Rubber nanocomposites via solution and melt intercalation.

Composite materials nowadays have displaced a lot of traditional ones in a variety of applications. Their light weight and enhanced properties are the main reasons for their growth. Optimization of the performance of these materials is a worldwide challenge. The reinforcing particle/matrix adhesion is of great importance, as it controls the final properties of the composite. Polymer-clay nanocomposites meet this demand due to a platelet-type dispersion of the clay in nanometer scale in the polymeric matrix. This field has raised significant scientific interest due to the improved physical properties that can be produced by adding a small fraction of clay (<10 wt. %) to the polymer matrix (refs. 1-5).

The best performance of these composites is achieved when the clay fillers are dispersed in the polymer matrix without agglomeration. An essential step in the formation of a nanocomposite is the delamination of the silicate layers. The layers of about 1 nm thickness consist of a central octahedral sheet of alumina fused between two external tetrahedral silica sheets. The spacing between these layers (galleries) is also at about 1 nm and is occupied by hydrated cations. The latter counter-balance the negative charge generated by the isomorphic substitution of some atoms forming the crystal.

The environment of the galleries is hydrophilic and thus not appropriate for the hydrophobic macromolecular chains to penetrate therein. The replacement of inorganic cations by organic onium ions (surfactant, intercalant) overcomes this character. The cationic head of the alkylammonium compound is tethered to the layers via columbic interactions, leaving the aliphatic tail to hover between the layers. The longer the surfactant chain length and the charge density of the clay, the further apart the clay layers will be forced (ref. 6). The presence of these aliphatic chains in the galleries renders the originally hydrophilic silicate organophilic. A value that characterizes each clay is the cation exchange capacity (CEC), expressed in meq/100 g, referring to the moderate negative charge of a sheet surface. The effect of the cationic exchange capacity has been checked for different types of clay (refs. 7 and 8). It was claimed that a totally exfoliated structure could be obtained with clays of optimum CEC (ca. 90 meq/100 g).

Various models have been developed to trace those parameters that force the macromolecular chain into the silicate layers and cause their delamination afterwards (refs. 9-11). The interplay of entropic and energetic factors determines the outcome of polymer intercalation/exfoliation. The entropic decrease of polymer confinement when entering between the galleries is over-compensated by the increase of the entropy of the tethered chains of the surfactant while the silicate layers are moving apart prior to final separation.

Vulcanized rubber compounds are usually reinforced by carbon black and/or inorganic fillers to improve the mechanical properties. Carbon blacks are excellent in reinforcement owing to the strong interaction with rubbers, but they often decrease the processability of rubber compounds at high volume loading. On the other hand, mineral fillers have a variety of shapes (e.g., fibrous, platy) suitable for reinforcement, but they have very poor interaction with rubber. Therefore, it is interesting to disperse layered silicates in rubber on a nanometer level. The required level of nanoreinforcement can be achieved at very low filler loading, which offers easier processing without any property penalties.

Production of rubber-clay nanocomposites

Note that the organophilic modification of the clay is not always necessary. To produce clay-rubber nanocomposites, the clay can be dispersed in water to which rubber latex is added (ref. 12). Melt intercalation of unmodified montmorillonite has been used in the case of a poly(ethylene oxide) based nanocomposite. The technique includes annealing at 80[degrees]C of a pelletized mixture of PEO and Li montmorillonite (CEC = 80 meq/100 g) for about six hours for intercalation (ref. 13). A theoretical approach for this phenomenon refers that the conditions are fulfilled if the unmodified montmorillonite is melt blended with a polymer mixture comprising the desired polymer matrix and a small amount of end-functionalized polymer, the terminal function of which can interact with the silicate layers (ref. 14).

Polymer layered silicate nanocomposites are currently prepared in the following ways:

* In-situ polymerization;

* intercalation of the polymer from a solution (solution intercalation);

* direct intercalation of the molten polymer (melt intercalation): and

* sol/gel technology.

A schematic representation of the nanocomposite preparation by different routes is given in figure 1 (ref. 15). Since most of the rubbers are available in solid (dry) and latex (solution) forms, melt and solution intercalations are considered to be the ideal methods for preparing rubber nanocomposites.

[FIGURE 1 OMITTED]

Solution intercalation

This method can be applied to dry forms of rubber (which can form solutions with organic solvents), as well as that in the latex form. From the dry form, rubber is dissolved in a suitable solvent along with the organophilic clays, or mixed together after dissolving in suitable solvents. After solvent removal the intercalated material is compounded with curatives and then vulcanized at specific temperature.

