Preparation of an oil-absorptive gel containing a 4-t-butyl styrene–EPDM–divinyl benzene graft polymer by reactive processing

Abstract

A 4-t-butyl styrene–ethylene propylene diene monomer–divinyl benzene (tBS–EPDM–DVB) gel was prepared by reactive hot-pressing by reacting an EPDM melt with tBS using DVB and bis(tert-butyl peroxy isopropyl)benzene (BIBP) as a crosslinker and initiator, respectively. The structure and properties of the tBS–EPDM–DVB gel were verified by Fourier transform infrared (FTIR) study, dynamic mechanical analysis (DMA), thermogravimetric analysis (TGA) and scanning electron microscopy (SEM). The results showed that tBS–EPDM–DVB gels have a microphase separated structure, enhanced stiffness and good thermal ability. The maximum value of oil absorbency to chloroform was 15.9 g g⁻¹. The compression set and hardness of the tBS–EPDM–DVB gels both increased with an increasing tBS/EPDM mass ratio. Moreover, the tBS–EPDM–DVB gels also maintained their integrity and showed low compression set values after oil absorption. We conclude that the new reactive processing method to prepare oil absorptive gels has remarkable potential in large-scale industrial applications.

---

1. Introduction

In past decades, oil absorptive polymers with three dimensional cross-linked networks have attracted the attention of scientists and engineers due to their potential use in oil pollution treatment.¹ Among them, alkylacrylate and aromatic polymers with non-polar side chains are the most commonly used materials for preparation of oil absorptive products.^2,3 Ethylene propylene diene monomer (EPDM) rubber has also been used to synthesize a series of oil absorptive gels,⁴ due to its non-polar nature⁵ and excellent resistance to heat, light, oxygen, and ozone.⁶

For example, Zhou et al.^7–9 successfully synthesized a series of EPDM-based oil absorptive polymers. Generally, it was found that the oil absorbency of Zhou’s products was high enough to meet the needs of commercial usage, but their products also suffered from drawbacks such as possessing undesirable shapes, in addition to being inconvenient to process, ship and store. To solve these problems, the same group synthesized bead-shape P(tBS-EPDM) oil gels using suspension polymerization. However, achieving desirable shapes of oil gels was at the expense of decreasing their maximum oil absorbency.¹⁰ And most importantly, the traditional organic solvents used in suspension and solution polymerization have detrimental effects on the environment and human health. Therefore, developing green and low-cost technology for the preparation of oil-absorptive gels through a synthesis route is of great urgency. But to date, most oil-absorptive gels have been synthesized using traditional methods. Recently, Essawy et al.¹¹ prepared an oil-absorptive polymeric nanocomposite through melt blending of EPDM and stearyl acrylate with LAPONITE® as a nanodispersed phase. However, the maximum toluene absorbency of the resulting polymeric nanocomposites was 2.74 g g⁻¹, which should be further enhanced in the future.

In this work, we provide a new reactive processing method to prepare tBS–EPDM–DVB oil absorptive gels, without using organic solvents. To investigate this feasibility, a series of tBS–EPDM–DVB gels were prepared by a reactive hot-pressing method, using tBS and EPDM as monomers, while DVB and BIBP were employed as a crosslinker and initiator respectively. This not only provides an environmentally-friendly and efficient method to produce oil-absorptive gels and reduces manufacturing costs, but also effectively widens the application field of oil absorptive gels since they can be manufactured into various forms of products by hot pressing. Moreover, our work also offers a practical method for the large-scale preparation of oil absorptive gels.

2. Experimental

2.1 Materials

Divinyl benzene (DVB; Fluka) and 4-t-butyl styrene (tBS; Aldrich Chem, USA) were each extracted with a 10% aqueous sodium hydroxide solution and water, dried over anhydrous sodium sulfate, and distilled under reduced pressure prior to use. Ethylene propylene diene copolymer (EPDM 7001) containing 73 wt% ethylene and 5 wt% ethylidenenorbornene was supplied by DuPont Dow Elastomer (USA). All other chemicals and reagents were of analytical grade and were used without further purification.

