TiO₂@Ag/P (VDF-HFP) composite with enhanced dielectric permittivity and rather low dielectric loss

Abstract

Ag-loaded TiO₂ hybrid particles (TiO₂@Ag) were synthesized as fillers by using an ethylene glycol reduction method. Based on this, the TiO₂@Ag/P(VDF-HFP) hybrid films were prepared by embedding TiO₂ nanoparticles in a poly(vinylidene fluoride-co-hexafluoropropylene) [P(VDF-HFP)] matrix. Morphology and thermal analysis show that the introduction of TiO₂@Ag does not disrupt the continuity of the polymer matrix, while the strong interaction between fillers and matrix influences the crystallization process of the composite. The adhesion of Ag to TiO₂ efficiently separates the Ag particles from connecting and suppresses the formation of a conductive network, and the ultra-small Ag nanoparticles “trap” the carriers due to coulombic blockade and quantum confinement effect, which results in low dielectric loss and electrical conductivity of the composites. As a consequence, the dielectric permittivity of polymer nanocomposites was enhanced by 300% over the P(VDF-HFP) matrix at a filler content of 30 vol% while maintaining a rather low dielectric loss (0.037 at 1 kHz), demonstrating promising applications in electronic devices.

---

Introduction

Polymer matrix dielectric nanocomposites have become research hotspots as new-generation dielectric materials, since they take advantage of high dielectric permittivity from ceramic fillers and high dielectric strength from the organic matrix.^1–7 Among them, the three phase system consisting of ceramic/conductive fillers/polymer matrix achieves excellent dielectric properties while keeping good processability in theory.^8–15 However, the conductive particles tend to aggregate and the ceramic fillers usually cause electric field distortion, which limits their practical applications. With the traditional simple mixing process of different components it is hard to control the distribution of conductive particles. As a consequence, the composite would exhibit a sharp increase in dielectric loss and conductivity and change from an insulator to conductor at higher filler concentrations. Therefore, it would make great sense to prevent the continuous contact of conductive particles and make them uniformly dispersed in polymer matrix. Nowadays, researchers develop a new strategy which is to load conductive particles on ceramic fillers to avoid them from aggregation and keep the film insulated. For instance, Zha et al.¹⁶ synthesized BaTiO₃@SnO₂ nanoparticles to prepare composite based on PVDF. When the filler concentration reaches to 45 vol%, high dielectric permittivity (90 at 1 kHz) was achieved, which is much higher than BT/PVDF while the dielectric loss is less than 0.6. Liu et al.¹⁷ reported the preparation of carbon nanotubes adhering BaTiO₃ nanoparticles (BT@CNTs) via a chemical vapor deposition (CVD). At the content of 10 vol%, the dielectric permittivity of the BT@CNTs/P(VDF-HFP) exhibits 27.7 at 50 Hz, which is more than triple that of the pure PVDF-HFP and PVDF-HFP/BT.

In this research, Ag-loaded TiO₂ hybrid particles were prepared via an ethylene glycol reduction method. The introduction of TiO₂ separates Ag nanoparticles from connecting and suppresses the formation of conductive network, and the ultra-small Ag nanoparticles “trap” the carriers due to Coulomb blockade and quantum confinement effect, which results in low dielectric loss and electrical conductivity of composites. The TiO₂@Ag/P(VDF-HFP) composite exhibit enhanced dielectric properties and the dielectric permittivity of polymer nanocomposites was enhanced by 300% over the P(VDF-HFP) matrix at a filler concentration of 30 vol% while maintaining a rather low dielectric loss (0.037 at 1 kHz).

Experimental

Materials

Principal materials include AgNO₃(AR, Sinopharm Chemical Reagent Co.,Ltd), titanium dioxide (TiO₂, 200 nm, anatase, Aladdin, China), polyvinylpyrrolidone (PVP, K30, GR), ethylene glycol (EG, Sinopharm Chemical Reagent Co.,Ltd), sodium chloride (NaCl, Sinopharm Chemical Reagent Co., Ltd), poly(vinylidene fluoride-co-hexafluoropropylene) (P(VDF-HFP), average M_w 455 000, average M_n 110 000, pellets, sigma, China), N,N-dimethyl formamide (DMF, Aladdin, China). All of the materials had not been disposed further, such as purification.

