Bi₂S₃/poly(vinylidene fluoride) composite with high dielectric constant and unusual low dielectric loss based on preferentially oriented fillers

Abstract

The main challenge in the field of capacitors is how to reconcile the contradiction of lowering the dielectric loss while maintaining a high dielectric constant of polymer based composites. In this work, a novel two-phase ferroelectric polymer composite, consisting of semiconducting bismuth sulfide (Bi₂S₃) nanorods and a poly(vinylidene fluoride) (PVDF) matrix, was fabricated by a sequence of casting, hot-stretching and hot-pressing techniques. Orderly polymer composites based on PVDF assembled with parallel aligned Bi₂S₃ nanorods were realized in the composites after a hot-stretching-pressing process. It’s interesting to note that those orderly polymer composites exhibit excellent dielectric properties. Results show that the composites with oriented structure have high dielectric constant and unusual low dielectric loss. The parallel aligned Bi₂S₃ nanorods along the tensile-strain direction could be responsible for the improvement of dielectric properties of the composites. This study also provides an extremely useful method to reduce the dielectric loss of similar kinds of composites.

---

1. Introduction

Polymer–matrix composites filled with inorganic particles such as ferroelectric ceramic^1,2 and metal particles^3,4 have attracted great attention due to their easy processing and flexibility for various potential applications such as embedded capacitors, actuators and printed circuit boards.^5–7 Xu et al. prepared semiconductor composites by mixing PVDF and various conductive fillers, which were investigated with regard to the dielectric constant as a function of frequency, temperature and the concentration of fillers.⁸ Poly(vinylidene fluoride) (PVDF), which is flexible, easy to process and shows excellent electric breakdown properties, has become the subject of many intensive investigations for its broad applications in electromechanical systems.^9,10

Many studies show that the dielectric constant would be greatly enhanced when metal and/or ceramic fillers were employed in polymer-based composites.^11–15 Dang et al. prepared high-K CCTO/PI functional hybrid films; a high dielectric constant (49) is obtained at 10² Hz when the concentration of the CCTO filler reaches 40 vol%.¹⁶ Researchers note, however, that the reliability is of great importance for a capacitor. Therefore, it is urgent to seek a kind of material which can achieve a high dielectric constant and a low dielectric loss at low content. One important approach to improve the dielectric properties of composites is to incorporate ceramic nanoparticles into the polymer matrix to enhance their relative permittivity. For instance, L. Zhang et al. have prepared BST–P(VDF–CTFE) 0–3 nanocomposites, for the composites with a 40 vol% of BST, a dielectric constant of 70 with a loss of 0.07 at 1 kHz is obtained at room temperature.¹⁷ Sodano et al. reported a BaTiO₃ nanowires/PVDF composite with a dielectric constant of about 44.3, when the concentration of the conducting filler is 30 vol%.¹⁸ However, the volume fraction of ceramic particles is usually very high, which will do great damage to the mechanical properties of the polymer based composites.¹⁹

To overcome the above shortcomings, inorganic conductors are often employed to increase the dielectric constant of the polymer matrix based on the percolation theory.²⁰ However, the corresponding loss and leakage currents of these composites are too high to meet the requirements of realistic applications. The one-dimensional fillers caused wide attention because of their large aspect ratio,²¹ so that the composites with these fillers can reach the percolation threshold at low content. Zhang et al. found that the fitting constant f_c increases with increasing frequency and they believed that this arises from the critical phenomenon, ε_eff ∝ f^γ−1, for composites close to the f_c.²² Nan et al. reported the dielectric properties of PVDF/Bi₂S₃ nanorod (nr-Bi₂S₃) composites, and the dielectric constant could reach 160.²³ However, it is found that the dielectric loss of these percolative composites is usually quite high. This is because the nanoscale fillers with a large aspect ratio are partially connected to each other and form a conductive network with the increase of conducting fillers, which can be blamed for the electric leakage loss. At the low frequency, the electric leakage could be considered to make the major contribution to the dielectric loss.²⁴ As a result, the dielectric loss of these percolative composites is usually high when the percolation threshold is achieved.²⁵ What really impacts the practical use of these percolative composites materials is dielectric loss, which is rarely mentioned and deserves more attention.

