Ultra-low dielectric closed porous materials via incorporating surface-functionalized hollow silica microspheres: preparation, interface property and low dielectric performance

Abstract

One effective route to reduce the dielectric constant is to directly incorporate hollow silica (HoSiO₂) microspheres into a polymeric matrix. However, the incompatibility between silica and hydrophobic polymers possibly results in interfacial defects and polarization, and thus a high dielectric loss. In this study, the HoSiO₂ microspheres were coated by polystyrene using a surface-initiated ATRP method in order to enhance interfacial property. TGA results indicated that the weight percentage of polystyrene in resulting microspheres (HoSiO₂@SI-PS) reached around 33 wt%. OM images showed that the thickness of the polystyrene layer reached around 2 μm. HoSiO₂@SI-PS microspheres with a weight percentage of 25% were incorporated into polyethylene (PE) to prepare the composites. The dielectric measurement results indicated that the dielectric constant of the composites was reduced to 2.05, while maintaining low dielectric loss at the level of 10⁻⁴. In comparison, when HoSiO₂@C-PS microspheres, which were prepared by conventional vinyl-initiated free radical polymerization, were incorporated into polyethylene, the dielectric loss was greatly elevated to 0.007. SEM images and water absorption experiments further revealed that the low dielectric loss of PE/HoSiO₂@SI-PS was related to the dense interfacial structure, strong interfacial interaction and low water absorption ability.

---

Introduction

Low-k interconnect dielectrics have received significant attention from the microelectronics industry because their use in integrated circuits (ICs) can lower propagation delays, cross-talk noise, and power dissipation in interconnects. In the past few years, with the continuous scaling down of individual devices within integrated circuits, ultralow dielectric (k < 2.0) materials have become the hot topic in this field.¹

One way to reduce the dielectric constant is incorporating nanopores into a polymeric or inorganic matrix because the dielectric constant of pore is 1.^2–6 In the past few years, a variety of porous low dielectric materials has been extensively studied.^7–21 These porous materials were generally prepared by incorporating decomposable groups or materials followed by thermal/UV or acid/alkaline treatment. The pores were facilely generated by removing the decomposable groups or moieties. Dang et al.²² prepared a nanoporous fluorinated polyimide by an in situ polymerization process in the presence of SiO₂ nanoparticles followed by the removal of SiO₂ nanoparticles using HF acid etching. The dielectric constant was decreased to 2.45. However, dielectric loss was at the level of 10⁻². Loo et al.²³ prepared ultralow dielectric materials derived from poly(D,L-lactide-b-pentafluorostyrene) diblock copolymers. An ordered porous structure was attainable by removing PLA from the nanostructure induced from microphase separation. In particular, the current low-k materials used in volume manufacturing, which are porous organic-added silica, are also prepared based on the above strategy. However, it should be noted that, although low dielectric constant is achievable for most of current porous materials, a high porosity is generally required and their pore structures are uncontrollable, opened or connected. As a result, these porous materials may exhibit poor water resistance, copper barrier performance and high dielectric loss.

The incorporation of hollow particles^24–26 or cage-like molecules^27,28 into a polymeric matrix provides an alternative route to porous materials. As compared with conventional porous materials, this method is facile and enables the formation of closed porous materials.²⁹ Hollow silica particles are ideal candidates with which to prepare this type of porous material because they are industrialized and easily attainable. However, as is well-known, silica particles have a poor compatibility with common polymer matrices, and thus poor interfacial properties. This is definitely unfavorable to maintaining a low dielectric loss of the matrices. A surface-functionalization method has been demonstrated to enhance mechanical, electrical/thermal conducting performance, etc. However, to the best of our knowledge, the effect of surface functionalization on low dielectric performances has not been reported before.