Preparation of the TiO₂@Ag hybrid structure particles

Ethylene glycol reduction method is available for preparing Ag-loading TiO₂ hybrid structure particles. In a typical procedure, 2 g TiO₂ powders were added into 20 ml ethylene glycol in a flask and sonicated for 30 min. An EG solution containing 0.7 M PVP was prepared by adding PVP into EG and stirred for a few minutes at 60 °C until the PVP was dissolved. At the same time, EG solution containing AgNO₃ was prepared. 2 mM NaCl was dissolved in EG solution. All of the solutions above were mixed together in the flask and placed in an oil bath at 160 °C and stirred for 55 min. After the temperature decreased to room temperature, the TiO₂–Ag suspension was obtained which was then rinsed and centrifuged several times with ethanol. At last, the obtained Ag-loaded TiO₂ hybrid structure particles were dried at 80 °C in a vacuum oven for 12 h.

Preparation of the TiO₂@Ag/P(VDF-HFP) films

For the preparation of TiO₂@Ag/P(VDF-HFP) composite films, P(VDF-HFP) was first dissolved in DMF and stirred for 3 h (mass to volume ratio P(VDF-HFP)/DMF = 1 : 8). Then the obtained hybrid nanoparticles were grinded and dispersed into the mixture of DMF and P(VDF-HFP) slowly by ultra-sonication at room temperature for 1.5 h and stirred for 24 hours to form a stable and homogenous suspension. The thin films were made by casting the as-synthesized mixture onto the glass substrate and heated in a vacuum oven at 80 °C for 12 h to remove the residual solvent.

Characterization

The phase composition of synthesized nanopowders was characterized by powder X-ray diffraction (Lab X, SHIMADZU, XRD-6000) analysis, using Cu Kα radiation in the 2θ range of 10–90°. Transmission electron microscopy (TEM) images and energy-dispersive spectroscopy (EDS) line-scan of coated nanoparticles were obtained from a Tecnai G2 F20 S-TWIN instrument operated at an accelerating voltage of 200 kV. X-ray photoelectron spectroscopy (XPS) was recorded using an ESCALAB 250XI X-ray photoelectron spectrometer (Thermo Fisher Scientific, USA). The elemental composition of the samples was determined by energy dispersive X-ray spectroscopy (EDS, Inca, UK). The cross section morphology of the composites was characterized on a SU-70 field emission scanning electron microscope (SEM) instrument with all samples fractured in the liquid nitrogen before testing. Differential scanning calorimetry (DSC) was performed using a PE-DSC 7 analyzer between 50 °C and 200 °C with a heating rate of 10 °C min⁻¹ under a nitrogen atmosphere.

The dielectric properties of the composites were performed on a broadband dielectric spectrometer (NOVOCONTROL GmbH, Germany) with Cu electrodes fabricated on both surfaces of the films. All the measurements were carried out in the frequency range of 10⁻¹ to 10⁷ Hz at several temperatures between 20 °C and 100 °C. The DC leakage current was measured using a Keithley 2400 Source Meter.

Results and discussion

The typical mechanism of reduction of silver ions by ethylene glycol can be represented by the following reactions:

2CH₂OH–CH₂OH → 2CH₃CHO + 2H₂O (1)

2Ag⁺ + 2CH₃CHO → CH₃COOCCH₃ + 2Ag + 2H⁺ (2)

The Ag was loaded on the TiO₂ particles through controlling the reaction rate. In this process, the crystal seeds of silver began to form at room temperature with a low reaction rate in the first step. While the reaction temperature was increased to 160 °C, the reduction of Ag⁺ was promoted. Cl⁻ helped control the concentration of Ag⁺ to reduce the reaction rate and the existence of PVP helped the loading.