Therefore, the main challenge is how to reconcile the contradiction of lowering the dielectric loss while maintaining high dielectric constant. There are many methods to lower the dielectric loss of nanocomposites, such as treating the nanoparticles with coupling agents;^26,27 fabricating nanocomposites via in situ olefin polymerization and employing core@shell structured fillers. Xie et al. successfully lower the dielectric loss of polymer nanocomposites by using core–satellite structured ultra-small silver (Ag) decorated barium titanate (BT) nanoassemblies.²⁸ Nowadays, structure design, as a new method to improve the dielectric properties of polymer composites has drawn wide attention. Our previous results showed that the addition of Bi₂S₃ could enhance the dielectric properties of the PVDF and polyethylene (PE).^29,30 We have also fabricated an orderly polymer composite with zinc flakes (ZFs) oriented parallel in the PVDF, which obtained a high dielectric constant of 176 but a low loss of 0.06 at 10³ Hz.³¹ Along this line, we expect to prepare one-dimensional directional composites by structural design, which have a high dielectric constant and a low dielectric loss at low content.

Tensile strain is usually employed to enhance the tensile strength and modulus of polymer composites.³² Very few studies have focused on the effect of tensile strain on dielectric properties. Yao et al. studied the effect of tensile strain on the dielectric properties of multiwall carbon nanotube (MWNT)/poly(vinylidenefluoride) (PVDF) nanocomposites, the maximum increment in the dielectric constant of the composite can reach 30% at 10² Hz relative to that before stretching.³³ Dang et al. have studied the effect of tensile strain on the morphology and dielectric properties in nanotube/polymer nanocomposites;³⁴ the dielectric constant and conductivity always decreased after the composites were stretched. So in this work, a novel but simple method (hot-stretching and hot-pressing) is developed to lower the dielectric loss of the two-phase Bi₂S₃/PVDF composites while maintaining a high dielectric constant. By designing such a one-dimensional well-aligned parallel structure, a significant reduction in dielectric loss of the novel two-phase Bi₂S₃/PVDF composite is observed, while the dielectric constant remains at a high level at a lower content.

2. Materials and methods

Bi₂S₃ nanorods were synthesized by a hydrothermal method.³⁵ All Bi₂S₃ nanorods in this work were treated by surfactant γ-aminopropyltriethoxysilane (KH550). The composites were prepared by a simple solution-casting, hot-stretching and hot-pressing method. First, a desired amount of Bi₂S₃ nanorods was ultrasonicated in 20 ml N,N-dimethylformamide (DMF) solvent for 30 min and at the same time a certain amount of PVDF was dissolved in N,N-dimethylformamide (DMF) at room temperature. Next, the PVDF solution was added into a Bi₂S₃ suspension solution while the latter was stirred. The mixture was further stirred for 2 h, and ultrasound was applied to the suspension more than one time in order to prevent the agglomeration of Bi₂S₃ nanorods. Then, the composite films were made by the solution casting machine at 80 °C, using glass as the substrate. The resultant films were stretched about 4 times along the casting direction at 160 °C. Finally, the Bi₂S₃/PVDF films were cut into 10 × 10 mm pieces, these little pieces of the oriented samples (O-S) were stacked up along the tensile strain direction (the disoriented samples (D-S) with no direction) in the hot-press mould; the bulk samples were made by a simple hot-pressing method at 200 °C, under 4 MPa (heated for 10 min at the same temperature, pre-pressed for 15 min, released for a while, and then re-pressed for 15 min, followed by cooling to room temperature under the same pressure). For electrical measurement, silver electrodes were painted on both sides of the samples.

The fractured surface morphology of the samples was examined by scanning electron microscopy (FEI Siron 200). The dielectric properties were measured by employing an Agilent 4294A Impedance analyzer in the frequency range of 100 Hz to 10 MHz at room temperature. The out-of-plane dielectric properties of the nanocomposites were studied in this work.

3. Results and discussion

Bi₂S₃ nanorods were synthesized by a hydrothermal method. The SEM micrograph of Bi₂S₃ nanorods and Bi₂S₃/PVDF composites are shown in Fig. 1. From Fig. 1a, the hydrothermal synthesized products mainly consist of various sizes of Bi₂S₃ nanorods with lengths of 0.4–1 μm and diameters of 80–100 nm, and the aspect ratio is about 10.

