Stable underwater superoleophobic conductive polymer coated meshes for high-efficiency oil–water separation

Abstract

Here we demonstrated that high-content nitrogen endowed polyaniline (PANI) and polypyrrole (PPy) have not only a doping–dedoping ability but also an inherent hydrophilic property. Combining their rigid chain structures, rough surfaces can be easily achieved by uniformly and firmly coating PANI and PPy onto the surfaces of stainless steel meshes via a simple modified dilute polymerization. Compared to the pristine meshes, the modified meshes become hydrophilic in air and superoleophobic underwater, and maintain stability in various harsh conditions. Taking advantage of this unique wetting feature, PANI and PPy coated meshes were used to separate mixtures of oil and water efficiently.

---

1 Introduction

Oil spill accidents and oily wastewater have caused great ecological disasters and economic losses.¹ To resolve these problems, scientists worldwide are trying to develop new technologies and materials. Recently, a large number of materials with special wettability, of broad variation in design, have been fabricated to separate oil and water with high efficiency and low power consumption. Initially, the development of superhydrophobic and superoleophilic materials was explored to realize the separation by absorbing or filtering oils from the oil–water mixtures.^2–5 However, these “oil-removing” materials are easily contaminated or obstructed by oils, resulting in a low separation efficiency. As for filtering oils, it is unfavorable for the separation of light oils because water with a higher density tends to form a barrier layer between light oils and substrates.

To overcome the above-mentioned disadvantages, materials with the opposite wettability (superhydrophilic and underwater superoleophobic) have been developed. To date, various organic hydrophilic materials have been introduced onto the surface of porous membranes, such as polyacrylamide,⁶ P2VP-b-PDMS block copolymer,⁷ chitosan,⁸ cellulose,⁹ poly(acrylic acid),¹⁰ poly(sulfobetaine methacrylate),¹¹ functionalized polyethylene terephthalate¹² and so on. These polymers show attractive advantages, such as flexibility, variability and robust combination with various substrates, that are attributed to their large amount of functional groups. However, most of these polymers have poor stability and fail to provide enough surface roughness, and both problems hinder the building of underwater superoleophobic membranes and application in oil–water separation.

In order to enhance the stability and surface roughness, inorganic materials are widely used to modify the polymer coatings.¹³ Furthermore, meshes coated by inorganics with a high surface energy have also been fabricated to achieve oil–water separation, such as ZnO, TiO₂, CaCO₃, zeolite, graphene oxide (GO), Cu(OH)₂, and so on.^14–19 Generally, the synthetic methods require multiple processes and special treatments to strengthen the interaction between substrates and coatings. Therefore, it is of great significance to develop a one-step and low energy-consuming method to fabricate underwater superoleophobic membranes with high stability, and to achieve high-efficiency oil–water separation.

Conductive polymers such as polyaniline (PANI) and polypyrrole (PPy) have been of intense interest due to their excellent physical and chemical properties, such as controllable conductivity, ease of synthesis, environmental stability, and their unique doping–dedoping and oxidation/reduction properties.^20,21 Besides, the rigidity of conductive polymers makes them act like inorganics. Making use of the compatible advantages of organic and inorganic materials, up to now, a great deal of research works have mainly concentrated on the doping–dedoping property of PANI and PPy to construct superhydrophobic surfaces.^22–25 In this work, we investigated the intrinsic hydrophilicity and other advantages of the conductive polymers, and fabricated PANI and PPy coated stainless steel meshes via a simple modified dilute polymerization. The prepared meshes show superhydrophilicity in air and superoleophobicity underwater, even in acidic, alkaline and salt solution. The PANI and PPy coated meshes, which consist of nanoscale rough coatings and microstructured porous substrates, can easily separate oils and water under gravitational force with a high separation efficiency and good recyclability.

2 Experimental section

2.1 Materials

Stainless steel meshes with different pore sizes were commercially available. Gasoline and diesel were purchased from a nearby filling station. The other chemicals were analytical-grade reagents and were used as received without further purification.