Fig. 1(a) shows the XRD patterns of TiO₂@Ag hybrid structure particles. Compared with pure TiO₂ powders, the production shows obvious peaks which correspond to silver, and the diffracted intensity of TiO₂ weakens after reaction. To further confirm the loading results of the composition, TEM analysis was performed on the nanoparticles. The HRTEM observation is shown in Fig. 1(b), from which can be found that both of the components showed crystallized phases. The lattice fringes allow for the identification of interplanar spacing. The fringe spacing of 0.35 and 0.236 nm matches the TiO₂ anatase (101) plane and Ag (111) plane, respectively.^18,19 Energy-dispersive spectroscopy line-scan microanalysis is presented in Fig. 1(d), which describes the composition distribution of the particle. Element Ti was enriched at the big particle with lower contrast. In contrast, Ag was in high content at the small particle with higher contrast. The two elements performed co-existence at the interface, which also indicated a growth of Ag particle over the TiO₂ grain during the synthesis process.

Fig. 1 (a) XRD patterns of TiO₂ and the as-prepared TiO₂@Ag nanoparticles. (b) HRTEM image of the TiO₂@Ag nanoparticle. (c) TEM image of the sample. (d) EDS results of line-scan micro analysis for the TiO₂@Ag nanoparticle.

XPS spectra were recorded to analyse the elemental compositions and chemical status of the obtained samples. Fig. 2(a) shows the XPS survey spectrum of TiO₂@Ag nanoparticles. The sample contains mainly Ti, Ag and O elements. A weak carbon peak is also visible, which may be introduced by the ex situ synthesis process and the transfer of the samples into the UHV chamber.⁸ Fig. 2(b)–(d) display the individual XPS spectra of elemental Ag, Ti, and O, respectively. The peaks observed at 367.8 eV and 373.8 eV (Fig. 2(b)) can be ascribed to Ag 3d_3/2 and Ag 3d_5/2 of the metallic silver. The gap of 6.0 eV between the BE of Ag 3d_5/2 and Ag 3d_3/2 peaks also indicates the metallic nature of Ag 3d states.²⁰ No peak corresponding to silver oxide species was found in XPS and XRD, illustrating that Ag existed as Ag⁰ mainly in the loading hybrid structure and the chemical state of Ag had not been changed during the reaction. The sample exhibits Ti 2p_1/2 (464.9 eV) and Ti 2p_3/2 (459.2 eV) peaks, as shown in Fig. 2(c), which is characteristic of the Ti⁴⁺ oxidation state according to reported XPS data.²¹ The binding energy of 5.7 eV for the splitting Ti 2p peaks also indicated the existence of TiO₂. The Ti 2p XPS spectra of pure TiO₂ are shown in Fig. 2(e). Compared with Fig. 2(c), the binding energy of Ti remained unchanged. The O 1s XPS spectra of the hybrid nanoparticle showed an asymmetric peak (Fig. 2(d)), which could be fitted with a nonlinear least squares fitting program using Gaussian function. After deconvolution, O 1s region was fitted into three peaks. The main peak belongs to Ti–O of TiO₂ (530.4 eV)²² while the minor peak is attributed to the surface hydroxyl group (531.8 eV).²³ Commonly, hydroxyl groups measured by XPS are ascribed to chemisorbed water.²⁴ H₂O was easily adsorbed on the surface of TiO₂ powders during the oil bath heating process. Moreover, this adsorption will be enhanced under the ultrahigh vacuum environment in the XPS system. The minimum peak observed is caused by the C–O group (533.2 eV).²⁵ The bonding energy corresponding to each peak of the O 1s spectra is increased 0.25 eV after reaction, as shown in Fig. 2(f), indicating that the bonding force of Ti–O, O–H, C–O(H) enhanced. Combined with the results from XRD and TEM, it can be concluded that the products consisted of Ag-loaded TiO₂ hybrid structure particles. The elemental composition of the as-prepared TiO₂@Ag nanoparticles was determined by energy-dispersive X-ray spectroscopy (EDS), as exhibited in Fig. S1.† The EDS data further confirm that the particles are composed of Ag, Ti and O, and the molar ratio of Ti/Ag is calculated to be about 4.7 : 1. The C and Pt peaks could be attributed to the carbon adhesive tape and sputtered Pt layer introduced during the SEM specimen preparation process.