Conventional vinyl-initiated polymerization is intrinsically a copolymerization between vinyl-functionalized SiO₂ and monomers.³⁰ As a result, the homopolymerization of monomers in solution is unavoidable. Thus, the grafting ratio of polymers is relatively low. In the past few years, surface-initiated free radical polymerization has demonstrated wide and important applications in the surface grafting of polymers on inorganic or organic particles.³¹ This method allows the synthesis of polymers in a controlled fashion, resulting in polymers with narrowly dispersed and controlled molecular weights. In this work, hollow SiO₂ microspheres (HoSiO₂) microspheres were surface functionalized by PS using surface-initiated ATRP method. The resulting HoSiO₂@PS microspheres were used to prepare PE composites. In addition, to show the characteristic features of this porous material, other PE composites were prepared by incorporating MPS-functionalized HoSiO₂ and HoSiO₂@PS microspheres prepared by conventional vinyl-initiated free radical polymerization (HoSiO₂@C-PS). The dispersion and interface properties of hollow nanoparticles in PE, the porous structure of the resulting composites, the corresponding low dielectric properties, water resistance and mechanical properties were compared and studied in detail.

Experimental

Chemicals

Hollow silica with an effective mean diameter of 20 μm was purchased from Guangzhou Chaotong Glass Product Trading Company Co. (Guangzhou, China). 3-Methacryloxypropyltrimethoxysilane (MPS) (>97%) and divinylbenzene (DVB, 80% divinylbenzene isomers) were purchased from Aladdin Chemistry Co. Ltd. Polyvinylpyrrolidone (PVP), triethylamine and dimethylaminopyridine (DMAP) were provided by Chengdu Kelong Chemical Reagent Factory. Polyethylene micropowder (PE, AR) was provided by Shanghai Youngling Electromechanical Technology Co., Ltd. (3-Glycidoxypropyl)trimethoxysilane (GPS) and N,N,N′,N′′,N′′-pentamethyldiethylenetriamine (PMDETA) were purchased from Aladdin and used as received. 2-Bromoisobutyryl bromide (2-BriB, 98%) was used as received from Beijing G&K Technology Co., Ltd. Copper(II) bromide (CuBr₂, 99.8%) was provided by Chinese Medicine Group Chemical Reagent Co. Ltd. Styrene with analytical purity was supplied by Chengdu Kelong Chemical Industry Co. (Chengdu, China). Prior to use, styrene was washed with 5% aqueous sodium hydroxide and water, then dried over anhydrous magnesium sulfate and vacuum distilled. 2,2-Azobisisobutyronitrile (AIBN) was purchased from Chengdu Kelong Chemical Industry Co. and was recrystallized from methanol prior to use. Copper(I) chlorine (CuCl, 99.99%) was purchased from Shanghai KeFeng Chemical Industry Co. (Chengdu, China). It was purified by ice acetic acid washing for 12 h followed by recrystallization from ethanol. Polyethylene micropowder (PE, AR) was provided by Shanghai Youngling Electromechanical Technology Co., Ltd.

Preparation of PS grafted HoSiO₂ microspheres by ATRP (HoSiO₂@SI-PS)

The surface-initiated ATRP (SI-ATRP) on HoSiO₂ was conducted according to a reported method which involves three steps: the preparation of hydroxyl-functionalized HoSiO₂, bromo-functionalized HoSiO₂ and surface-initiated polymerization.³² A typical experiment is as follows: the silica suspension (1.6 g HoSiO₂ suspended in 2.4 g NaOH solution (pH = 11)) and (3-glycidoxypropyl)trimethoxysilane (GPS) (6.7 mL, 0.03 mol) were added to a two-necked round bottom flask with a mechanical stir bar and a reflux condenser. The mixture was stirred for 15 min at room temperature, then was refluxed at 100 °C and kept stirring for 24 h. The reaction mixture was then cooled down to room temperature and precipitated into ice methanol. The precipitate was purified by centrifugation in methanol and THF four to five times at 8000 rpm. The solids were dried at 45 °C in a vacuum oven.

The hydroxyl-functionalized HoSiO₂ (OH-HoSiO₂, 1.0 g) was dispersed in 17 mL THF (the concentration was kept below 65 mg mL⁻¹). The solution was charged into a two-necked round bottom flask with a constant pressure funnel and was oscillated for 15 min. Then, triethylamine (1.5 mL) and DMAP (0.001 mol, 0.11 g) were added. The solution was oscillated for 30 min. 2-Bromoisobutyryl bromide (0.007 mol, 0.9 mL) was then added drop by drop at 0 °C. After the addition was completed, the solution was slowly heated up to room temperature and oscillated for 48 h. Then, the colloids were precipitated by adding drop wise to a methanol/H₂O mixture (3 : 1 vol). The bromo-functionalized silica particles (Br-HoSiO₂) were recovered by repeated washing and centrifugation with methanol. The recovered particles were dried overnight at 45 °C in a vacuum oven.