Mulhaput and coworkers prepared rubber nanocomposites based on butadiene rubber (BR) and styrene/butadiene rubber (SBR) containing organophilic-layered silicates (refs. 8 and 16). Here, organophilic silicates were swollen in a rubber/ toluene solution. The increase in interlayer distance of the silicates was monitored by wide angle x-ray scattering (WAXS). The interlayer distance increased from 1.26 nm for montmorillonite to 2.59 nm for the organoclay; where it ranged from 3.59 nm to approximately 6 nm for the rubber swollen organoclay. Fully vulcanized nanocomposites were prepared by compounding the rubber-swollen silicates with rubber chemicals in a three-roll, mill followed by vulcanization at 165[degrees]C for 35 minutes in a hot press under vacuum. The excellent dispersion of organoclay (exhibiting intercalated and partially exfoliated layers) in rubber was demonstrated by transmission electron microscopy (TEM) and atomic force microscopy (AFM).

Toyota's research group prepared montmorillonite cation-exchanged with amine terminated butadiene oligomer (ATBN) in a solvent mixture of N,N'-dimethylsulfoxide, ethanol and water (ref. 17). After this, organophilic montmorillonite was blended with acrylonitrile/butadiene rubber (NBR) by roll milling, and the rubber was vulcanized with sulfur. According to TEM observations, the silicate layers were well dispersed in the robber matrix. The tensile stress at 100% elongation of this rubber-clay nanocomposite, containing 10 phr of montmorillonite, was equal to that of the rubber containing 40 phr of carbon black. In this rubber-clay nanocomposite, the permeability of hydrogen and water decreased by 70% by means of adding 3.9 vol. % montmorillonite.

Giannelis et al. presented the relationship between nanostructure and properties in polysiloxane layered silicate nanocomposites (ref. 18). Solvent uptake in dispersed nanostructure was dramatically decreased as compared to conventional composites. The swelling behavior and the modulus were related to the excess amount of bound rubber formed in the nanocomposites compared to the conventional composites.

Mark et al. established the conditions for dispersing clay nanolayers into both cis- 1,4-polyisoprene (IR) and epoxidized natural rubber (ENR) (ref. 19). Incorporation of the clays into the elastomers was achieved by mixing the components in a standard internal mixer or by mixing their dispersions produced in toluene or methyl ethyl ketone solvents. X-ray diffraction studies indicated intercalation of IR and ENR into the silicate interlayers, followed by exfoliation (delamination) of the silicate layers. The reinforcing effect depended strongly on the extent of dispersion of the silicate layers.

Nanocomposites from rubber lattices

As most of the rubbers exist in the latex form and layered silicates can be easily dispersed in water, production of nanocomposites from latices is rather easy. Here, latex could be blended with a clay-water dispersion without coagulation. Wang and coworkers produced nanocomposites from SBR and styrene/vinylpyridene latices by a coagulation method (refs. 12 and 20). Although they noticed some level of intercalation, the properties of the resulting nanocomposites were not promising.

We tried the conventional latex compounding technique lot producing nanocomposites from NR latex (ref. 21). Two types of layered silicates were selected for latex incorporation, namely sodium fluorohectorite (synthetic) and sodium bentonite (natural origin). Dispersions of the layered silicates were prepared and compounded with natural rubber latex, along with dispersions of other latex chemicals, for vulcanization. An inert filler (commercial clay) loaded natural rubber composite was used as the reference material. The major properties of the vulcanized films will be discussed.

Note that there is only a marginal increase in moduli for all composites, except with the fluorohectorite modified one (figure 2). Figure 3 depicts the dynamic mechanical spectra (dynamic storage modulus E') as a function of temperature for the composites. There is a remarkable increase in storage modulus for both of the layered silicate filled composites compared to the commercial clay. The storage modulus below the glass transition temperature ([T.sub.g] = -70[degrees]C) has been increased considerably for fluorohectorite and bentonite when compared to commercial clay.

[FIGURE 3 OMITTED]

The dispersion of layered silicate in the composites was observed by TEM and is illustrated in figures 4a and b, which represent commercial clay and fluorohectorite filled NR composites, respectively. In commercial clay filled composites, the filler exists as large particles. Recall that this clay was not a layered version. The exfoliation and dispersion of fluorohectorite silicate can be better understood from figure 4b. Here, clay layers are visible as regions of dark narrow bands within the polymer (skeleton or house-of-cards structuring). Even though the layers are 'ceramic' in nature, because of their very large aspect ratio and nanometer thickness, they behave mechanically more like sheets of paper rather than rigid plates. This increased flexibility (elastic nature) of the layers con tributes to the elasticity of the rubber. It has been reported that exfoliated clay layers orient along the strain direction in elastomers (ref. 16).