Fig. 2 XPS spectra of the nanoparticle (a) survey spectrum (b) Ag 3d region (c) Ti 2p region. (Inset: Ti 2p region of pure TiO₂) (d) O 1s region. (Inset: O 1s region of pure TiO₂).

Fig. 3 shows the SEM images of the freeze-fractured cross sections of the films. Fig. 3(a) exhibits the microstructure of the TiO₂@Ag/P(VDF-HFP) hybrid film at 15 vol% concentration, which shows that the P(VDF-HFP) matrix forms a continuous phase without obvious voids. The TiO₂@Ag nanoparticles are embedded in the P(VDF-HFP) matrix and are dispersed relatively homogenously. Fig. 3(b) illustrates the microstructure of the TiO₂@Ag/P(VDF-HFP) hybrid film with a concentration of 30 vol%. It can be observed that TiO₂@Ag nanoparticles are intensively distributed in the matrix.

Fig. 3 SEM images of the cross-sections of films (a) the film filled with 15 vol% TiO₂@Ag (b) the film filled with 30 vol% TiO₂@Ag. Filled inset: typical film thickness.

DSC was adopted to investigate the effect of TiO₂@Ag nanoparticles on the crystallization behavior of the P(VDF-HFP) matrix after erasing the thermal history. In Fig. 4(a), it can be observed that the melt temperature (T_m) of the nanocomposite is lower than that of the pure P(VDF-HFP). Along with the fillers content increasing, the T_m of the film decreased. It indicates that the nanoparticles promoted the melting process of the P(VDF-HFP) matrix. As is shown in Fig. 4(b), the crystallization temperatures (T_c) first shifted toward higher region then back toward the lower one with increasing TiO₂@Ag contents. This result can be explained by the competition between the heterogeneous nucleation effect and the hindrance of the polymer chain movement.²⁶ It has been proven in previous study that the well-dispersed nanofillers in the polymer matrix can act as nucleation sites to promote the crystallization of polymer matrix without changing its structure.¹ On the other hand, the nanofillers inhibit the movement of the polymer chain segments.²⁷ Hence this phenomenon is a balance of the two effects.

Fig. 4 (a) DSC melting curves (b) DSC cooling curves of pure P(VDF-HFP) and TiO₂@Ag/P(VDF-HFP) nanocomposites with different loading levels of TiO₂@Ag nanoparticles.

Fig. 5(a) presents the dependence of the dielectric permittivity of TiO₂@Ag/P(VDF-HFP) hybrid films on frequency at room temperature. With filler TiO₂@Ag content increasing, the dielectric constant of the composites increases due to the enhancement of Maxwell–Wagner–Sillars (MWS) interfacial polarization, which occurs when there is an accumulation of charge carriers at the interfaces of heterogeneous systems.^28,29 For PVDF-HFP/30 vol% of TiO₂@Ag composite, the dielectric permittivity achieves 26.6 (1 kHz), which is nearly three times larger than that of pure P(VDF-HFP). It is notable that though the silver nanoparticles reach to high loading level and the hybrid fillers have connected as a network, the films keep insulated. In fact, TiO₂ particles in the system separate the Ag nanoparticles from connecting and thus inhibit the formation of conductive network. Although some particle clusters may exist, there is no obvious influence on the film.³⁰

Fig. 5 Frequency dependence of (a) dielectric permittivity (b) dielectric loss (tan δ) and (c) electrical conductivity of TiO₂@Ag/P(VDF-HFP) hybrid films at various volume fractions at room temperature.