2.2 Fabrication of PANI nanofibers and PANI coated meshes

We adopted a modified dilute polymerization method to prepare PANI nanofibers.²⁶ In a typical procedure, ammonium persulfate (APS) was dissolved in 15 mL of 1 M aqueous solution of different acids and then rapidly mixed with 15 mL of aniline dissolved in a 1 M solution of the same acid. The reaction solution contained 0.02 M aniline and 0.01 M APS. The acids included perchloric acid (HClO₄), hydrochloric acid (HCl), sulfuric acid (H₂SO₄), and nitric acid (HNO₃). The mixed solution was shaken vigorously for about 1 min, and then left still at room temperature for 5 hours. After aniline polymerization, the dark-green precipitates were purified by centrifugation at 9000 rpm, and then re-dispersed in water.

Stainless steel meshes with different pore sizes were cleaned with ethanol and distilled water in an ultrasonic cleaner to remove any possible impurities. Afterward, the clean meshes were immersed into the above-mentioned reaction solution under stirring. When the dopant acid was HCl, particularly, the aniline polymerization reaction was performed in an ice bath in order to control the reaction rate and facilitate the uniform coating of PANI. After aniline polymerization, the obtained PANI coated meshes were thoroughly washed with water and dried at 60 °C in a drying oven.

2.3 Fabrication of PPy coated meshes

PPy coated meshes were prepared in a 1 M HClO₄ solution via a similar route, except that pyrrole and APS were used as the monomer and oxidant, respectively.

2.4 Characterizations

The optical images of the PANI aqueous solution and the movies of the separation of the oils and water were obtained using a digital camera (Sony, DSC-HX200). The morphology of uncoated and conductive polymer coated meshes was observed by a field emission scanning electron microscope (FESEM, JEOL JSM-6701F). The chemical compositions of uncoated and PANI coated meshes were investigated by X-ray photoelectron spectroscopy (XPS, Thermo Scientific ESCALAB 250Xi). The UV-visible absorption spectra were obtained using a UV-visible spectrophotometer (Agilent Cary 60). The morphology of the as-synthesized PANI was measured with a transmission electron microscope (TEM, FEI Tecnai G2 TF20). The TEM samples were fabricated by dropping the PANI aqueous solution onto a copper grid, and allowing it to dry in a vacuum oven. Water contact angles (WCAs) and oil contact angles (OCAs) were obtained using a JC20001 contact angle system (Zhongchen digital equipment Co. Ltd. shanghai, China) by measuring five different positions on the same sample. The WCAs were measured on the sample stage directly. Meanwhile, the OCA measurements were performed in cubic glassware filled with the aqueous phase. The chemical oxygen demand (COD) in the filtrates was measured according to U.S. Environmental Protection Agency method 8000 using a digital reactor block DRB 200 (HACH).

3 Results and discussion

In our previous work, we coated GO onto the surface of stainless steel meshes via a simple adsorption method, in which a high-concentration GO aqueous solution was required.¹⁸ The GO coated meshes showed superoleophobicity underwater due to the hydrophilic property of GO. In addition, S. Roth and H. J. Park proposed that most conductive polymers are related to graphene and could be cut out from graphene sheets.²⁷ Inspired by their viewpoints, we suggested that conductive polymers (especially high-content N-doped conductive polymers) could possess inherent hydrophilicity similar to O-doped graphene (GO). In order to verify our viewpoint, we prepared PANI as a typical N-doped conductive polymer via a very simple route (dilute polymerization²⁶ or rapid mixing²⁸). Fig. 1a shows the UV-visible absorption spectra of doped and dedoped PANI in aqueous solution. It is clearly seen that the as-prepared PANI is in the doped state, showing two typical absorption bands at about 415 and 830 nm. After adding ammonia solution, the PANI suspension turns blue and transforms to a dedoped state with blue-shifted absorption bands at 340 and 620 nm, in agreement with the literature.^26,29 Note that this was observed for a transparent and highly stable colloidal suspension during the whole process of aniline polymerization. Besides, the obtained PANI can be well re-dispersed in water (see insets in Fig. 1a), and kept stable in aqueous solution after standing for one week (Fig. 1b) due to the presence of electric double layers around the PANI chains.²⁹ The ability to form a stable PANI aqueous solution demonstrates that PANI possesses excellent hydrophilicity. Furthermore, we dropped the PANI aqueous solution onto a copper grid. According to TEM observations, PANI nanofibers of 50–100 nm in diameter were homogeneously deposited on the copper grid (Fig. 1c and d), further indicating the good dispersion of PANI in aqueous solution and its hydrophilic property.

Fig. 1 (a) UV-visible absorption spectra and photographs (inset) of doped (1) and dedoped (2) PANI. (b) PANI aqueous solution before and after standing for one week at room temperature. (c and d) TEM images of PANI at different magnification.