Br-HoSiO₂ (1 g), Cu(I)Cl (0.043 g, 0.43 mmol) and Cu(II)Br₂ (0.008 g, 0.0347 mmol), styrene (23 mL, 0.2 mol) and PMDETA (100 μL, 0.476 mmol) were sequentially added to another flask under a nitrogen atmosphere. The flask was then placed in an oil bath at 90 °C for 6 h. The mixture was diluted with THF and precipitated into an excess of ice methanol followed by filtration and washing. The final silica particles (HoSiO₂@PS) were dried at 45 °C in a vacuum oven.

Preparation of PS-grafted HoSiO₂ by conventional radical polymerization (HoSiO₂@C-PS)

The silica hollow microspheres (3 g) were dried in the oven and dispersed in ethanol (50 mL). The PH of this solution was tuned to 3–4 by adding glacial acetic acid. Subsequently, MPS (0.45 g, 8.9 mmol) was added. The solution was stirred for 2 h at room temperature and was oscillated continuously for 20 h at 30 °C. The mixture was filtered and the filtrate was washed with ethanol three to four times. The MPS-modified silica particles were dried in a vacuum oven at 45 °C.

0.50 g of MPS-HoSiO₂ microspheres were suspended in 35 mL of ethanol. Then, St (0.9 g, 0.0086 mol), PVP (0.06 g, 0.002 mmol) and AIBN (0.02 g, 0.12 mmol) were sequentially added. The reaction was conducted at 65 °C for 4 h. The resultant microspheres were purified by repeated centrifugation in toluene and ethanol and dried in a vacuum oven at 45 °C.

Preparation of PE composites

HoSiO₂@SI-PS microspheres were mixed with PE powders (the weight ratio of PE/HoSiO₂@SI-PS is 3 : 1) at ambient temperature. Subsequently, the mixture was charged into a stainless mold with a diameter of 2.5 cm. The samples were hot pressed at 150 °C for 1.5 h under a pressure of 60 N, and then the film was cooled down to the room temperature. Finally, composited pellets with dimensions of 6.25 cm² areas and about 1.5 mm thickness were obtained. The PE, PE/HoSiO₂, PE/MPS-HoSiO₂ and PE/HoSiO₂@C-PS pellets were prepared using the same method (Scheme 1).

Scheme 1 Preparation routes to PE composites.

Measurements

Fourier transform infrared (FTIR) measurements were conducted on a Nicolet FTIR 5700 spectrophotometer. Scanning electron microscopy analysis (SEM) images were obtained by means of a Zeiss Ultra 55 (Germany) scanning electron microscope, and the microstructure was observed under 15 kV accelerating voltage. Optical microscopy analysis (OM) images was obtained by means of an Olympus BX051 Optical microscope (Japan). Thermogravimetric analysis (TGA) was performed on a TGA Pyris 1, PE in flowing argon at a heating rate of 10 °C min⁻¹. The water adsorption was determined by the following method. The composite films were cut into square bricks with dimensions of 15 mm and immersed in water. The film was taken out at a given time point, then dried and weighed. The water adsorption percentage was calculated by the following equation:

W = (B − G)/G × 100%

where W denotes the water resistance, and B and G denote, respectively, the weight of the film immersed and dried.

Mechanical properties were measured by Dynamic Mechanical Analysis (DMA) on a Q800 (TA Instruments) working in the tensile mode. The sample dimensions were 1.5 (thickness) × 7 (width) × 10 (effective length) mm³ and tests were performed under isochronal conditions at 1 Hz and the temperature was varied between −145 °C and 110 °C at a heating rate of 7 °C min⁻¹. The dielectric permittivity and dielectric loss of the sample pallets were measured by an Agilent 4294A Impedance Analyzer at various frequencies and ambient temperatures. Prior to measurement, the pellets were cut into cubic sheets and Cu electrodes were magnetron sputtered on both sides. The dielectric permittivity was calculated by the following equation:

ε_r = (C × d)/(ε₀ × S)

where C, d and S denote, respectively, the capacitance, thickness and surface area of composite film. ε₀ denotes the permittivity of free space and equals 8.854 × 10⁻¹² F m⁻¹.