Fig. 1 SEM images of (a) Bi₂S₃ nanorods; and cross-sectional views of oriented or disoriented Bi₂S₃/PVDF composites with different f_Bi2S3: (b) 2% O-S; (c) 2% D-S; (d) 4% O-S; (e) 2% O-S (another direction); (f) 10% O-S.

The composite films were fabricated by a casting procedure at 80 °C. The cross-section micrograph Fig. 1b and c of the composite samples demonstrate that the Bi₂S₃ nanorods are homogeneously dispersed into the PVDF matrix without serious aggregation at low Bi₂S₃ nanorod content, i.e., f_Bi2S3 = 2 vol%. However, as f_Bi2S3 increases, some Bi₂S₃ aggregates emerge, as shown in Fig. 1d (f_Bi2S3 = 4 vol%) and Fig. 1f (f_Bi2S3 = 10 vol%). The films were stretched along the casting direction at 160 °C. After stretching, the nanorods had preferential orientation in the direction parallel to the tensile strain, which could be clearly seen in the cross-section micrograph Fig. 1b, d, and f.

As shown in Fig. 1e, numerous cross surface Bi₂S₃ nanorods could be observed, and whole nanorods can barely be seen in the picture because this cross-section is perpendicular to the tensile-strain direction. The holes in this picture should be attributed to the extraction of Bi₂S₃ nanorods, because the size of the hole is basically consistent with the diameter of the Bi₂S₃ nanorods and there are no clear holes in Fig. 1b. By contrast, no preferential orientation of Bi₂S₃ nanorods could be observed in the disoriented samples (D-S) (see Fig. 1c). Due to the big difference in the tensile strength and modulus of the PVDF matrix and Bi₂S₃ nanorods fillers, the stretching process will do great damage to the interface between the matrix and the fillers of the nanocomposite and decreased the dielectric properties. But there are no clear defects to be seen in Fig. 1b at low Bi₂S₃ nanorods content, which shows that the hot-stretching process can fix the defects at high temperature above PVDF viscoelastic temperature. In the hot-stretching process, we call this phenomenon the effect of “self-repairing” when the matrix material (PVDF) is at its viscoelastic temperature.

From Fig. 1 we can draw a conclusion that the tensile strain played a crucial role in deciding the morphology of Bi₂S₃/PVDF nanocomposites. After that, a pressure at 200 °C was added along the direction perpendicular to the film plane when the film was cut into pieces, and then were stacked up along the tensile strain direction (the disoriented samples with no direction) into the hot-press mould. The hot-pressing progress could effectively reduce the bubble defect in the composite.

Fig. 2 shows the dependences of dielectric constant, dielectric loss and conductivity of the Bi₂S₃/PVDF composites with different concentrations of Bi₂S₃ nanorods with oriented structure and disoriented structure on frequency at room temperature. A conclusion can be drawn that the dielectric constant, dielectric loss and conductivity of the composites always increased with the increasing Bi₂S₃ nanorods, whether the Bi₂S₃ nanorods are oriented or not. The dielectric constant of the composites always decreased with the increasing frequency, while the conductivity is increased. At a low frequency range, according to the interfacial polarization or Maxwell–Wagner–Sillars effect,³⁶ because of the different relaxation time of the PVDF and Bi₂S₃ nanorods lots of free charge carriers generated by surface polarization are blocked at the interfaces between Bi₂S₃ nanorods and the PVDF matrix, and lots of dipoles form on the Bi₂S₃ nanorods. The inertia of the formed dipoles causes the polarization spend to take much more time than other dielectric processes, thus the interfacial polarization occurs at low frequency and dielectric constant and loss decrease rapidly when the frequency increases. At a high frequency range, the electrical properties of the composites are dominated by the PVDF. Therefore, the dielectric constant is comparatively independent of frequency and is generally lower than that at low frequency, and the dielectric properties of the Bi₂S₃/PVDF nanocomposites change little with tensile strain at high frequency.

Fig. 2 Frequency-dependence of: (a) dielectric constant, (b) dielectric loss and (c) conductivity of the oriented and disoriented Bi₂S₃/PVDF composites with different f_Bi2S3 at room temperature.

Fig. 3 shows the schematic images of how tensile-strain affects the Bi₂S₃ nanorods distribution in polymer composites. The (a) and (a′) are disoriented and oriented samples at low content filling, while (b) and (b′) are disoriented and oriented samples at high content filling, respectively. Because the Bi₂S₃ nanorod was an anisotropic rod material with a large aspect ratio, the nanorods are rearranged along the tensile-strain direction when the Bi₂S₃/PVDF composites are stretched. This mechanism has been verified by the SEM micrographs as shown in Fig. 1.