Fig. 5(b) shows the dependence of dielectric loss of P(VDF-HFP)/TiO₂@Ag hybrid films on frequency. It can be seen that the dielectric loss of neat P(VDF-HFP) and P(VDF-HFP)/TiO₂@Ag followed the same trend with increasing frequency. The dielectric loss of the composites decreases lightly with increasing frequency at lower frequencies (below 10 kHz), which is attributed to MWS interfacial polarization, and then increases sharply at higher frequencies. The composite films maintained a low dielectric loss (0.037) at low frequencies, even though the filler concentration is as high as 30 vol%. It can be observed that the dielectric losses of samples containing hybrid particles at low frequencies exhibit slight increase compared to neat P(VDF-HFP). It is interesting that the dielectric loss of the composite at high frequency (above 100 kHz) declines with increasing filler concentration. It is generally believed that the dielectric response at high frequency is mainly associated with dipolar relaxation, while interfacial polarization and conductivity are contributed significantly to the dielectric response at low frequency. The dielectric loss peak is attributed to two kinds of molecular motions. One is micro-Brownian moments of the amorphous chain segments and another one is molecular motion on the interfaces (such as amorphous/crystalline interfaces and a/b interfaces).³¹ Thus, the electrical conductivity increase with increasing the filler content, which is presented in Fig. 5(c), is responsible for the increase in dielectric loss of hybrid composites at low frequency. However, at high frequencies, it is helpful to reduce dielectric loss by introducing hybrid nanoparticles in certain content due to the Coulomb blockade and quantum confinement effects of ultra-small Ag nanoparticles.¹⁶ The DC leakage current densities (J_L) are measured to demonstrate the insulating properties of the composite films. As shown in Table S1,† the leakage current density declines at a low loading level of TiO₂@Ag, and then increases as the content of fillers further increases. It is worth mentioning that the leakage current densities are only slightly higher than that of pure P(VDF-HFP) even at a high volume fraction, implying that that electron transport can be prevented by the insulating polymer matrix.

The electric modulus is used to study the dielectric relaxation processes of the composites and investigate the influence of fillers on relaxation behavior of polymer matrix. As shown in Fig. 6, compared with the pure P(VDF-HFP), the interfacial polarization relaxation peaks of TiO₂@Ag/P(VDF-HFP) composites shift to low-frequency. The interfacial polarization in P(VDF-HFP) derives from charge accumulation at the boundary between the lamellar crystal and interlamellar amorphous regions. The shift observed indicates that the ultra-small Ag nanoparticles suppress the space charge accumulation at the boundary. This phenomenon may be caused by the following two reasons. On one hand, the nanofillers performed as pinning center for the polymer chain segment and inhibited its movement. On the other hand, as size of the Ag nanoparticles reduced to a certain value (generally about several nanometers), it could exhibit the Coulomb blockade and quantum confinement effects.³² We would take the example of an individual particle to demonstrate the electrostatic energy change. During the charging process, an electron transfers from a particle to an adjacent particle, resulting in the charge of the first particle by one hole and the second particle by one electron.³³ Here, the energy needed for the electron–hole transfer by tunneling is so-called Coulomb blockade energy. The quantum confinement effect occurs when the diameter of the particle decreases to a size comparable with the wavelength of the electron wave function and it is larger than thermal energy. At the same time, it caused the increase of energy difference between energy states and band gap, which exhibits different electrical properties from usual case. In TiO₂@Ag/P(VDF-HFP) nanocomposites, the ultra-small Ag nanoparticles act as Coulomb islands and hinder the movement of the electrons. They even perform as trapping sites for decelerating electron and space charges, which results in low electrical conductivity.

Fig. 6 Frequency dependence of electrical modulus (M′′) of (a) pure P(VDF-HFP) (b) TiO₂@Ag/P(VDF-HFP) (20 vol%) composites measured at different temperatures.