2.2 Reactive processing procedures

The tBS–EPDM–DVB gel was prepared using the following steps: EPDM was firstly mixed with a given amount of BIBP for approximately 10 min at 80 °C using a double roller mixer. The resulting mixtures were soaked in tBS for different amounts of time to prepare a series of tBS/EPDM/BIBP blends with various mass ratios of tBS to EPDM. Then, a determined amount of DVB crosslinker was added directly into the tBS/EPDM/BIBP blends, and allowed to stand for 24 h under airtight conditions. Finally, the processed samples were hot pressed at 180 °C under 20 MPa for 10 min and then cold pressed at room temperature under 10 MPa for 3 min into sheets of a suitable thickness and size for analysis. For comparison, a pure EPDM sample was also made by the hot and cold pressing process described above. The notations and formulations of samples are listed in Table 1, where ES is the abbreviation for tBS–EPDM–DVB, and the two digits in the notation indicate the tBS content used.

Table 1 Notations and formulations of the tBS–EPDM–DVB gels

Sample	EPDM^a (wt%)	tBS^a (wt%)	DVB^a (wt%)	BIBP^a (wt%)
a The concentration was based on the total weights containing tBS and EPDM monomers.
EPDM	100	0	0	0
ES10	90	10	3	1
ES20	80	20	3	1
ES30	70	30	3	1
ES40	60	40	3	1
ES50	50	50	3	1

---

2.3 Solvent extraction

The residual monomers and the uncross-linked copolymers were removed by extraction with tetrahydrofuran (THF) and hexane in a Soxhlet apparatus, finally the residual gels were dried in a vacuum oven at 60 °C by determining their weight loss until no further significant amount of solvent could be extracted. The content of the gel was calculated using the following eqn (1):

2.4 Fourier transform infrared spectrometer analysis

Fourier transform infrared (FTIR) spectra of the EPDM and tBS–EPDM–DVB gels were recorded on a Nicolet 560 FTIR spectrometer (USA). The spectra were collected for each sample from 4000 to 400 cm⁻¹ at a resolution of 4 cm⁻¹ over 20 scans. The tBS and DVB samples were identified using the KBr disc method, and other tBS–EPDM–DVB gels were analyzed by the film method.

2.5 Dynamic mechanical analysis (DMA)

Dynamic mechanical measurements were carried out on a DMA (TA, Q800) instrument under a nitrogen atmosphere, over a temperature range of −70 °C to 100 °C at a rate of 3 °C min⁻¹, a frequency of 1 Hz, with a three-point bending loading mode at an amplitude of 15 μm. Specimens for the DMA measurements were cut as rods with dimensions of approximately 20 × 10 × 1 mm³. The glass transition temperature (T_g) was defined as the tan δ peak temperature.

2.6 Thermogravimetric analysis

The thermogravimetric analysis (TGA) was performed on a TGA Q50 (TA Instruments) analyzer at a heating rate of 20 °C min⁻¹ from 30 to 600 °C. EPDM and the tBS–EPDM–DVB gels (approximately 10–15 mg) were examined under an air flow.

2.7 Scanning electron microscopy

Scanning electron microscopy (SEM) was used to view the fracture surfaces of all of the samples for comparative assessment of the tBS–EPDM–DVB gel morphology, using a Jeol JSM-5900LV electron microscope at an accelerating voltage of 20 kV. All samples (cross section) were coated with a gold film using an Emscope SM300 coater set at 20 mA for 2 min.

2.8 Oil-absorption tests

Oil-absorptivity of all the tBS–EPDM–DVB gels was determined using ASTM F726-12. A 0.1 g amount of gel was put in a stainless steel mesh (4 × 4 × 2 cm³), and immersed in different organic solvents (hexane, toluene, chloroform, diesel oil, and cyclohexane). The swollen gels were taken out at regular time intervals, wiped with tissue paper gently, weighed on a balance and replaced into the same bath. This procedure of swelling and weighing was continued until the gel achieved a constant final weight. Measurements were carried out in triplicate. The equilibrium oil-absorbency of the cross-linked polymer Q_eq can be calculated using the following eqn (2):where W₀ is the initial weight of the gel and W_e is the weight at equilibrium.

2.9 Compression set measurements

Compression set measurements (ASTM D395)¹² was performed on a circular disk shaped specimen of 29 ± 0.5 mm diameter and 12.5 ± 0.5 mm thickness using a compression molding method. The test specimen was placed between the plates of the compression device with the spacers on each side, allowing sufficient clearance for bulging of the rubber when compressed. The bolts are tightened so that the plates are drawn together uniformly until they are in contact with the spacers. The percentage of the compression employed is 25% of the initial thickness. Then the assembled compression device was placed for 22 h at room temperature. After completion of the compression, the specimen was removed from the device and allowed to recover for 30 min, and the final thickness was measured by an electronic digital caliber with 0.01 mm accuracy. Each result was obtained by repeating the test with three samples. The calculation of compression set follows eqn (3):where t₀ is the initial thickness of the specimen; t₁ is the final thickness of the specimen; and t_s is the thickness of the spacer bar used.