Fig. 3 Schematic images of Bi₂S₃ nanorods distribution in polymer composites at low content filling [(a) and (a′)] and high content filling [(b) and (b′)] of Bi₂S₃ nanorods with disoriented structure (a) and (b) and oriented structure (a′) and (b′), respectively.

It is well known that the enhancement of the dielectric constant in the low frequency range is mainly due to the contribution of interfacial polarization associated with the entrapment of space charges at the interfaces of the fillers and matrix. Therefore, the anisotropic dielectric constant could be ascribed to the anisotropic intensity of interfacial polarization. According to the theory of dielectrics, the polarization intensity P is defined as³¹

where μ stands for the dipolar moment, V corresponds to the volume of the sample, q and l are denoted as the positive or negative charge and the displacement between positive and negative charges under external electric field, respectively. In the case of interfacial polarization, the positive or negative space charge q and volume V of a given cubic sample are considered as constants independent of direction.

As shown in Fig. 3, after stretching, the distance between the Bi₂S₃ nanorods is greatly reduced in the direction perpendicular to the tensile strain. Therefore, the displacement l of the oriented samples is obviously higher than that of the disoriented samples in the perpendicular direction, which leads to the stronger intensity of interfacial polarization and a higher dielectric constant. Simultaneously, with the increase of Bi₂S₃ nanorods, a fast increase of dielectric constant can be found in the disoriented samples (see Fig. 4a).

Fig. 4 At 1048 Hz, the dependence of (a) dielectric constant, and the inset in (a) presents the best fitting of dielectric constants of disoriented samples according to eqn (1); (b) dielectric loss and (c) conductivity of the oriented and disoriented Bi₂S₃/PVDF composites on the f_Bi2S3, measured at room temperature.

According to percolation theory, when f_Bi2S3 is less than f_c, the dielectric constant (ε) near the percolation threshold can be characterized by the power law,^20,37 as follows:

ε(f_Bi2S3) ∝ (f_c − f_Bi2S3)^−s as f_Bi2S3 < f_c (2)

where s is the dielectric critical exponent. For disoriented samples, the inset in Fig. 4a shows that the linear fit of the experimental data according to eqn (2) and f_c = 10.9% and s = 0.80, and s agrees well with the universal value (0.7–1). As for oriented samples, the percolation threshold is very large, which can be clearly seen in Fig. 4a.

It can be seen from Fig. 4 that, at 1048 Hz, with increasing concentration of Bi₂S₃ nanorods, the differences in dielectric constant and dielectric loss between the oriented and disoriented composites become more significant. At 1048 Hz, for f_Bi2S3 = 4 vol%, the dielectric loss of the disoriented sample is 0.21472 which is about 3 times higher than that of oriented sample (0.05561). However, when f_Bi2S3 increases to 10 vol%, the dielectric loss of the disoriented sample is 1.672, which is almost 9 times of that of the oriented sample (0.1975).

The conductive network is very important to improve the dielectric constant as well as the dielectric loss of composites. When f_Bi2S3 is low, most of the Bi₂S₃ nanorods are fully dispersed in the matrix without serious aggregation, which may be due to the solvent evaporation process. Tensile strain cannot do much under these circumstances, so the change of the dielectric properties of these two series composites is not obvious (see Fig. 4). However, the Bi₂S₃ nanorods begin to connect to each other and form a conductive network with increasing volume content. In this case, the conductive network is destroyed by the tensile strain, because the Bi₂S₃ nanorods prefer to stay along the tensile strain direction and parallel to each other when the tensile strain is employed, which decreases their chance to connect each other.