Fig. 7 shows the frequency dependence of dielectric properties of TiO₂@Ag/P(VDF-HFP) hybrid films incorporated with fillers obtained at various reaction times. As shown in Fig. 7(a), the reaction time has an obvious influence on the dielectric properties of the films. When reaction time is lower than 15 min, the dielectric permittivity of composites rises along with the increasing reaction time. The composite exhibits the highest dielectric permittivity (∼31.3) at 30 vol% of TiO₂@Ag nanoparticles for the reaction time of 15 min. Nevertheless, further increase of the reaction time results in the decline of dielectric constant. It is also observed in Fig. 7(b) that at the low-frequency range, dielectric loss decreases with increasing reaction time. Fig. 8 shows the TEM images of TiO₂@Ag nanoparticles synthesized at different reaction time. As reaction time increases, Ag particles grow bigger and some new ultra-small Ag nanoparticles are produced. It can be inferred that with increasing reaction time, Ag nanoparticles which are already in the system grow and new ultra-small ones appear, providing more trapping sites for decelerating electron and space charges, which results in low dielectric loss. However, if the reaction time is too long, overgrown Ag particles with large sizes would appear, and the abovementioned effect will be weakened.

Fig. 7 Frequency dependence of (a) dielectric permittivity (b) dielectric loss (tan δ) and of TiO₂@Ag/P(VDF-HFP) hybrid films at reaction time concentration at room temperature.

Fig. 8 TEM images of TiO₂@Ag nanoparticles synthesized at different reaction time (a) 5 min (b) 15 min (c) 55 min.

Conclusions

In summary, Ag-loaded TiO₂ hybrid structure particles have been successfully synthesized by an ethylene glycol reduction method as functional fillers. Hybrid films with homogeneously dispersed TiO₂@Ag nanoparticles in a ferroelectric P(VDF-HFP) matrix were initially prepared. Thermal analysis indicated that the introduction of TiO₂@Ag did not disrupt the continuity of the polymer matrix, while the strong interaction between fillers and matrix changed the crystallization temperature of the composite. The adhesion of Ag to TiO₂ efficiently separates the Ag particles from connecting and suppresses the formation of conductive network, and the ultra-small Ag nanoparticles “trap” the carriers due to Coulomb blockade and quantum confinement effect, which results in low dielectric loss and electrical conductivity of composites. As a consequence, the dielectric permittivity of polymer nanocomposites was enhanced by 300% over the P(VDF-HFP) matrix at a filler content of 30 vol% while maintaining a rather low dielectric loss (0.037 at 1 kHz). Our research finds a feasible approach to the fabrication of PVDF-based polymer composites with enhanced dielectric properties, which are supposed to be used as smart functional materials for integrated capacitors.

Acknowledgements

The authors gratefully acknowledge the financial support from the National High Technology Research and Development Program of China (863 Program) (No. 2013AA030701) and the Fundamental Research Funds for the Central Universities (No. 2015QNA4007).