2.10 Hardness measurements

The hardness of all samples was measured using a durometer provided with an “A” scale. The test procedure was in line with ASTM D2240. Reported hardness values are the average of readings taken at five different locations on each sample at room temperature.

3. Results and discussion

Our suggested approach for preparing oil-absorptive gels by reactive processing is illustrated in Fig. 1A. From the point of molecular design, EPDM and tBS, a flexible long chain macromolecule and a rigid short chain macromolecule respectively, were chosen as the main monomers to improve the structure of oil absorptive gels and to enhance the oil absorbency and mechanical properties. A state of homodispersity can be easily formed by melt mixing because of the hydrophobic–hydrophobic interaction of the tBS and EPDM matrix. As a result, in the presence of the radical initiator (BIBP) and crosslinker (DVB), the grafting polymerization of tBS onto EPDM can take place during melt pressing at an elevated temperature caused by the increasing hydrophobic interactions, ultimately leading to the formation of an interpenetrating polymer network.

Fig. 1 (A) Schematic illustration of the process for the preparation of tBS–EPDM–DVB gels by reactive processing. (B) Possible chemical reaction mechanism for the tBS–EPDM–DVB gels.

3.1 Characterization

The structure of the synthesized tBS–EPDM–DVB gels with different tBS content were indicated by the IR spectra. As shown in Fig. 2, the backbone of characteristic sorption bands of EPDM at 723 cm⁻¹ corresponding to a rocking vibration of –(CH₂)_n–, 1378 cm⁻¹ corresponding to a symmetric C–H stretching vibration of –CH₃, 1465 cm⁻¹ corresponding to a scissoring vibration of –CH₂–, 2852 cm⁻¹ corresponding to a stretching vibration of the saturated hydrocarbon backbone of aliphatic symmetric C–H and 2921 cm⁻¹ corresponding to a stretching vibration of the saturated hydrocarbon backbone of aliphatic alkyl asymmetric symmetric C–H can be observed. The characteristic absorption bands of tBS occurred at 838 cm⁻¹ corresponding to a C–H bending vibration of the benzene ring, 902 cm⁻¹ and 989 cm⁻¹ corresponding to a C–H bending vibration of –CH=CH₂, 1630 cm⁻¹ corresponding to a C=C stretching vibration of –CH=CH₂, and 3030 cm⁻¹ and 3087 cm⁻¹ corresponding to a C–H stretching vibration of –CH=CH₂. All peaks mentioned above were observed in the IR spectra of the tBS–EPDM–DVB gels, however, the peaks assigned to the vinyl bending and stretching vibration peak of pure tBS at 902 cm⁻¹, 989 cm⁻¹ and 1630 cm⁻¹ almost disappeared. This observation may suggest that the cross-linking reaction between tBS and EPDM happened during the reactive processing. Furthermore, the appearance of peaks at 693 cm⁻¹ and 773 cm⁻¹, which correspond to the out-of-plane deformation of tBS, further supports the fact that tBS exists in the interior of the tBS–EPDM–DVB gels. Therefore, we conclude that crosslinking tBS–EPDM–DVB gels were achieved through the chemical reaction of EPDM, tBS, and DVB using the reactive processing method. A possible chemical reaction mechanism for the tBS–EPDM–DVB gel is shown in Fig. 1B.

Fig. 2 FTIR spectra of EPDM, tBS, and tBS–EPDM–DVB gels with different tBS content.

