Morphology control of porous epoxy resin by rod-coil block oligomer: a self-assembly-induced phase separation by diphenyl fluorene-modified silicone epoxy

Abstract

Self-assembly of amphiphilic rod-coil polymers into well-ordered structures has attracted significant interest over the last decade. An especially attractive application of rod-coil polymer self-assembly is the formation of porous materials. A major problem in this process is the expensive synthesis of well-defined block copolymers, and the difficulty of up-scaling for industrial applications. In this study, a robust and economical synthesis route was successfully proposed to prepare a new amphiphilic rod-coil block oligomer combining flexible silicone and rod diphenyl fluorene segments other than diblock copolymers. Porous epoxy monolith was prepared via self-assembly-induced phase separation using silicone epoxy modified by diphenyl fluorene as a porogen. The morphology of the cured resin was examined by SEM and the compatibility and phase separation were studied by SEM-EDS. The results showed that the pore structure of the cured epoxy monolith was controlled by the concentration of the porogen, but the average pore size remained stable. The uniform voids and particles dispersed in the epoxy matrix are of the order of 1 μm or larger.

---

1. Introduction

Due to a wide range of fascinating properties and potential applications, research interest in porous materials is continuously increasing.^1–3 Porous polymers with the advantages of low density, good thermal and electrical insulation and high specific surface area, have found applications in versatile fields such as electrolysis in fuel cells,⁴ separation membranes⁵ and tissue engineering.⁶ Many methods have been developed to produce different morphologies of porous polymers, such as fiber-bonded nonwoven production, thermally and chemically induced phase separation, freeze-drying, gas foaming, porogen leaching, fused deposition modeling and electrospinning. Among these strategies, the method associated with the use of a solvent as porogen is a robust way to manufacture porous monolith, in which the porous morphologies are formed due to a phase separation mechanism induced in a polymer-solvent system. However, due to environmental concerns, the use of solvents and other harmful porogens should be avoided.

Amphiphilic macromolecule self-assembly into well-defined supramolecular structures has attracted a lot of scientific interest over the past decades.^7–10 These ordered structures result from the balance of repulsive interactions between dissimilar segments and the conformational entropy loss of the dissimilar blocks. Moreover, amphiphilic block copolymer self-assembly is very useful for making macroporous polymers, especially materials with tailored pore size, well-defined pore architectures and long-range order. Although self-assembly methodology has many merits, it has limitations. The synthesis of well-defined block copolymers is often hard to scale up and expensive. From an application point of view, the development of simple and scalable procedures for the construction of porous polymers is particularly appealing.

Epoxy resins have been widely applied in polymer industry fields as structural adhesives, coatings, encapsulation material, insulating materials and composite matrix due to their superior electrical and mechanical properties, low shrinkage, good cohesiveness, and excellent moisture and chemical resistance in comparison with many materials.^11–15 Thus, the design of porous epoxy resins has encouraged a lot of scientific and industrial effort to realize this functional technology. In fact, porous epoxy monoliths have been produced by many groups with different methods. Kiefer¹⁶ prepared porous epoxy by use of low molecular weight liquids, Loera¹⁷ obtained porous epoxy by a thermal oxidative degradation method, Guo¹⁸ produced porous epoxy by removing hyperbranched polymer and Liu⁸ designed porous epoxy by the self-assembly of block copolymers. Harmful low molecular weight liquids and expensive amphiphilic block copolymers make porous epoxy difficult for wide applications. The objective of this work is to establish a general methodology to manufacture porous cross-linked epoxy with controlled porosity. The strategy is based on self-assembly-induced phase separation other than using solvents as porogen. Especially, silicone epoxy modified by diphenyl fluorene was designed as the amphiphilic porogen so that silicone can provide the driving force for phase separation and diphenyl fluorene units can maintain compatibility with epoxy matrix. Furthermore, the robust synthesis of the silicone epoxy provides a practical technique for preparing an amphiphilic oligomer porogen by avoiding the expensive synthesis of amphiphilic block copolymers.