References

1. N. X. Xu, L. Hu, Q. L. Zhang, X. R. Xiao, H. Yang and E. J. Yu, ACS Appl. Mater. Interfaces, 2015, 7, 27373–27381.
2. E. Q. Huang, J. Zhao, J. W. Zha, L. Zhang, R. J. Liao and Z. M. Dang, J. Appl. Phys., 2014, 115, 194102.
3. D. R. Wang, Y. R. Bao, J. W. Zha, J. Zhao, Z. M. Dang and G. H. Hu, ACS Appl. Mater. Interfaces, 2012, 4, 6273–6279.
4. L. J. Fang, W. Wu, X. Y. Huang, J. L. He and P. K. Jiang, Compos. Sci. Technol., 2015, 107, 67–74.
5. M. Li, X. Y. Huang, C. Wu, H. P. Xu, P. K. Jiang and T. Tanaka, J. Mater. Chem., 2012, 22, 23477–23484.
6. H. X. Tang and H. A. Sodano, Appl. Phys. Lett., 2013, 102, 063901.
7. P. H. Hu, Y. Shen, Y. H. Guan, X. H. Zhang, Y. H. Lin and Q. M. Zhang, Adv. Funct. Mater., 2014, 24, 3172–3178.
8. X. R. Xiao, H. Yang, N. X. Xu, L. Hu and Q. L. Zhang, RSC Adv., 2015, 5, 79342.
9. C. Pecharroman, F. E. Betegon, J. F. Bartolome, S. L. Esteban and J. S. Moya, Adv. Mater., 2001, 13, 1541–1544.
10. T. Wei, C. Q. Jin, W. Zhong and J. M. Liu, Appl. Phys. Lett., 2007, 91, 222907.
11. X. W. Liang, S. H. Yu, R. Sun, S. B. Luo and J. Wan, J. Mater. Res., 2012, 27, 991–998.
12. Y. Li, X. Y. Huang, Z. W. Hu, P. K. Jiang, S. T. Li and T. Tanaka, ACS Appl. Mater. Interfaces, 2011, 3, 4396–4403.
13. C. Wu, X. Y. Huang, L. Y. Xie, X. F. Wu, J. H. Yu and P. K. Jiang, J. Mater. Chem., 2011, 21, 17729.
14. Y. Yang, H. L. Sun, D. Yin, Z. H. Lu, J. H. Wei, R. Xiong, J. Shi, Z. Y. Wang, Z. Y. Liu and Q. Q. Lei, J. Mater. Chem. A, 2015, 3, 4916.
15. L. Yang, J. H. Qiu, H. L. Ji, K. J. Zhu and J. Wang, Composites, Part A, 2014, 65, 125–134.
16. J. W. Zha, X. Meng, D. R. Wang, Z. M. Dang and R. Y. Li, Appl. Phys. Lett., 2014, 104, 072906.
17. Z. D. Liu, Y. Feng and W. L. Li, RSC Adv., 2015, 5, 29017.
18. W. H. Gong, Y. H. Chen and L. M. Ke, Trans. Nonferrous Met. Soc. China, 2011, 21, 2044–2048.
19. J. G. Yu, J. F. Xiong, B. Chen and S. W. Liu, Appl. Catal., B, 2005, 60, 211–221.
20. X. C. Wang, J. C. Yu, C. Ho and A. C. Mak, Chem. Commun., 2005, 17, 2262–2264.
21. C. Y. Flore, C. Diaz, A. Rubert, G. A. Benitez, M. S. Moreno, M. A. F. L. de Mele, R. C. Salvarezza, P. L. Schilardi and C. Vericat, J. Colloid Interface Sci., 2010, 350, 402–408.
22. U. Diebole, Methods Surf. Charact., 2006, 4, 145–171.
23. H. Y. Li, D. J. Wang, P. Wang, H. M. Fan and T. F. Xie, Chem.–Eur. J., 2009, 15, 12521–12527.
24. J. G. Yu, L. J. Zhang, B. Cheng and Y. R. Su, J. Phys. Chem. C, 2007, 111, 10582–10589.
25. G. Silversmit, H. Poelman, D. Depla, N. Barrett, G. B. Marin and R. D. Gryse, Surf. Interface Anal., 2006, 38, 1257–1265.
26. S. H. Liu, J. W. Zhai, J. W. Wang, S. X. Xue and W. Q. Zhang, ACS Appl. Mater. Interfaces, 2014, 6, 1533–1540.
27. K. Yang, X. Y. Huang, L. J. Fang, J. L. He and P. K. Jiang, Nanoscale, 2014, 6, 14740.
28. M. Rahimabady, M. S. Mirshekarloo, K. Yao and L. Lu, Phys. Chem. Chem. Phys., 2013, 15, 16242.
29. Q. Wang and L. Zhu, Polym. Phys., 2011, 49, 1421–1429.
30. S. B. Luo, S. H. Yu, R. Sun and C. P. Wong, ACS Appl. Mater. Interfaces, 2014, 6, 176–182.
31. F. X. Guan, J. L. Pan, J. Wang, Q. Wang and L. Zhu, Macromolecules, 2010, 43, 384–392.
32. L. Y. Xie, X. Y. Huang, B. W. Li, C. Y. Zhi, T. Tanaka and P. K. Jiang, Phys. Chem. Chem. Phys., 2013, 15, 17560.
33. I. Balberg, J. Appl. Phys., 2011, 110, 061301.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra08259a

---