Results and discussions

Preparation of PS-grafted HoSiO₂ microspheres

PS-grafted HoSiO₂ microspheres were prepared by surface-initiated atom transfer radical polymerization of styrene on HoSiO₂. This method allowed the synthesis of polymers in a controlled fashion, resulting in polymers with narrowly dispersed and controlled molecular weights. Hollow silica microspheres with a diameter of 20 μm were used. The FTIR spectrum of the particles prepared from ATRP showed characteristic adsorption bands of phenyl and alkyl structures in the ranges of 1400–1600 and 2900–3020 cm⁻¹, respectively (Fig. 1c). This result indicated that PS chains were successfully grafted onto the surface of hollow silica microspheres. The SEM image showed a rough surface of microspheres (Fig. 2d). This phenomenon also supported the surface coating of polystyrene. OM images showed that the thickness of the spherical shell of HoSiO₂ microspheres is about 0.5 μm (Fig. 3a). For HoSiO₂@SI-PS microspheres, this increased to around 2.5 μm (Fig. 3b). Thus, the estimated thickness of the coated polymer was about 2 μm. The TGA curve of HoSiO₂@SI-PS particles showed the characteristic decomposition of PS with T_5% around 380 °C (Fig. 4). The residual weight was around 67.0 wt%, which revealed that the weight percentage of HoSiO₂ was around 67.0 wt%, and thus the weight percentage of grafted PS was around 33.0 wt% (Fig. 5).

Fig. 1 FT-IR spectra of HoSiO₂, HoSiO₂@C-PS and HoSiO₂@SI-PS.

Fig. 2 SEM images of HoSiO₂ (a), MPS-HoSiO₂ (b), HoSiO₂@C-PS (c) and HoSiO₂@SI-PS (d).

Fig. 3 OM images of HoSiO₂ (a) and HoSiO₂@SI-PS (b) microspheres.

Fig. 4 TGA curves of HoSiO₂@C-PS and HoSiO₂@SI-PS.

Fig. 5 SEM images of cross-sections of PE composites. (a) PE/HoSiO₂; (b) PE/MPS-HoSiO₂; (c) PE/HoSiO₂@C-PS; (d) PE/HoSiO₂@SI-PS.

As a comparison, a conventional method, which employs free radical polymerization of vinyl-functionalized HoSiO₂, was also used to prepare HoSiO₂@PS (HoSiO₂@C-PS). The FTIR spectrum of HoSiO₂@C-PS showed an absence of the adsorption band of phenyl groups in the range of 3000–3020 cm⁻¹ and around 1500 cm⁻¹. SEM images of HoSiO₂@C-PS showed a smooth surface like unmodified HoSiO₂. These results implied that only a small amount PS was grafted onto the surface of HoSiO₂. This was further demonstrated by the TGA curves of HoSiO₂@C-PS, which showed a much high residual weight ratio of around 99%. One possible explanation for the low grafting ratio is that the large size of silica microspheres lowers the activity of vinyl groups.

Preparation and interfacial property of PE/HoSiO₂@SI-PS composites

The PE/HoSiO₂@SI-PS composite with HoSiO₂@SI-PS at a weight ratio of 3 : 1 was prepared by mechanical blending, followed by hot pressing. To make a comparison, PE/HoSiO₂, PE/MPS-HoSiO₂ and PE/HoSiO₂@C-PS composites with identical matrix/filler ratios were also prepared (Scheme 1). The morphology of the composites was characterized by SEM images of the cross-section surface. From the SEM image of PE/HoSiO₂@SI-PS, one can see that the HoSiO₂@SI-PS microspheres were uniformly dispersed in PE. No serious destruction of the microspheres was observed. In other words, HoSiO₂ microspheres had adequate mechanical stability to maintain their hollow structure well after hot pressing. As a consequence, the pores were completely closed and isolated. Of particular importance, the cross-section morphology of the PE/HoSiO₂@SI-PS composite showed an apparent irregular plastics deformation morphology and very rough fracture surfaces. The microspheres were well covered by the PE matrix so that they hardly peeled off during fracture of the composites. These phenomena suggested that the PE chain possibly penetrated into PS shells during hot pressing, thus leading to a strong interfacial interaction.