2. Experimental

Materials

9,9-Bis-(4-hydroxyphenyl)-fluorene (BHPF), dichlorodiphenylsilane (DPS), 4,4′-diaminodiphenylsulfone (DDS), tetrabutyl ammonium bromide (TBAB) and epichlorohydrin (ECH) were obtained from Sigma-Aldrich. Bisphenol A epoxy resin E51 was provided by Sinopharm Chemical Reagent Co., Ltd. 1,4-Dioxane was distilled over CaH₂ under reduced pressure. Other chemical agents and organic solvents were used without further purification. All reactions were performed under a nitrogen atmosphere.

Characterization

FTIR spectra were recorded on a Nicolet Nexus 470 FTIR spectrometer in the range of 4000–500 cm⁻¹. ¹H-NMR characterization was carried out on a Bruker Ultra Shield Plus-400 NMR spectrometer. DSC measurements were evaluated on a NETZSCH 204C differential scanning calorimeter under a constant flow of nitrogen at 20 mL min⁻¹. The dynamic scanning experiments ranged from 25 to 350 °C at heating rates of 10 °C min⁻¹. Thermogravimetric analysis (TGA) was performed on a TA Q500 thermogravimetric analyzer at a heating rate of 10 °C min⁻¹ from 25 to 700 °C under a nitrogen atmosphere at a flow rate of 60 mL min⁻¹. The dynamic mechanical thermal properties of the epoxy thermosets were carried out with a TA Q800 dynamic mechanical analyzer. The samples (17.5 × 10.0 × 3.17 mm) were loaded in a single-cantilever mode at a heating rate of 3 °C min⁻¹ and a frequency of 1 Hz under an air atmosphere. SEM was used to study the morphology of the samples. The fracture surfaces were obtained by fracturing samples under cryogenic conditions using liquid nitrogen and were examined with a Hitachi S4800 after coating with gold by vapor deposition using vacuum sputtering. Moisture absorption was determined as follows: the rectangular samples (20 × 10 × 4 mm) were dried under vacuum at 100 °C for 12 h, cooled to the ambient temperature, weighed and placed in 100 °C water for a period of time and then re-weighed. The moisture absorption was calculated as the weight gain percentage: moisture absorption% = (W_t − W_o)/W_o × 100%, where W_t is the weight of the sample after immersion in 100 °C water for a period of time, W_o is the initial weight of the sample after placing in a vacuum oven for 12 h.

Synthesis of silicone epoxy resin containing diphenyl fluorene (EGF)

7.008 g (20 mmol) 9,9-bis-(4-hydroxyphenyl)-fluorene and 60 mL anhydrous 1,4-dioxane were added into a round-bottomed flask and heated to 40 °C. Then, 2.10 mL (10 mmol) dichlorodiphenylsilane was added drop wise to the solution. Hydrogen chloride was expelled by bubbling dry nitrogen. The reaction was stirred at 40 °C for 4 hours and heated to 70 °C, followed by addition of 7.83 mL (100 mmol) ECH, 0.645 g (2 mmol) TBAB and 1.6 mL aqueous NaOH solution (50 wt%). The reaction was stirred for another 4 hours and cooled to room temperature. Excess ECH and solvent were removed and the residue was washed with water and ethanol leading to a white precipitate. Pure silicone epoxy resin containing diphenyl fluorene was obtained by re-crystallization from acetone–ethanol (5 : 1, v/v) in 65% yield.

Preparation of the cured epoxy resin casting

Epoxy resins, E51, EGF and E51/EGF mixtures were cured with DDS. EGF was mixed with E51 in different amounts, which formed five curing systems, i.e., E51/DDS, F_0.1E_0.9/DDS, F_0.2E_0.8/DDS, F_0.3E_0.7/DDS, and EGF/DDS, where E and F stand for epoxy resins E51 and EGF, respectively.