Because of the inherent polymer attributes, we can observe abundant PANI homogeneously and strongly coated on the walls of the reaction vessels after the aniline polymerization, including hydrophilic glass beakers and hydrophobic plastic centrifuge tubes. In other words, regardless of the properties of the substrates immersed in the aniline polymerization solution (hydrophilic or hydrophobic), PANI can be easily grown on the substrates and adheres tightly, similar to mussel-inspired polydopamine.³⁰ Distinctively, the surfaces modified by polydopamine suffer from insufficient roughness and poor stability. For example, Xu et al. prepared superhydrophilic meshes combining polydopamine coatings with linear poly(acrylic acid), in which the polydopamine coated meshes are only hydrophilic and not superoleophobic underwater due to the insufficient roughness.¹⁰ Note that the surfaces modified by PANI with rigid chain structures exhibit great roughness. Fig. 2 presents SEM images of neat and PANI coated stainless steel meshes. As for the neat meshes, a typical image in Fig. 2a shows knitted porous structures with an average pore diameter of about 38 μm (400 mesh size). In addition, the surface of a neat mesh is clean and smooth (see the inset in Fig. 2a). As shown in Fig. 2b–d, PANI has been uniformly coated onto the surface of the neat meshes. The pore size of the meshes is slightly decreased after coating with PANI. The magnified images of the surface of the PANI coated meshes show a large number of short PANI nanofibers, resulting in a very rough surface. Furthermore, we prepared PANI coated meshes using different acidic solutions to study the effect of acids on the surface morphology and wettability of PANI coated meshes; the acids included HClO₄, HCl, H₂SO₄ and HNO₃. It was found that micro/nanoscale hierarchical rough coatings on all modified meshes can be obtained though some differences exist (Fig. S1, ESI†).

Fig. 2 SEM images of neat (a) and PANI coated (b, c and d) meshes with an average pore diameter of about 38 μm. The inset in (a) shows the magnified view of the smooth surface on a neat mesh. PANI coated meshes were fabricated in a HClO₄ solution.

As is well known, chemical composition and micro/nano structure play important roles in the wettability of solid surfaces. The surface composition of the PANI coated meshes was investigated using XPS measurements as shown in Fig. 3. For the pristine meshes, C and O peaks are mainly observed. In contrast, for the PANI coated meshes a new peak appears at about 400 eV attributed to the presence of N element from the PANI coatings. Moreover, Fig. 3b shows the fine N 1s spectra of nitrate doped and dedoped PANI coated meshes. The nitrate doped PANI coated meshes prepared in a HNO₃ solution exhibit two peaks at around 407 and 400 eV ascribed to two kinds of N element (nitrate ion and benzenoid amine). After dedoping with ammonia solution, the dedoped PANI coated meshes have only one peak at about 400 eV due to the removal of nitrate ions on the dedoped PANI chains. The XPS measurements further confirm the successful coating of PANI on the meshes. A large number of the polar components on the PANI chains, such as benzenoid amines and doped acids, can strongly interact with water, leading to the inherent hydrophilicity of PANI.

Fig. 3 (a) XPS full survey spectra of neat (bottom) and PANI coated (above) meshes. (b) N 1s fine spectra of nitrate doped (above) and dedoped (bottom) PANI coated meshes.

Afterward, we systematically investigated the wettability of the neat and PANI coated meshes using the contact angle measurements. It is clearly seen that the WCAs of the neat meshes are about 120°, as shown in Fig. 4a. Compared to the neat meshes, both doped and dedoped PANI coated meshes are hydrophilic (Fig. 4b, c and 5). Notably, the water droplet can spread completely on the surface of the doped PANI coated meshes, suggesting that the doped PANI coated meshes show more hydrophilicity than the dedoped samples with WCAs around 45°. In any case, all these data reveal that the PANI coatings have a significant impact on the surface wetting property of neat meshes.

Fig. 4 Photographs of water droplets on neat (a), doped (b) and dedoped (c) PANI coated meshes in air. (d) OCAs in water of neat and doped PANI coated meshes. The insets are the representative photographs of oil droplets on neat (left) and PANI coated (right) meshes in water. The PANI coated meshes were prepared in a HClO₄ solution.