[FIGURE 4 OMITTED]

Melt intercalation

Melt intercalation of high polymers is a powerful new approach to synthesize polymers reinforced by layered silicates. This method is broadly applicable to a range of commodity polymers. Pristine layered silicates usually contain hydrated Na+ or K+ ions. Ion exchange reactions with cationic surfactants, including ammonium ions, render the normally hydrophilic surface organophilic, which makes intercalation of many engineering polymers possible. The role of the alkyl ammonium cations in the organosilicates is to lower their surface energy and to improve their wetting-out by the polymers. Additionally, the alkylammonium cations can provide functional groups that can react with the polymer or initiate a polymerization or grafting of monomers. The outcome is improved strength characteristics, due to a strong interracial interaction between the silicate layers and the polymer.

Organoclay reinforced ethylene/propylene/diene (EPDM) compounds were prepared by mixing EPDM with organoclay, followed by melt blending and vulcanization, and various accelerators were tested (ref. 22). In this EPDM-clay hybrid, the silicate layers were exfoliated using suitable accelerators. This hints that the surfactant of the organoclay might have entered into a complexation reaction with the curatives. The tensile strength of the EPDM-clay, at a clay loading of 4 wt. %, was two times higher than that of the neat EPDM. The gas permeability of the hybrid decreased to 30% as compared with neat EPDM. In our previous work, devoted to studying the curing reaction and mechanical property profile of organoclay reinforced SBR, a strong acceleration effect was found for the organoclay (ref. 23). This was traced to a possible complexation in which zinc, sulfur and amine groups (the latter from the organoclay) are participating. The adverse trends deduced from the curing behavior and mechanical properties were "harmonized" by a model structure. According to the proposed morphology, the clay layers are encapsulated by a more-crosslinked rubber phase and dispersed in a matrix of less-crosslinked rubber.

Epoxidized natural rubber (ENR) nanocomposites

Rubber recipe and its curing

ENR with 50 mol % epoxidation (ENR-50), showing a Mooney viscosity of 140 [ML(1+4) at 100[degrees]C]. was used. As pristine layered clay and silicate, sodium-bentonite (interlayer distance: 1.24 nm) and sodium-fluorohectorite (interlayer distance: 0.94 nm) were used and denoted later as B and F, respectively. Note that the aspect ratio of F is about two-fold that of B (ref. 24). The organoclays were montmorillonite-based ones intercalated by octadecylamine (interlayer distance: 2.10 nm) and by methyl-tallow-bis-2-hydroxyethyl quaternary ammonium salt (interlayer distance: 1.85 nm). For comparison purposes, compounds with amorphous silica (Ultrasil VN2 GR from Degussa) were also made. The remaining compound ingredients are listed in table 1. The rubber mixes were prepared on a laboratory two roll mixing mill. The samples were then cured at 150[degrees]C in an electrically heated hydraulic press to their respective cure times, [t.sub.90]. The [t.sub.90] values were derived from oscillating disc rheometer (MDR) measurements.

Cure characteristics

Table 1 lists the cure parameters derived from the MDR measurements. The cure time ([t.sub.90]) values indicate that both pristine (B, F) and organophilic clays (MMT-ODA, MMT-MTH) accelerated the vulcanization. In the case of the organophilic clays, this effect, observed also for other rubbers (refs. 23 and 25), is likely linked to transition metal complexing in which sulfur and amine-groups of the intercalants participate (ref. 24). This aspect, however, has not been studied yet. The reason behind the t90 reduction for B and F fillers, when compared to S, may be similar, i.e.. of a complexing nature. The cure rate. as defined ([t.sub.90]-[t.sub.2]), is not affected by the silicates, except the MMT-ODA. The increase in maximum and minimum torques, as well as in their difference, compared to the silica-filled compound already suggests some reinforcement for both pristine and organophilic layered silicates. Note that this is possible only if the silicates are intercalated and/or exfoliated.

Dynamic (DMTA) response

Figures 5a and b display the course of the storage modulus (E') and mechanical loss factor (tan [delta]) as a function of temperature (T). The ranking of the fillers in respect to the stiffness in the temperature range below the glass transition temperature of the ENR is: S < MMT-MTH < B< MMT-ODA [approximately equal to] F.

[FIGURE 5 OMITTED]

More interesting is the information one can derive from the tan [delta]-T traces in the glass transition ([T.sub.g]) region. It is noteworthy that the smaller the [T.sub.g] peak, the higher the reinforcing efficiency of the related filler is. Accordingly, a high reinforcement effect can be predicted for fluorohectorite (F) and for the organoclays (MMT-ODA, MMT-MTH). The more or less pronounced doubling and tripling of the Tg peak (as additional peak and/or shoulder on the main peak) suggests that at least a part of the ENR molecules are less mobile. Reduced chain mobility, owing to physical adsorption of the ENR molecules on the filler surface, causes a height reduction of the Tg peak. The multiplication of the Tg peak (that could be explicitly shown by de-convolution techniques) hints at several ENR populations with different chain mobilities. It is intuitive that this behavior should be a product of the silicate intercalation/ exfoliation by ENR. If it is so, then the tan [delta]-T traces may deliver further insight into the intercalation/exfoliation stage of the silicates. Albeit this has been suggested in previous works (refs. 12 and 14), is not yet addressed by studies in detail. In order to shed light on this issue, the dispersion state of the silicates has to be characterized first.