Molten epoxy resins were completely degassed, followed by addition of curing agents. Thoroughly mixed epoxy resins were cast into a preheated steel mold (220–230 °C), which was coated with a mold release agent. Degasification was carefully conducted again, and the mixtures were cured in their respective optimal curing conditions, decided by the dynamic DSC traces of epoxy/curing agent compositions. Cured resins were then demoulded and cut into suitable sizes for measurement.

3. Results and discussion

Synthesis of EGF

Scheme 1 presents the chemical structure and the procedure for the synthesis of silicone epoxy resin containing diphenyl fluorene, which is denoted as EGF. Pure EGF was obtained by recrystallization from acetone–ethanol as a white powder. The melting point of EGF is 180 °C, which is determined by DSC.

Scheme 1 Synthesis routes of silicone epoxy resin containing diphenyl fluorene (EGF).

Fig. 1 shows the ¹H NMR spectrum of EGF in CDCl₃. The chemical shifts at 7.34–7.76 are attributed to the protons on fluorine, and the peaks at 6.75–7.19 are attributed to the protons on the phenyl groups. The characteristic chemical shifts of oxirane can be clearly seen at 4.12, 3.91, 3.30, 2.87 and 2.71.¹⁹ Fig. 2 shows the FTIR spectra of the monomer BHPF and the epoxy precursor EGF. The phenolic hydroxyl group absorption peak at 3481 cm⁻¹ decreases because of the reaction of the OH group with epichlorohydrin.²⁰ The characteristic absorption peak of phenyl-oxygen-silicon at 1245 cm⁻¹ is evident,²¹ and the absorption peak of oxirane at 913 cm⁻¹ is also clear. In addition, the peaks located at 1034 cm⁻¹ are attributed to the stretching vibrations of C–O–C, the splitting peaks at 1119 and 1111 cm⁻¹ indicate the phenyl–silicon–phenyl stretching vibration, and the absorption peak of –CH₂– at 2925 cm⁻¹ is observed in Fig. 2. The result of the ¹H-NMR spectrum is in agreement with the FTIR spectrum of EGF, which further confirmed the successful synthesis of the silicone epoxy resin containing diphenyl fluorene (EGF). According to the results of the epoxy equivalent weight titration, the EEW of the synthesized silicone epoxy containing diphenyl fluorene EGF is 500 g eq⁻¹.

Fig. 1 ¹H NMR spectrum of silicone epoxy containing diphenyl fluorene (EGF) in CDCl₃.

Fig. 2 FTIR spectra of BHPF and EGF oligomer.

Morphologies of cured epoxy resins

Scanning electron microscopy (SEM) was used to examine the cryogenically fractured surface of the neat epoxy and EGF/E51 blends to reveal the texture and morphology of the phase-separated system. The glassy fractured surface in the SEM photograph of the neat epoxy resin shows ripples that are due to brittle fracture. Owing to the higher compatibility between the fluorene-epoxy blocks and the neat epoxy, the fracture surface of the sample F_0.1E_0.9 (Fig. 3b) is homogeneous throughout the bulk of the material; however, the fractured surface has been changed significantly. In the cases of F_0.2E_0.8 (Fig. 3c) and F_0.3E_0.7 (Fig. 3d), the fracture surface consists of three distinct phases: globular particles and voids dispersed in a continuous epoxy matrix. More voids and particles were observed in cured EGF (Fig. 3e), but the morphology is still similar to that of F_0.2E_0.8 and F_0.3E_0.7. All the fracture surfaces of modified epoxy networks show an obvious stress-whitened zone. Stress whitening is due to the scattering of visible light from the layer of the scattering centers which in this case are voids. Uniform distribution of the voids throughout the matrix is very important for toughening, as it allows the yielding process to operate throughout the matrix.

Fig. 3 SEM images of the section of cured (a) neat epoxy, (b) F_0.1E_0.9, (c) F_0.2E_0.8, (d) F_0.3E_0.7 and (e) EGF (scale bar, 50 μm).