There is a gradual enhancement of dielectric constant with increasing f_Bi2S3 with low loading of Bi₂S₃ nanorods. However, when f_Bi2S3 > 0.08, both the dielectric constant and dielectric loss of the disoriented composites have a dramatic enhancement, while the oriented composites don’t have. With the increasing Bi₂S₃ nanorods, the conductive phase Bi₂S₃ nanorods began to connect, and form the conductive network where the dielectric constant abruptly increases. When the conductive network formed, the leakage current increased dramatically, so the dielectric loss also abruptly increased. Because of the tensile strain, the phenomenon of oriented composites did not appear. The oriented composites need many more Bi₂S₃ nanorods to form the conductive network, which means the hot-stretching process delayed the percolation threshold of the composites. This could be explained by the rearrangement of Bi₂S₃ nanorods in the composites which destroyed the network structures. In this case, numerous clusters were broken off, and subsequently the Bi₂S₃ nanorods were rearranged along the tensile-strain direction in a preferential orientation. That is, numerous Bi₂S₃ nanorods in the composites were parallel to the tensile-strain direction after stretching. As a result, the dielectric constant and conductivity of the Bi₂S₃/PVDF nanocomposites decreased significantly.

The change in dielectric constant with the filled content of Bi₂S₃ nanorods at different frequencies is shown in Fig. 5. The dielectric constant increases with increasing f_Bi2S3 at all considered frequencies, which is due to the conductivity increase of semiconductive Bi₂S₃ nanorods. As the frequency increases, the dielectric constant increases slowly for these two kinds of composites. At low frequency, the molecules have enough time for polarization. However, at high frequency, the polarization of molecules does not have enough time to catch up with the change in electrical field frequency, which leads to the weak dependence of dielectric constant on the filled content.²⁹ Comparing Fig. 5a with 5b, we can conclude that the dielectric constant of the D-S deceased more intensely than that of O-S, as we have mentioned before.

Fig. 5 The change in dielectric constant with the filled content of Bi₂S₃ nanorods at different frequencies.

4. Conclusions

In summary, orderly polymer composites based on PVDF assembled with parallel aligned Bi₂S₃ nanorods were prepared by the simple and robust technique of a hot stretch-pressing process. It’s interesting to note that those orderly polymer composites indicate excellent dielectric properties with high dielectric constant and unusual low dielectric loss. The dielectric loss of the two-phase Bi₂S₃/PVDF composite decreased after the composites were stretched, while the dielectric constant still remains at a high level. The dielectric loss of the 4 vol% oriented sample (O-S, f_Bi2S3 = 4 vol%) is lowered from 0.21472 to 0.05561 at 1048 Hz due to the orientation of the Bi₂S₃ nanorods. Meanwhile the dielectric constant of the 4 vol% oriented sample still keeps as high as 15.6. The abrupt decrease in the dielectric loss of the percolative composite in the low frequency range could be attributed to the rearrangement of Bi₂S₃ nanorods along the tensile-strain direction in the Bi₂S₃/PVDF nanocomposites. In the hot-stretching process, the nanocomposite materials have the effect of “self-repairing” when the matrix material (PVDF) is at its viscoelastic temperature. This study provides a useful but simple method to reduce the dielectric loss of the composites materials which are constituted by a thermoplastic polymer matrix and fillers with a large aspect ratio.

Acknowledgements

The work was supported by the National Natural Science Fund Innovation Group (No. 51221163), the Research Fund for Doctor Station sponsored by the Ministry of Education of China (20111102110035) and the Fundamental Research Funds for the Central Universities.

