–7317.
20. J. J. Li, Y. H. Liu, J. B. Luo, P. X. Liu and C. H. Zhang, Langmuir, 2012, 28, 7797–7802.
21. M. Hirano, K. Shinjo, R. Kaneko and Y. Murata, Phys. Rev. Lett., 1997, 78, 1448–1451.
22. A. Socoliuc, R. Bennewitz, E. Gnecco and E. Meyer, Phys. Rev. Lett., 2004, 92, 134301.
23. C. Q. Sun, Y. Sun, Y. G. Ni, X. Zhang, J. S. Pan, X. H. Wang, J. Zhou, L. T. Li, W. T. Zheng, S. S. Yu, L. K. Pan and Z. Sun, J. Phys. Chem. C, 2009, 113, 20009–20019.
24. U. Raviv and J. Klein, Science, 2002, 297, 1540–1543.
25. J. J. Li, C. H. Zhang and J. B. Luo, Langmuir, 2011, 27, 9413–9417.
26. J. J. Li, C. H. Zhang, L. Sun, X. C. Lu and J. B. Luo, Langmuir, 2012, 28, 15816–15823.
27. J. J. Li, L. R. Ma, S. H. Zhang, C. H. Zhang, Y. H. Liu and J. B. Luo, J. Appl. Phys., 2013, 114, 114901.
28. J. J. Li, C. H. Zhang, L. Sun and J. B. Luo, Tribol. Trans., 2013, 56, 141–147.
29. N. Sahai, Environ. Sci. Technol., 2002, 36, 445–452.
30. B. Damiens, C. H. Venner, P. M. E. Cann and A. A. Lubrecht, J. Tribol., 2004, 126, 105–111.
31. T. Lubrecht, D. Mazuyer and P. Cann, C. R. Acad. Sci., Ser. IV: Phys., Astrophys., 2001, 2, 717–728.

---