For typical CTBN-toughened epoxy, the fracture surface consists of dispersed particles and continuous matrix.²² Interestingly, the fracture surface of EGF-modified epoxy resin shows completely different morphologies. At low concentrations of EGF resin, cured epoxy resin consists of continuous matrix and a small amount of voids. With increased amounts of EGF resin, particles appear as a third phase dispersed in the matrix, which suggests precipitated EGF particles in the epoxy matrix. In addition, more and more voids and particles are developed with increased concentrations of EGF resin. Despite the morphology of the cured resins undergoing a dramatic change, the size of the voids and particles remains unchanged. This may result from the greater solubility of EGF resin in the epoxy. Note that voids and particles are of the order of 1 μm or larger. The small size of the discrete particles with a unimodal distribution seems to indicate that phase separation started in the gelation region; thus, particle growth was not possible because of the diffusion restriction existing after gelation of the epoxy matrix.

Void and particle formation model

Si was found by EDS spectra in each cured EGF/E51 matrix, and the Si content in the epoxy matrix shows a slight change when the EGF resin content was increased up to 30% in the formulation (Fig. 5a–c). This indicates that some EGF molecules can dissolve in the epoxy matrix. In addition, the EGF shows a good compatibility with E51 at low concentrations of EGF resin (<10 wt%). This result is in agreement with the morphology of the F_0.1E_0.9 (Fig. 3b). The compatibility should mainly be attributed to the diphenyl fluorene segment. Morphology study shows that the F_0.1E_0.9 sample is homogeneous throughout the bulk of the material (Fig. 3b). In fact, increased magnification shows a small amount of void in the epoxy matrix (Fig. 4b). This tendency of cured composite resin to produce more and more voids with increasing EGF content can be seen in Fig. 3 and 4. The morphology changes of cured epoxy composite resins indicate that excess amounts of EGF resin lead to phase separation. The silicone's low surface energy and, in turn, the huge group of the diphenyl fluorene leads to a strong repulsive interaction between the silicone and fluorene blocks. Hence, the EGF molecules are strongly segregated and assemble into micelles. Voids are formed when silicone groups aggregate at the interior surface of the micelle and particles appear when silicone groups aggregate at the outside surface of the micelle. The migration and aggregation of silicone blocks were confirmed by the EDS results (Fig. 5d–f). The increase in Si in dispersed particles with increasing EGF resin concentration indicates that the particles are composed of EGF and E51. This is consistent with the decrease in S in dispersed particle surfaces (Fig. 5d–f and Table 1). The decrease in S should be attributed to the lower crosslink density since the EGF resin shows a higher epoxy equivalent weight than that of E51. On the other hand, abundant Si aggregated on the surface of the particle hinders the detection of the embedded S inside the particle. Morphology analysis and EDS spectra illustrate the mechanism and the formation process of the phase separation between the EGF resin and epoxy matrix. Three stages can be clearly found as shown in Scheme 2. In the beginning, epoxy resin can accommodate small amounts of EGF (<10 wt%) to form a homogeneous phase. Then, voids originate from the aggregation of excess EGF molecules. In the final stage, silicone-rich particles appear as the third phase in the final morphology of this phase-separation blend.

Fig. 4 SEM micrographs of fracture surface of cured epoxy resins (a) E51, (b) F_0.1E_0.9, (c) F_0.2E_0.8, (d) F_0.3E_0.7 and (e) pure EGF resin (scale bar, 5 μm).

Fig. 5 SEM micrographs and EDS spectra of the epoxy matrix (left) and dispersed particles (right).

Scheme 2 Aggregation and phase separation of cured EGF/E51 resin.