References

1. D. Yu, N. Xu, L. Hu, Q. Zhang and H. Yang, J. Mater. Chem. C, 2015, 3, 4016–4022.
2. H. Hao, M. Liu, H. Liu, S. Zhang, X. Shu, T. Wang and Z. Y. A. M. Cao, RSC Adv., 2015, 5, 8868–8876.
3. W. Zhou, J. Zuo and W. Ren, Composites, Part A, 2012, 43, 658–664.
4. D. Bhadra, S. C. Sarkar and B. K. Chaudhuri, RSC Adv., 2015, 5, 36924–36932.
5. W. Jillek and W. K. C. Yung, Int. J. Adv. Manuf. Technol., 2005, 25, 350–360.
6. Z. Dang, J. Yuan, S. Yao and R. Liao, Adv. Mater., 2013, 25, 6334–6365.
7. M. Panda, V. Srinivas and A. K. Thakur, Appl. Phys. Lett., 2011, 99, 42905.
8. H. Xu, H. Xie, D. Yang, Y. Wu and J. Wang, J. Appl. Polym. Sci., 2011, 122, 3466–3473.
9. Y. Zhen, J. Arredondo and G. Zhao, Open J. Org. Polym. Mater., 2013, 03, 99–103.
10. Z. Dang, J. Yuan, S. Yao and R. Liao, Adv. Mater., 2013, 25, 6334–6365.
11. H. A. Ávila, L. A. Ramajo, M. S. Góes, M. M. Reboredo, M. S. Castro and R. Parra, ACS Appl. Mater. Interfaces, 2013, 5, 505–510.
12. Z. Dang, J. Yuan, J. Zha, T. Zhou, S. Li and G. Hu, Prog. Mater. Sci., 2012, 57, 660–723.
13. C. W. Beier, J. M. Sanders and R. L. Brutchey, J. Phys. Chem. C, 2013, 117, 6958–6965.
14. J. Li, S. I. Seok, B. Chu, F. Dogan, Q. Zhang and Q. Wang, Adv. Mater., 2009, 21, 217–221.
15. X. Xiao, H. Yang, N. Xu, L. Hu and Q. Zhang, RSC Adv., 2015, 5, 79342–79347.
16. Z. Dang, T. Zhou, S. Yao, J. Yuan, J. Zha, H. Song, J. Li, Q. Chen, W. Yang and J. Bai, Adv. Mater., 2009, 21, 2077–2082.
17. L. Zhang, P. Wu, Y. Li, Z. Y. Cheng and J. C. Brewer, Composites, Part B, 2014, 56, 284–289.
18. H. Tang, Z. Zhou and H. A. Sodano, ACS Appl. Mater. Interfaces, 2014, 6, 5450–5455.
19. D. Wang, T. Zhou, J. Zha, J. Zhao, C. Shi and Z. Dang, J. Mater. Chem. A, 2013, 1, 6162–6168.
20. C. W. Nan, Y. Shen and J. Ma, Annu. Rev. Mater. Res., 2010, 40, 131–135.
21. R. L. Poveda and N. Gupta, Mater. Des., 2014, 56, 416–422.
22. L. Zhang, P. Bass and Z. Y. Cheng, Appl. Phys. Lett., 2014, 105, 42905.
23. N. Ning, X. Bai, D. Yang, L. Zhang, Y. Lu, T. Nishic and M. Tian, RSC Adv., 2014, 4, 4543–4551.
24. B. P. Sahoo, K. Naskar, R. Choudhary, S. Sabharwal and D. Tripathy, J. Appl. Polym. Sci., 2012, 124, 678–688.
25. M. Panda, V. Srinivas and A. K. Thakur, Appl. Phys. Lett., 2008, 92, 132905.
26. P. Kim, N. M. Doss, J. P. Tillotson, P. J. Hotchkiss, M. Pan, S. R. Marder, J. Li, J. P. Calame and J. W. Perry, ACS Nano, 2009, 3, 2581–2592.
27. P. Kim, S. C. Jones, P. J. Hotchkiss, J. N. Haddock, B. Kippelen, S. R. Marder and J. W. Perry, Adv. Mater., 2007, 19, 1001–1005.
28. L. Xie, X. Huang, B. Li, C. Zhi, T. Tanaka and P. Jiang, Phys. Chem. Chem. Phys., 2013, 15, 17560.
29. Y. Deng, Y. Zhang, Y. Xiang, G. Wang and H. Xu, J. Mater. Chem., 2009, 19, 2058.
30. Y. Deng, N. Li, Y. Wang, Z. Zhang, Y. Dang and J. Liang, Mater. Lett., 2010, 64, 528–530.
31. Y. Zhang, Y. Wang, Y. Deng, Y. Guo, W. Bi, M. Li, Y. Luo and J. Bai, Appl. Phys. Lett., 2012, 101, 192904.
32. Y. Gao, Q. Fu, L. Niu and Z. Shi, J. Mater. Sci., 2015, 50, 3622–3630.
33. S. Yao, J. Yuan, T. Zhou, Z. Dang and J. Bai, J. Phys. Chem. C, 2011, 115, 20011–20017.
34. Z. Dang, S. Yao and H. Xu, Appl. Phys. Lett., 2007, 90, 12907.
35. J. Lu, Q. Han, X. Yang, L. Lu and X. Wang, Mater. Lett., 2007, 61, 3425–3428.
36. Z. Dang, L. Wang, Y. Yin, Q. Zhang and Q. Lei, Adv. Mater., 2007, 19, 852–857.
37. Z. Dang, L. Z. Fan, Y. Shen and C. W. Nan, Chem. Phys. Lett., 2003, 369, 95–100.

---