3.2 Effect of reaction conditions on the gel fraction of tBS–EPDM–DVB gel

The effects of the mass ratio of tBS to EPDM, and of the BIPB and DVB concentration on the gel fraction of synthesized tBS–EPDM–DVB gel are shown in Fig. 3. Fig. 3A shows the effect of tBS content on the gel fraction of the synthesized gel. The reactions with various tBS content (from 0 to 50 wt% of the total monomer weight) were carried out in BIBP 1 wt% and DVB 3 wt% at 180 °C under 20 MPa for 10 min. The gel fraction decreased with increasing tBS content, which was probably caused by the decreasing diene content of the graft or crosslinking site in EPDM as the EPDM content decreases. On the other hand, the lower the concentration of EPDM, the fewer active centers,¹³ and thus, the higher amount of PtBS homopolymer and poly(tBS-co-DVB) copolymer. Fig. 3B shows the effect of the initiator concentration on the gel fraction of the synthesized gel. The reactions were performed in tBS 30 wt% and DVB 3 wt% at 180 °C under 20 MPa for 10 min. The gel fraction was increased up to 99% with increasing BIBP concentration, and then decreased a little, which may be attributed to the fact that the few active centers exist mainly on the EPDM chains in lower initiator concentrations. When the initiator concentration was too high, more initiators would give more active centers and monomers could move easily and had more chance to be initiated to form a homopolymer.⁸ Fig. 3C shows the effect of DVB concentration on the gel fraction of the synthesized gel. The reactions were performed in tBS 30 wt% and BIBP 1 wt% at 180 °C under 20 MPa for 10 min. As Fig. 3C shows, the gel fraction increased with an increase in the mass ratio of DVB to monomers in the feed. According to the literature,¹⁴ as DVB has two double bonds, a higher DVB content can more easily promote graft polymerization and chemical crosslinking. It is known that the lower the degree of crosslinkage, the more the gel swells,⁸ so that all experiments followed were performed with a moderate DVB content (3 wt%).

Fig. 3 Effect of mass ratio of tBS to EPDM (A), BIPB (B) and DVB (C) concentration on the gel fraction of synthesized tBS–EPDM–DVB gel.

3.3 Dynamic mechanical thermal analysis

All of the samples were subjected to microstructural analysis by means of T_g determination through dynamic mechanical analysis (DMA) (Fig. 4A). As determined by the maximum of the tan δ peak (Fig. 4B), the effect of the tBS content on the T_g of pure EPDM is shown in Fig. 4C, a general trend toward a rise of T_g with the increasing tBS content is perceptible. The increase in T_g may suggest that the cross-linking reaction between tBS and EPDM happened during the polymerization.

Fig. 4 (A) DMA curves of EPDM, ES10, ES20, ES30, ES40 and ES50. (B) tan δ curves of EPDM, ES10, ES20, ES30, ES40 and ES50. (C) The glass transition temperature (T_g) of EPDM, ES10, ES20, ES30, ES40 and ES50, which is defined as the peak of the tan δ curve.

Generally, cross-linking increases the glass transition temperature of a polymer and the change in T_g depends upon the degree of cross-linking, because the bridges between macromolecules impede chain gliding and retard the glass transition.^15,16 It can be deduced that the increased T_g of the tBS–EPDM–DVB gel was due to EPDM and tBS forming a cross-linked network structure. This structure increased the steric hindrance, enhanced the barrier to internal rotation and made the flexibility of the EPDM long main chain decrease.

Fig. 5 shows the dynamic storage modulus as a function of temperature for the EPDM and ES gels with different tBS content, which clearly indicates that the storage modulus (E′) of all the ES gels increased with increasing tBS content. This means that the incorporation of tBS into the EPDM matrix remarkably enhances the stiffness and load bearing capability of the material, which is probably due to the reinforcement effect of a rigid phenyl structure and restrictions in the chain mobility. Furthermore, it should be noted that neat EPDM is readily broken at high temperature (about 100 °C). In contrast, all of the ES gels are more flexible and not broken at 100 °C, which probably due to the formation of networks.

Fig. 5 Storage modulus (E′) curves of EPDM, ES10, ES20, ES30, ES40 and ES50.

3.4 Thermal analysis

The weight loss of the samples due to thermal degradation was obtained from the thermogravimetric analysis (TGA/DTG) results (Fig. 6). These results provided helpful information regarding the composition and phase structure of the samples. As shown in Fig. 6A, the curves show that the initial degradation temperatures of the tBS–EPDM–DVB gels slightly decreased with the increase in the tBS content, which was probably due to the earlier degradation of tBS. In addition, it can be seen that only one major degradation step at around 450 °C involved the decomposition of the tBS–EPDM–DVB gel when the tBS content was no greater than 30 wt% (Fig. 6B). However, the DTG curves show two peaks for ES40 and ES50, which are probably attributed to the intrinsic phase segregation between the hard benzene rings and amorphous EPDM. By comparing both the TGA and DTG traces, it is evident that thermal degradation of EPDM and the tBS–EPDM–DVB gels was different, which also supported the conclusion that grafting of tBS onto the EPDM has occurred.