Table 1 Abundance of element S and Si in continuous matrix and dispersed particles

Sample	Matrix		Particle
	S (%)	Si (%)	S (%)	Si (%)
E51	2.31	—	—	—
F0.1E0.9	2.30	2.97	2.78	0.57
F0.2E0.8	2.32	2.35	0.81	2.92
F0.3E0.7	2.35	2.32	0.66	4.88

---

Thermal properties of cured epoxy resins

After curing in a DSC cell up to 300 °C, each sample was allowed to cool to room temperature and subjected to a second run. From the DSC trace, obtained in the second run, the T_g was determined. The T_g of neat epoxy E51 cured with DDS is 218.9 °C and, as expected, modified silicone epoxy has a lower T_g since the silicone blocks increase the flexibility of the system. The depression of the T_g of modified epoxy resin indicates that some silicone blocks are dissolved in the epoxy system. Despite silicone blocks becoming aggregated as micelles with an increase in the concentration of added EGF, the addition of EGF resin reduces the T_g of the cured EGF/E51 blends in a linear fashion. Our observations agree with previous work on similar networks (Fig. 6).^7,8

Fig. 6 Glass transition temperature of the cured EGF/E51 blends and the neat epoxy network. The inset is T_g of cured resins.

The thermal stability and the degradation behavior of the cured systems were investigated by thermogravimetric analysis (TGA) under a nitrogen atmosphere. The weight loss as a function of temperature and the thermal property parameters for the cured epoxy systems are shown in Fig. 7 and Table 2, respectively. T_d,5% and T_d,10% show that E51 epoxy resin has a slightly higher initial degradation temperature than that of mixed resins. This might be attributed to the higher crosslink density of E51 resins. As the chain breakage mainly occurs intermolecularly in the beginning of decomposition, the crosslink density plays an important role in deciding the initial decomposition temperature.²³ On the other hand, the initial degradation temperature of mixed epoxy resin increases with the increase in EGF content. This should be attributed to rigid diphenyl fluorene and Si–O bonds, which have better thermostability. Table 2 shows a higher Y_c value of mixed resins compared with E51/DDS, and Y_c values change consistently with increasing content of EGF, which could obviously be the contribution of the rigid diphenyl fluorene structure to the crosslink network. In addition, the higher Y_c values of EGF cured systems are enhanced owing to the Si–O bond, which has higher bond dissociation energy and better stability in comparison with C–C and C–O bonds.^24–27

Fig. 7 TGA curves of cured EGF/E51 blends and the neat epoxy network.

Table 2 Thermal stability parameters for different epoxy/DDS systems^a

Samples	Td,5% (°C)	Td,10% (°C)	Tmax (°C)	Yc (%, 600 °C)
a Td,5%: temperature of 5% weight loss; Td,10%: temperature of 10% weight loss; Tmax: temperature of maximum rate of weight loss; Yc: char yield under nitrogen atmosphere.
F0.3E0.7/DDS	361.77	377.58	406.83	18.55
F0.2E0.8/DDS	360.78	375.42	402.87	17.42
F0.1E0.9/DDS	358.61	374.06	402.67	15.54
E51/DDS	366.81	378.94	407.60	14.55

---

Dynamic mechanical observations were carried out to analyze the dynamic elastic modulus and the occurrence of molecular mobility transitions such as glass transition temperature.²⁸ Temperature dependence curves of the storage modulus (E′) and tan δ values for different cured epoxy systems are presented in Fig. 8 and 9, respectively. In the glassy region, the E′ value of the cured E51 system was lower than that of modified epoxy. The cured systems with higher content of EGF resins appear to retain their E′ value better than other cured systems. This result might be attributed to the higher rigidity of the diphenyl fluorene skeleton in the chain backbone of EGF resin.²³ The diphenyl fluorene-based polymer (EGF resin) has a structure in which a bulky fluorene unit protrudes vertically from the polymer main chain. This chemical structure of four phenyl rings connected to a quaternary carbon leads to severe rotational hindrance of the phenyl groups, which in turn reduces the rotational mobility of the main chain.²⁹ However, this situation was reversed when it came to the rubbery region. This is consistent with the depression in T_g of the cured resins determined by DSC and DMA. The T_g values (taken as the maximum of the tan δ curve at 1 Hz) of cured E51, F_0.1E_0.9, F_0.2E_0.8 and F_0.3E_0.7 are 236.6, 226.3, 217.2 and 213.8 °C, respectively. With more and more inclusion of EGF resin, the peak of tan δ shifts to lower temperature. This may be attributed to the flexibility of silicone blocks. Another reason for the peak shifting could be the decrease in the cross-linking density of the epoxy upon the incorporation of EGF resin.²⁸

Fig. 8 Storage modulus versus temperature of cured EGF/E51 blends and the neat epoxy network.