From the SEM images of the PE/HoSiO₂, PE/MPS-HoSiO₂ and PE/HoSiO₂@C-PS composites, it can be seen that a considerable number of microspheres were peeled off. In addition, no plastics deformation and rough fracture surfaces were observed. These results pointed to a weak interfacial interaction. Besides, some microspheres were broken into small pieces; however, no irregular holes were observed. This phenomenon indicated that the breaking of the microspheres should take place during the fracture of the composites rather than in the hot pressing process. Apparently, the microspheres without a PS coating exhibited poor mechanical stability.

Water resistance and mechanical properties of PE/HoSiO₂@SI-PS composites

The water adsorption curves with time are shown in Fig. 6. It was found that the PE/HoSiO₂@SI-PE composite exhibits a much lower water adsorption as compared with that of PE/HoSiO₂. A possible explanation for this is that when HoSiO₂ microspheres are grafted with PS, their surface hydrophibility is greatly reduced. The low water adsorption is also possibly related to the dense interface, which inhibits the diffusion of water. In addition, the water adsorption of the PE/HoSiO₂@SI-PE composite shows an increase for 12 h and almost keeps a constant value (0.2%) after that point. This value is slightly higher than that of PE and is comparable with current industrialized low dielectric materials, demonstrating the good water resistance of PE/HoSiO₂@SI-PE composite.

Fig. 6 Water absorption of PE, PE/HoSiO₂ and PE/HoSiO₂@SI-PS composites with the temperature at 30 °C. The content of the nanoparticles in the composites was 25 wt%.

Fig. 7 shows the storage modulus of the pure PE, PE/HoSiO₂ and PE/HoSiO₂@SI-PS composites. In theory, when incorporating pores in the matrix, the modulus generated is decreased. To our surprise, the storage moduli of the PE/HoSiO₂ and PE/HoSiO₂@SI-PS composites were much higher than that of pristine PE, indicating a higher stiffness. This high stiffness could be ascribed to the high stiffness of the silica structure itself and also to a reinforcing effect of the HoSiO₂ microspheres.

Fig. 7 DMA curves of pure PE, PE/HoSiO₂ and PE/HoSiO₂@SI-PS composites.

Dielectric property of PE/HoSiO₂@SI-PS composites

In most of the current studies, the focus is generally a low dielectric constant of the materials, while the dielectric loss is ignored. In fact, for future low dielectric use in interlayered/interlined packaging and high-frequency related applications, both of these are equally significant. Definitely, in theory incorporating hollow SiO₂ microspheres into a polymer matrix will effectively reduce the dielectric constant. However, at the same time, it will generate a heterogeneous structure or introduce impurities, as a result leading to elevated dielectric loss. Thus, the control of the interfacial structure is important to optimize the dielectric constant and loss. To show the impact of the control of interfacial structure on low dielectric properties, the dielectric properties of PE/HoSiO₂, PE/MPS-HoSiO₂ and PE/HoSiO₂@C-PS, in which the HoSiO₂ has different coating structures, were compared and investigated in this work.

Table 1 summarizes the dielectric constant and loss of the PE/HoSiO₂, PE/MPS-HoSiO₂, PE/HoSiO₂@C-PS and PE/HoSiO₂@SI-PS composites at 10 MHz with a (functionalized) HoSiO₂ weight ratio of 1 : 3. One can find that when HoSiO₂ and modified HoSiO₂ microspheres were incorporated into PE, the dielectric constants were commonly reduced. PE/HoSiO₂@SI-PS composites had relatively high dielectric constants as compared with those of PE/HoSiO₂. As indicated above, the HoSiO₂ in HoSiO₂@SI-PS is around 67 wt%. As a result, the real weight percentage of HoSiO₂ in PE/HoSiO₂@SI-PS was relatively lower than that in PE/HoSiO₂. This means that the porosity of PE/HoSiO₂@SI-PS was definitely lower than that of PE/HoSiO₂, thus leading to a relatively high dielectric constant for the PE/HoSiO₂@SI-PS composite. Besides, although the weight percentages of HoSiO₂ in PE/MPS-HoSiO₂ and PE/HoSiO₂@C-PS composites were close to that in PE/HoSiO₂ (TGA curve of HoSiO₂@C-PS indicated a 99% residual ratio), the dielectric constants of PE/MPS-HoSiO₂ and PE/HoSiO₂@C-PS composites were apparently higher than that of PE/HoSiO₂. One possible explanation for this is that some impurities were absorbed on the surface of HoSiO₂ owing to the presence of acrylate groups and uncovered hydroxyl groups (HoSiO₂ microspheres were treated by acid to form more hydroxyl groups prior to surface functionalization).