Fig. 6 TGA/DTG curves of EPDM, ES10, ES20, ES30, ES40 and ES50.

3.5 Morphology

The morphology of the internal sections of the samples was imaged using scanning electron microscopy (SEM), in which the light regions correspond to the hard phase (tBS) and the darker regions correspond to the soft phase (EPDM) (Fig. 7). The SEM images showed the presence of microphase separated structures with a homogenous distribution of hard phase (tBS) in the EPDM matrix and good interface adhesion, supporting the effectiveness of the reactive processing method adopted for the preparation of the tBS–EPDM–DVB gels. The hard phase gradually increased with the increase in the tBS content, indicating an increase in stiffness for all of the ES gels. When the tBS content of tBS–EPDM–DVB gels reached 50 wt%, the fracture surface showed a network-like co-continuous morphology. Moreover, the microscaled hard segments gave rise to an increased surface area to volume ratio and no large pores have been seen in the tBS–EPDM–DVB networks, which are good for oil-absorbing ability¹⁷ and oil holding.¹⁸ These properties make the tBS–EPDM–DVB gels more suitable for oil absorption applications.

Fig. 7 SEM images of the fracture surface morphology of EPDM, ES10, ES20, ES30, ES40 and ES50.

3.6 Oil absorbency

The effect of tBS content on the oil-solvent absorbency is shown in Fig. 8A. It was observed that the oil-solvent absorbency of the tBS–EPDM–DVB gels increased with greater amounts of tBS in the beginning, which could be explained by the fact that the structure of 4-tert-butyl on tBS gives some space effect in the network. The higher the tBS content, the greater the number and size of the cavities in the network, and so more solvent could be held. When tBS content was at 30 wt%, the oil absorbency reached a maximum. When the tBS content surpasses 30 wt%, the oil absorbency decreases, which is probably due to the formation of a dense network structure of the tBS–EPDM–DVB gel. The addition of tBS can increase the polymeric gels’ rigidity which is caused by phenyl group in tBS structure. Oil absorbencies of the tBS–EPDM–DVB gels with different tBS content as a function of chloroform immersion time are shown in Fig. 8B. All curves show an increase with increasing immersion time and level off at 12 h. As the tBS content increased, the oil absorption rate increased at first and then decreased, which correlates well with the effect of tBS content on the oil-solvent absorbency. It is worth mentioning that the pure EPDM sample soon loses its integrity after oil absorption, and finally dissolves in the oil-solvents. Furthermore, the optimum oil absorbency of tBS–EPDM–DVB gel prepared by hot-pressing was lower than that of the tBS–EPDM–DVB gel prepared by suspension polymerization,⁸ which is probably due to the fact that reactive melt processing leads to intensive hydrophobic–hydrophobic interactions resulting in occupation of the cavities or other free spaces between the chains. A similar phenomenon also occurred in the research of Essawy et al., when they prepared an oil-absorptive polymeric nanocomposite through a melt blending method.¹¹ Although the hot-pressing method did not improve the oil absorbency of the tBS–EPDM–DVB gel as well as the suspension polymerization method, it did provide an environment-friendly and efficient method to produce oil-absorptive gels.

Fig. 8 (A) Variation of oil absorbency with different tBS contents in various oil-solvents. (B) Oil absorbency of tBS–EPDM–DVB gels with different mass ratios of tBS to EPDM as a function of chloroform immersion time.

3.7 Compressive properties

To determine the effect of the tBS content on the compression set behavior of the tBS–EPDM–DVB gel, compression set tests were carried out before and after oil absorption. As is shown in Fig. 9, it is clear that the compression set of dry tBS–EPDM–DVB gel increases with increasing tBS content. This is because the crosslinking density of the tBS–EPDM–DVB gel increases with an increase in tBS content, thus decreasing the mobility of the gel chains, which may consequently induce significant stiffness in the resulting gels.¹⁹ As a result of compression of the rubber specimens to a definite amount (25% strain), the enormous crosslinks try to resist the compression which is expressed as an increase in the stiffness of the tBS–EPDM–DVB gels. Some crosslinks have been broken during the resistance, so the number of crosslinks responsible for the strain recovery is fewer than the number of crosslinks responsible for resisting compression when the load is relieved, resulting in the specimen being unable to recover to its initial thickness. As expected, when increasing the crosslinking density by increasing the tBS content, the chance to break more crosslinks increases which leads to a high compression set percentage. The lower the compression set percentage, the better the material resists permanent deformation under a given deflection and temperature. Since a rough compression set value of 80% is linked to the onset of leakage upon application, the material with a low compression set is better for practical use. It is worth mentioning that the tBS–EPDM–DVB gels maintain their integrity after oil absorption and also show low compression set values, indicating that the mechanical properties of the gels are significantly enhanced by the incorporation of rigid tBS.