Fig. 9 Tan δ versus temperature of cured EGF/E51 blends and the neat epoxy network.

It is well known that absorbed moisture has a detrimental effect on the mechanical properties of epoxy resins, especially at elevated temperature. Therefore, moisture absorption is an important parameter to evaluate whether epoxy resin has high performance or not. Fig. 10 shows the moisture uptake of the cured epoxy resins over varying times. By comparing the curves of different epoxy/diamine systems, it can be clearly seen that the EGF/DDS system had lower moisture absorption than the E51/DDS system, and increasing content of EGF in the mixed resins, the relevant systems increasingly showed lower water uptake ability and took less time to achieve the plateau, which corresponds to the water uptake at equilibrium. The results above showed that the EGF/DDS system possessed a better moisture resistance than the E51/DDS system. It was thought that the decrease in absorbed amount of moisture was attributed to the introduction of silicone and the fluorene ring into epoxy. Silicone is known to possess hydrophobic nature and excellent moisture resistance.^30–32 When the Si–O bond was introduced into epoxy, the system would show better moisture resistance. During the curing process, hydrophilic hydroxyl groups were generated through the ring opening of epoxy groups.^29,33 As the average volume of the fluorene skeleton is larger than that of the methyl group in biphenyl A epoxide, it led to a decrease in free volume and cross-link density in the network of cured systems, which produced fewer hydroxyl groups and resulted in an improved hydrophobicity of the EGF systems. In addition, EGF had a higher content of phenyl groups compared with E51, thus leading to an increase in non-polarity of the polymer molecule, which could further depress the moisture absorption.³⁴ As a result, the EGF/DDS curing system showed significantly lower moisture uptake relative to the E51/DDS curing system and the incorporation of EGF could effectively improve the moisture resistance of the blend-epoxy curing systems.

Fig. 10 Moisture absorption curves of different epoxy/DDS systems.

4. Conclusions

Amphiphilic silicone epoxy containing diphenyl fluorene was successfully synthesized and used as a porogen for cured epoxy resin. The morphology of cured epoxy resin can be easily controlled by the concentration of amphiphilic silicone epoxy modified by diphenyl fluorene. The compatibility between the EGF and epoxy at low loading (<10 wt%) was elucidated by the SEM images. Higher EGF loading led to obvious phase separation and porous morphology. The migration and aggregation of silicone units was observed by SEM-EDS. Self-assembly-induced phase separation establishes a general methodology to prepare a porous epoxy monolith.

Acknowledgements

The authors gratefully acknowledge the financial support from the National Natural Science Foundation of China (no. 51363004), Guangxi Commission of Science and Technology (2012GXNSFAA053212), Guangxi Small Highland Innovation Team of Talents in Colleges and Universities, and Guangxi Funds for Specially-appointed Expert.