Table 1 Dielectric constant and loss of PE, PE/HoSiO₂, PE/MPS-HoSiO₂, PE/HoSiO₂@C-PS and PE/HoSiO₂@SI-PS composites

Samples	Filler/PE weight ratios	Dielectric constant^a	Dielectric loss^a
a Dielectric constant and loss were recorded at 10 MHz.
PE	0	2.40	0.0002
PE/HoSiO2	1 : 3	1.81	0.0014
PE/MPS-HoSiO2	1 : 3	2.14	0.0080
PE/HoSiO2@C-PS	1 : 3	2.10	0.0075
PE/HoSiO2@SI-PS	1 : 3	2.05	0.0008

---

Attractively, the dielectric loss of the PE/HoSiO₂@SI-PS composite, which was at the level of 10⁻⁴, was very close to that of PE. In contrast, the dielectric loss of PE/HoSiO₂@C-PS and PE/HoSiO₂ composites reached the level of 10⁻³. This result indicated that the dielectric loss is linked with the PS shell structure. When HoSiO₂ microspheres were grafted with an adequate amount of PS chains, the interface depolarization can be inhibited to a large degree. In addition, the surface was protected from the absorption of polar matters such as water. This is also beneficial for decreasing dielectric loss.

Conclusions

HoSiO₂ microspheres were surface-functionalized by polystyrene employing a surface-initiated ATRP method. When these PS-coated HoSiO₂ microspheres were incorporated into PE with a weight ratio of 1 : 3, the dielectric constant was reduced from 2.40 to 2.05. More attractively, the dielectric loss can also reach the level of 10⁻⁴. Comparing experiments indicated that when MPS-functionalized HoSiO₂ or derived PS-functionalized HoSiO₂ from conventional vinyl-initiated free radical polymerization was used, the dielectric losses of the resulting PE composites were greatly enlarged. SEM images and water adsorption experiments revealed that the low dielectric loss was linked with dense interfaces and good water resistance. These results pointed to the significance of surface modification on enhancing a low dielectric property. DMA results further showed that PE/HoSiO₂@SI-PS composite had a higher storage modulus as compared with PE. The enhanced low dielectric property, water resistance and mechanical property by employing surface grafting provides a new insight for the structure design of low dielectric materials. Further study on revealing the relationship between grafting density/chain length and low dielectric property is on-going in our group, possibly providing more detailed information on how to design interfacial structure to improve low dielectric performance.

Acknowledgements

This work was supported through a grant from the Open Project of the State Key Laboratory Cultivation Base for Nonmetal Composites and Functional Materials of Southwest University of Science and Technology (11zxfk26), outstanding youth Project of the Southwest University of Science and Technology (13zx9106) and the Innovation Team Project of the Department of Education of Sichuan Province (13TD0022). We thank Dr Pro. Hongtao Yu for discussion and measurement of low dielectric properties.