Fig. 9 Compression set before and after oil absorption of EPDM, ES10, ES20, ES30, ES40 and ES50.

3.8 Hardness test

As shown in Fig. 10, the hardness of ES significantly increased with increasing tBS content, which was mainly due to the fact that the rigidity of the ES network was greatly enhanced by incorporation of tBS hard segments. The trend of changes in the hardness of ES with the increasing of tBS content correlates well with the trends of the storage modulus.

Fig. 10 Hardness of EPDM, ES10, ES20, ES30, ES40 and ES50.

4. Conclusion

In conclusion, an oil absorptive gel was successfully prepared via reactive hot-pressing, in which the crosslinking reaction between tBS and EPDM happened in the presence of an initiator and crosslinker. It was found that the optimized reaction conditions for maximum oil absorbency of tBS–EPDM–DVB gel were 30 wt% tBS, 1 wt% BIBP and 3 wt% DVB. The highest oil absorptivity of the gel was about 15.9 g g⁻¹ in chloroform, and the oil-absorption saturation time was 24 h under the best processing conditions. Despite the fact that the oil absorbency is lower than commercial products, the feasibility of the reactive processing method for the preparation of oil absorptive gels has been demonstrated, which provides a practical way for large-scale preparation of oil absorptive gels.

Acknowledgements

This work was financially supported by the State Key Laboratory of Polymer Materials and Engineering of China (Grant No. sklpme2015-2-06).

Notes and references

1. H. X. Jin, B. Dong, B. Wu and M. H. Zhou, Polym.-Plast. Technol. Eng., 2012, 51, 154.
2. M.-H. Zhou, S.-H. Kim, J.-G. Park, C.-S. Ha and W.-J. Cho, Polym. Bull., 2000, 44, 17.
3. M. H. Zhou and W. J. Cho, J. Appl. Polym. Sci., 2003, 89, 1818.
4. B. Wu, M. Zhou and D. Lu, Iran. Polym. J., 2006, 15, 989.
5. S. Davis, W. Hellens and H. Zahalka, in Polymeric Materials Encyclopedia, ed. J. C. Salamone, CRC Press, New York, 1996, vol. 4, p. 2264.
6. K. G. Park, J. G. Park, C. S. Ha and W. J. Cho, J. Appl. Polym. Sci., 1999, 74, 3259.
7. M. H. Zhou and W. J. Cho, Polym. Int., 2001, 50, 1193.
8. M. H. Zhou and W. J. Cho, J. Appl. Polym. Sci., 2002, 85, 2119.
9. B. Wu and M. H. Zhou, Polym.-Plast. Technol. Eng., 2008, 47, 483.
10. B. Wu and M. H. Zhou, J. Appl. Polym. Sci., 2009, 112, 2213.
11. H. Essawy, M. Essa and Z. Abdeen, J. Appl. Polym. Sci., 2010, 115, 385.
12. ASTM D 395, Standard test method for rubber property-compression set of vulcanized rubber, Annual Book of ASTM Standards, Philadelphia, 1995.
13. K. Srinivasan, N. Radhakrishnan and M. K. Pillai, J. Appl. Polym. Sci., 1989, 37, 1551.
14. A. Matsumoto, Polymeric Materials Encyclopedia, ed. J. C. Salamone, CRC, Boca Raton, FL, 1996.
15. H. Hou, B. Maranesi, J.-F. Chailan, M. Khadhraoui, R. Polini, M. L. Di Vona and P. Knauth, J. Mater. Res., 2012, 27, 1950.
16. M. L. Di Vona, E. Sgreccia, M. Tamilvanan, M. Khadhraoui, C. Chassigneux and P. Knauth, J. Membr. Sci., 2010, 354, 134.
17. Z. Xue, Y. Cao, N. Liu, L. Feng and L. Jiang, J. Mater. Chem., 2014, 2, 2445.
18. Y. Liu, B. Wu, M. Zhou and J. Ding, Mater. Sci. Technol., 2010, 26, 1288.
19. M. van der Schuur and R. J. Gaymans, Polymer, 2005, 46, 6862.

---