Notes and references

1. K. V. Peinemann, V. Abetz and P. F. Simon, Nat. Mater., 2007, 6(12), 992–996.
2. J. H. Lee, L. F. Wang, S. Kooi, M. C. Boyce and E. L. Thomas, Nano Lett., 2010, 10(7), 2592–2597.
3. K. H. Wong, M. H. Stenzel, S. Duvall and F. Ladouceur, Chem. Mater., 2010, 22(5), 1878–1891.
4. Y. Liu, J. Y. Lee, E. T. Kang, P. Wang and K. L. Tan, React. Funct. Polym., 2001, 47(3), 201–213.
5. P. L. Drzal, A. Halasa and P. Kofinas, Polymer, 2000, 41(12), 4671–4677.
6. Y. H. Ng and L. Hong, J. Appl. Polym. Sci., 2006, 102(2), 1202–1212.
7. F. Meng, Z. Xu and S. Zheng, Macromolecules, 2008, 41(4), 1411–1420.
8. F. Meng, S. Zheng, H. Li, Q. Liang and T. Liu, Macromolecules, 2006, 39(15), 5072–5080.
9. M. H. Wang, Y. F. Yu, X. G. Wu and S. J. Li, Polymer, 2004, 45(4), 1253–1259.
10. Y. F. Yu, M. H. Wang, W. J. Gan, Q. S. Tao and S. J. Li, J. Phys. Chem. B, 2004, 108(20), 6208–6215.
11. H. Lee and K. Neville, Handbook of epoxy resins, McGraw-Hill, New York, 1967.
12. G. Lubin, Handbook of composites, Nostrand Reinhold, New York, 1982.
13. P. Chen and D. Z. Wang, Applications of epoxy resins, Academic, Beijing, 2001.
14. P. Khurana, S. Aggarwal, A. K. Narula and V. Choudhary, Polym. Int., 2003, 52(6), 908–917.
15. E. Sharmin, L. Imo, S. M. Ashraf and S. Ahmad, Prog. Org. Coat., 2004, 50(1), 47–54.
16. J. Kiefer, J. G. Hilborn, J. A. E. Manson and Y. Leterrier, Macromolecules, 1996, 29(11), 4158–4160.
17. A. G. Loera, F. Cava, M. Dumon and J. P. Pascault, Macromolecules, 2002, 35(16), 6291–6297.
18. Q. P. Guo, A. Habrard, Y. Park, P. J. Halley and G. P. Simon, J. Polym. Sci., Part B: Polym. Phys., 2006, 44(6), 889–899.
19. M. Pramanik, S. K. Mendon and J. W. Rawlins, Polym. Test., 2012, 31(5), 716–721.
20. H. C. Lin, J. C. Chiang and C. S. Wang, J. Appl. Polym. Sci., 2003, 88(11), 2607–2613.
21. J. H. Lady, G. M. Bower, R. E. Adams and F. P. Byrne, Anal. Chem., 1959, 31(6), 1100–1102.
22. R. Thomas, A. Boudenne, L. Ibos, Y. Candau and S. Thomas, J. Appl. Polym. Sci., 2010, 116(6), 3232–3241.
23. Z. Dai, Y. Li, S. Yang, N. Zhao, X. Zhang and J. Xu, Eur. Polym. J., 2009, 45(7), 1941–1948.
24. L. Pauling, The nature of the chemical bond, Cornell University Press, Ithaca, New York, 3rd edn, 1960.
25. P. H. Sung and C. Y. Lin, Eur. Polym. J., 1997, 33(6), 903–906.
26. C. L. Homrighausen and T. M. Keller, Polymer, 2002, 43(9), 2619–2623.
27. M. Spontón, D. Estenoz, G. Lligadas, C. Ronda, M. Galià and V. Cádiz, J. Appl. Polym. Sci., 2012, 126(4), 1369–1376.
28. J. Y. Lee, J. Jang, S. S. Hwang, S. M. Hong and K. U. Kim, Polymer, 1998, 39(24), 6121–6126.
29. S. Kazama and M. Sakashita, J. Membr. Sci., 2004, 243(1–2), 59–68.
30. J. C. Cabanelas, S. G. Prolongo, B. Serrano, J. Bravo and J. Baselga, J. Mater. Process. Technol., 2003, 143–144, 311–315.
31. U. Lauter, S. W. Kantor, K. Schmidt-Rohr and W. J. Macknight, Macromolecules, 1999, 32(10), 3426–3431.
32. J. Chojnowski, M. Cypryk, W. Fortuniak, M. Ścibiorek and K. Rózga-Wijaset, Macromolecules, 2003, 36(11), 3890–3897.
33. S. Swier, G. Van Assche, W. Vuchelen and B. V. Mele, Macromolecules, 2005, 38(6), 2281–2288.
34. K. Xu, M. C. Chen, K. Zhang and J. W. Hu, Polymer, 2004, 45(4), 1133–1140.

---