Fig. 5 The wettability of different acid doped (red) and dedoped (blue) PANI coated meshes in air (sparse) and in water (dense).

To further study the wetting change of the stainless steel meshes after the modification with PANI, the underwater OCAs were measured comprehensively by immersing the meshes into water. Fig. 4d and 5 show the underwater OCAs of uncoated, doped and dedoped PANI coated meshes for a series of oil droplets, including toluene, diesel, gasoline, hexane, chloroform and dichloroethane. The typical photographs of the oil droplets are shown in the insets on the right in Fig. 4d. Compared to the uncoated meshes with OCAs of about 120°, the OCAs of both doped and dedoped PANI coated meshes are above 150°, demonstrating that PANI coated meshes become superoleophobic underwater. In addition, the pore size of the stainless steel meshes is another important factor that may influence the wettability. Thus, we tested the dichloroethane contact angles of uncoated and PANI coated meshes with the pore diameters in the range of 38–425 μm in aqueous solution (Fig. S2, ESI†). All the dichloroethane contact angles of the PANI coated meshes are almost 150° and obviously higher than those of the neat meshes, indicating that the underwater superoleophobic property of the PANI coated meshes is not affected by the pore diameters.

Moreover, the measurements of underwater oil adhesion were carried out by using a dichloroethane droplet as a probe oil to contact and leave the surface of uncoated and PANI coated meshes. Fig. 6 shows that the oil droplet can completely detach from the surface of the PANI coated meshes without leaving any residue. On the contrary, the oil droplet adheres strongly to the surface of the neat meshes, indicating that the PANI modification can reduce the affinity for oil. The low oil adhesion of the PANI coated meshes was further confirmed by the sliding state experiments. It can be clearly seen that the dichloroethane droplet tightly sticks to the surface of the neat meshes with a big tilting angle. In contrast, the oil droplet can easily roll along the surface of the PANI coated meshes even at a small tilting angle (Fig. 7 and Movie S1, ESI†). These results demonstrate that the PANI coatings endow the stainless steel meshes with underwater superoleophobic and ultralow oil-adhesion characteristics in the water–oil–solid three phase systems. This is ascribed to the micro/nanoscale hierarchical rough structures and massive hydrophilic amino functional groups of the PANI coatings. The excellent wettability is favorable for preventing the obtained meshes from fouling and becoming blocked up by oils during the oil–water separation processes.

Fig. 6 Dynamic underwater oil-adhesion measurements on uncoated (a) and PANI coated (b) meshes.

Fig. 7 The rolling states of the oil droplets (about 5 μL) on the surface of neat (a–c) and PANI coated (d–f) meshes. Hexane (a and d) and dichloroethane (b, c, e and f) were used to test the meshes with a pore diameter of 38 μm (a, b, d and e) and 425 μm (c and f).

As for the practical applications in industrial wastewaters, the ability to maintain the superoleophobic property underwater in the corrosive solutions is a key issue for the filtering membranes. Hence, the corrosion resistance of the as-prepared PANI coated meshes was studied by immersing the modified meshes into corrosive solutions, such as acidic, basic and saline solutions. Fig. 8 shows the representative photographs of the oil droplets on the surface of PANI coated meshes in the corrosive solutions. The OCAs for either heavy oils or light oils in both acidic and basic solutions are around 150°, and the low oil-adhesion characteristics are also preserved. Although the light oil droplets in the saline solution show an oval image with an OCA of 140° due to the higher buoyancy force of the high-density saline solution, the light oil droplets likewise easily roll off the surface of the PANI coated meshes (Movie S2, ESI†). We propose that the PANI coatings can survive in harsh conditions, and their underwater superoleophobicity can be maintained.

Fig. 8 Photographs of the dichloroethane and hexane droplets on the PANI coated meshes which are submerged in 0.1 M HCl (left), 1 M NaCl (middle), and 0.1 M NaOH (right) aqueous solutions.

In addition to PANI, we also investigated the wetting property of PPy as another important N-doped conductive polymer. In an identical way, PPy coated meshes were prepared by the modified dilute polymerization method. Using APS as the oxidant, pyrrole monomers were in situ polymerized onto the surface of the stainless steel meshes. The SEM images shown in Fig. S3 (ESI†) certify that a large number of PPy small granules with an average diameter of about 80 nm aggregate, resulting in micro/nanoscale rough structures. In addition, the OCAs of the PPy coated meshes were measured in aqueous solution (Fig. S4, ESI†). The meshes with the PPy coatings display underwater superoleophobicity, similarly to PANI. It is suggested that N-doped conductive polymers including PANI and PPy possess an intrinsic hydrophilic property and can be used for constructing underwater superoleophobic films which have great potential for practical application in oil–water separation.