References

1. W. Volksen, R. D. Miller and G. Dubois, Chem. Rev., 2010, 110, 56–110.
2. Y. Boontongkong and R. E. Cohen, Macromolecules, 2002, 35, 3647.
3. L. M. Bronstein, S. N. Sidorov, P. M. Valetsky, J. Hartmann, H. Col-fen and M. Antonietti, Langmuir, 1999, 15, 6256.
4. R. S. Kane, R. E. Cohen and R. Silbey, Chem. Mater., 1999, 11, 90.
5. V. Sankaran, J. Yue, R. E. Cohen, R. R. Schrock and R. J. Silbey, Chem. Mater., 1993, 5, 1133.
6. B. H. Sohn, J. M. Choi, S. I. Yoo, S. H. Yun, W. C. Zin, J. C. Jung, M. Kanehara, T. Hirata and T. Teranishi, J. Am. Chem. Soc., 2003, 125, 632.
7. H. Zhang, Q. G. Du, J. Zhou, X. L. Zhang, H. T. Wang and W. Zhong, Eur. Polym. J., 2008, 44, 1095–1101.
8. B. F. Zhang and Z. G. Wang, Chem. Mater., 2010, 22, 2780–2789.
9. B. H. Yang, H. Y. Xu, Z. Z. Yang and X. Y. Liu, J. Mater. Chem., 2009, 19, 9038–9044.
10. C. H. Ji and F. Meng, Macromolecules, 2007, 40, 2079–2085.
11. C. M. Lew, R. Cai and Y. S. Yan, Acc. Chem. Res., 2010, 43, 210–219.
12. T. L. Bucholz, S. P. Li and Y. L. Loo, J. Mater. Chem., 2008, 18, 530–536.
13. S. Oh, J. K. Lee, P. Theato and K. K. Char, Chem. Mater., 2008, 20, 6974–6984.
14. G. D. Fu, E. T. Kang, Z. L. Yuan, K. G. Neoh, D. M. Lai and A. C. H. Huan, Adv. Funct. Mater., 2005, 15, 315–322.
15. W. C. Wang, E. T. Kang, R. H. Vora, K. G. Neoh, C. K. Ong and L. F. Chen, Adv. Mater., 2004, 16, 54–57.
16. T. M. Long and T. M. Swager, J. Am. Chem. Soc., 2003, 125, 14113–14119.
17. B. H. Yang, H. Y. Xu, Z. Z. Yang and C. Zhang, J. Mater. Chem., 2010, 20, 2469.
18. T. Fukumaru, T. Fujigaya and N. Nakashima, Polym. Chem., 2012, 3, 369–376.
19. M. Seino, W. Wang, J. E. Lofgreen, D. P. Puzzo, T. Manabe and G. A. Ozin, J. Am. Chem. Soc., 2011, 133, 18082–18085.
20. G. D. Fu, Z. Shang, L. Hong, E. T. Kang and K. G. Neoh, Adv. Mater., 2005, 17, 2622–2626.
21. F. Goethals, I. Ciofi, O. Madia, K. Vanstreels, M. R. Baklanov, C. Detavernier, P. van der Voort and I. van Driessche, J. Mater. Chem., 2012, 22, 8281.
22. J. W. Zha, H. J. Jia, H. Y. Wang and Z. M. Dang, J. Phys. Chem. C, 2012, 116, 23676–23681.
23. T. L. Bucholz, S. P. Li and Y. L. Loo, J. Mater. Chem., 2008, 18, 530–536.
24. G. F. Zhao, T. Ishizaka, H. Kasai, M. Hasegawa, T. Furukawa, H. Nakanishi and H. Oikawa, Chem. Mater., 2009, 21, 419–424.
25. S. P. Mukherjee, D. Suryanarayana and D. H. Strope, J. Non-Cryst. Solids, 1992, 783, 147–148.
26. D. X. Zhuo, A. G. Gu, Y. Z. Wang, G. Z. Liang, J. T. Hu, L. Yuan and W. Yao, Polym. Adv. Technol., 2012, 23, 1121–1128.
27. W. H. Zhang, J. D. Xu, X. S. Li, G. H. Song and J. X. Mu, J. Polym. Sci., Part A: Polym. Chem., 2014, 52, 780–788.
28. Y. W. Chen, L. Chen, H. R. Nie and E. T. Kang, J. Appl. Polym. Sci., 2006, 99, 2226–2232.
29. M. Ree, J. H. Yoon and K. Y. Heo, J. Mater. Chem., 2006, 16, 685–697.
30. G. D. Fua, G. L. Li, K. G. Neohb and E. T. Kang, Prog. Polym. Sci., 2011, 36, 127–167.
31. S. K. Kumar, N. Jouault, B. Benicewicz and T. Neely, Macromolecules, 2013, 46, 3199–3214.
32. G. L. Chakkalakal, M. Alexandre, C. Abetz, A. Boschetti-de-Fierro and V. Abetz, Macromol. Chem. Phys., 2012, 213, 513–528.

---