Taking advantage of the underwater superoleophobic and low oil adhesion properties, oil–water separation is conducted using PANI coated meshes as illustrated in Fig. 9a. The PANI coated meshes were fixed between two glass tubes wrapped with Teflon tape. The PANI coated meshes were pre-wetted with water in advance, then the mixture of hexane and water was poured into the top tube. It was observed that water with a higher density passed through the modified meshes quickly, but hexane was retained above it (Movie S3, ESI†). Note that the separation process just relies on the gravitational force, which is cost saving. Likewise, the PPy coated meshes can be also used to separate oils and water. In contrast, the neat meshes cannot realize oil–water separation. Moreover, a series of oil–water mixtures can be separated effectively by using the same method, such as hexane, toluene, diesel and gasoline, and the water flux is more than 10 L m⁻² s⁻¹ (Fig. S5a, ESI†). Simultaneously, the recyclability of the PANI coated meshes was tested by repeating the separation process. As a result, the prepared meshes still maintain high water flux, underwater superoleophobicity and low oil adhesion after 70 cycles of oil–water separation (Fig. S5b and S6, ESI†), demonstrating an excellent stability against recycling for long term use, which well fulfils the requirements for oil–water separation on a large scale.

Fig. 9 (a) Oil–water separation process of PANI coated meshes with a pore diameter of 38 μm. The PANI coated meshes were fixed between two glass tubes and the oil–water mixtures were poured into the top tube. To allow clear observation, the water was dyed with methylene blue and the hexane was dyed with Sudan IV. (b) Oil–water separation efficiency of various light oils.

The separation efficiency of a variety of oils was expressed by the water permeation coefficient (R/%), which was calculated as follows: R(%) = m_separation/m_initial. Where m_separation represents the weight of water after each separation, and m_initial denotes the weight of water in the initial oil–water mixtures. The separation efficiency of the PANI coated meshes is above 98% (Fig. 9b). Besides, no visible oil residues were observed in the collected water. The oil content in the filtrates was also analysed using COD measurements. The measured content was about 8, 25, 58 and 109 mg L⁻¹ for hexane, diesel, toluene and gasoline, respectively (Fig. S7, ESI†). Together, these results demonstrate the high-efficiency oil–water separation of the conductive polymer coated meshes.

4 Conclusions

In summary, we have demonstrated the inherent hydrophilic property of high-content N-doped conductive polymers, such as PANI and PPy. Based on this, PANI and PPy have been successfully coated on the surface of stainless steel meshes by a simple modified dilute polymerization, leading to a micro- and nanoscale hierarchical rough surface. As a result, the obtained meshes show hydrophilicity in air and superoleophobicity in water, even in various harsh conditions due to the high stability of the conductive polymers. Taking advantage of the underwater superoleophobic and low oil adhesion properties, the conductive polymer coated meshes can simply and efficiently separate oils from various oil–water mixtures just under a gravity-driven force, and maintain excellent stability against recycling for long term use. As a consequence, N-doped conductive polymers are deemed to be promising candidates for constructing underwater superoleophobic films for practical applications such as oil–water separation, marine antifouling, microfluidic devices, and so on.

Acknowledgements

This work was supported by the National Nature Science Foundation of China (nos 21203217 and 11172301), the “Western Light Talent Culture” project and the “Top Hundred Talents” program of Chinese Academy of Sciences.

Notes and references

1. J. L. Schnoor, Environ. Sci. Technol., 2010, 44, 4833–4833.
2. L. Feng, Z. Zhang, Z. Mai, Y. Ma, B. Liu, L. Jiang and D. Zhu, Angew. Chem., Int. Ed., 2004, 116, 2046–2048.
3. J. Li, L. Shi, Y. Chen, Y. B. Zhang, Z. G. Guo, B. L. Su and W. M. Liu, J. Mater. Chem., 2012, 22, 9774–9781.
4. P. Mishra and K. Balasubramanian, RSC Adv., 2014, 4, 53291–53296.
5. E. Bormashenko, S. Balter, Y. Bormashenko and D. Aurbach, Colloids Surf., A, 2012, 415, 394–398.
6. Z. X. Xue, S. T. Wang, L. Lin, L. Chen, M. J. Liu, L. Feng and L. Jiang, Adv. Mater., 2011, 23, 4270–4273.
7. L. B. Zhang, Z. H. Zhang and P. Wang, NPG Asia Mater., 2012, 4, e8.
8. S. Zhang, F. Lu, L. Tao, N. Liu, C. Gao, L. Feng and Y. Wei, ACS Appl. Mater. Interfaces, 2013, 5, 11971–11976.
9. F. Lu, Y. N. Chen, N. Liu, Y. Z. Cao, L. X. Xu, Y. Wei and L. Feng, RSC Adv., 2014, 4, 32544–32548.
10. L. X. Xu, N. Liu, Y. Z. Cao, F. Lu, Y. N. Chen, X. Y. Zhang, L. Feng and Y. Wei, ACS Appl. Mater. Interfaces, 2014, 6, 13324–13329.
11. Q. S. Liu, A. A. Patel and L. Y. Liu, ACS Appl. Mater. Interfaces, 2014, 6, 8996–9003.
12. W. D. Liu, X. Y. Liu, J. Z. Fangteng, S. L. Wang, L. P. Fang, H. Z. Shen, S. Y. Xiang, H. C. Sun and B. Yang, Nanoscale, 2014, 6, 13845–13853.
13. L. P. Xu, J. Zhao, B. Su, X. L. Liu, J. T. Peng, Y. B. Liu, H. L. Liu, G. Yang, L. Jiang, Y. Q. Wen, X. J. Zhang and S. T. Wang, Adv. Mater., 2013, 25, 606–611.
14. D. L. Tian, X. F. Zhang, Y. Tian, Y. Wu, X. Wang, J. Zhai and L. Jiang, J. Mater. Chem., 2012, 22, 19652–19657.
15. L. B. Zhang, Y. J. Zhong, D. Cha and P. Wang, Sci. Rep., 2013, 3, 2326.
16. P. C. Chen and Z. K. Xu, Sci. Rep., 2013, 3, 2776.
17. Q. Wen, J. C. Di, L. Jiang, J. H. Yu and R. R. Xu, Chem. Sci., 2013, 4, 591–595.
18. F. Zhang, W. B. Zhang, Z. Shi, D. Wang, J. Jin and L. Jiang, Adv. Mater., 2013, 25, 4192–4198.
19. Y. Dong, J. Li, L. Shi, X. B. Wang, Z. G. Guo and W. M. Liu, Chem. Commun., 2014, 50, 5586–5589.
20. A. G. MacDiarmid, Angew. Chem., Int. Ed., 2001, 40, 2581–2590.
21. D. Li, J. X. Huang and R. B. Kaner, Acc. Chem. Res., 2009, 42, 135–145.
22. Y. Zhu, D. Hu, M. X. Wan, L. Jiang and Y. Wei, Adv. Mater., 2007, 19, 2092–2096.
23. W. X. Liang and Z. G. Guo, RSC Adv., 2013, 3, 16469–16474.
24. X. Y. Zhou, Z. Z. Zhang, X. H. Xu, F. Guo, X. T. Zhu, X. H. Men and B. Ge, ACS Appl. Mater. Interfaces, 2013, 5, 7208–7214.
25. J. An, J. F. Cui, Z. Q. Zhu, W. D. Liang, C. J. Pei, H. X. Sun, B. P. Yang and A. Li, J. Appl. Polym. Sci., 2014, 131, 7208–7214.
26. N.-R. Chiou and A. J. Epstein, Adv. Mater., 2005, 17, 1679–1683.
27. S. Roth and H. J. Park, Chem. Soc. Rev., 2010, 39, 2477–2483.
28. J. X. Huang and R. B. Kaner, Angew. Chem., Int. Ed., 2004, 43, 5817–5821.
29. J. Li, L. H. Zhu, W. Luo, Y. Liu and H. Q. Tang, J. Phys. Chem. C, 2007, 111, 8383–8388.
30. H. Lee, S. M. Dellatore, W. M. Miller and P. B. Messersmith, Science, 2007, 318, 426–430.

---

Footnote

† Electronic supplementary information (ESI) available: Fig. S1–S7 and Movies S1–S3. See DOI: 10.1039/c5ra01681a

---