Dispersion of the layered silicates

The XRD spectra of the ENR mixes are depicted in figure 6. As expected, the amorphous silica (S) does not show any crystallographic peak. The rather small peak of bentonite (B) did not change its position, so it is hardly intercalated by ENR (note that the interlayer distance agrees with that of the initial bentonite powder, = 1.24 nm). For the fluorohectorite filled ENR, two peaks are discernible at [approximately equal to] 0.9 nm and 1. 15 nm. The former one agrees with the initial layer distance of F, whereas the latter demonstrates the onset of some intercalation. MMT-MTH possesses three peaks at 1.30, 1.60 and 3.27 nm. Recall that two of them are below the initial value of this organoclay (= 1.85 nm). So, here instead of exfoliation and confinement, re-aggregation of the silicate layers took place. The shoulder at low 2[theta] value (= 3.27 nm), on the other hand, ix evidence of pronounced intercalation with ENR. The XRD trace for the MMT-ODA filled ENR also shows two peaks, one at 1.42 nm and the other at 3.20 nm interlayer distances. Recall that the former interlayer distance is far below the initial value of this organoclay (= 2.10 nm). Accordingly, a part of the organoclay is re-aggregated (a reduction of the layer distance is apparent). Based on the other peak, however, MMT-ODA was definitely intercalated by ENR.

[FIGURE 6 OMITTED]

Figures 7a and b show characteristic TEM pictures of the F-filled ENR. Figure 7 supports that fluorohectorite should exhibit a very high aspect ratio, as the particle length ix ca. 2,000 nm. The image in figure 7a demonstrates further that no delamination of F took place. What figure 7b actually reveals ix that the stacked silicate particles were partially peeled off due to the locally acting shear stress field. ENR molecules could penetrate in between the peeled layers, and their mobility became strongly hampered. The TEM picture in figure 8, taken from the ENR mix with MMT-MTH, seems to be in concert with XRD results. Large silicate aggregates (laying edge-on) and intercalated/exfoliated platelets (laying flat-on in the related TEM pictures) are simultaneously present. The scenario is similar to the ENR nanocomposite with MMTODA (figures 9a and 9b). In this system, exfoliated, intercalated and aggregated organoclay platelets and particles can be found. This finding corroborates the prior conclusions based on the XRD measurements. Note that XRD does not deliver any experimental evidence for exfoliation, as in this case no long range ordering of the silicate layers exists.

[FIGURES 7-9 OMITTED]

The remaining question on the role of the organoclays is: What is the reason for the clay confinement (re-aggregation) when, parallel to that, intercalation and even exfoliation occurs? This could be due to possible interactions between ENR, organophilic intercalant and sulfuric curatives. The epoxy groups of ENR can react, or at least form hydrogen bonds, with both amine (ODA) and hydroxyl-groups (MTH) of the intercalants. This is likely favoring the exfoliation of the or-ganoclay. The amine functionality of the intercalant, when removed from the interlayer to participate in the formation of a zinc-based complex (along with sulfur), should result in a close-up of the silicate galleries. Simultaneous intercalation and confinement were observed also for natural rubber (NR) stocks where reactions via the epoxide groups are excluded (ref. 24). This suggests that for the confinement of the organoclay, the possible reaction between the intercalant and the sulfuric curatives is responsible. If this assumption is correct, then the recipe of sulfur curing, and thus in general the curing system, are crucial in view of the intercalation/exfoliation behavior. This aspect, to which a contribution was already made (ref. 22), has to be studied in the future.

Mechanical properties

The mechanical properties of the ENR (nano)composites are summarized in table 2. The ultimate tensile properties (strength, elongation) do not reflect the tendency expected by considering the dispersion of the silicates. Intercalation/exfoliation of the silicates should result in enhanced strength and reduced elongation, such as is obvious for MMT-ODA. The tear strength, which also proved to be a useful indicator for clay dispersion (ref. 26), gives another ranking for the fillers, viz. S [approximately equal to] B [approximately equal to] MMT-MTH < F < MMT-ODA. Based on the resilience and hardness values, a further ranking can be deduced for the layered silicates. Nevertheless, the best mechanical response was achieved in most cases by MMT-ODA. Note that this silicate was present in exfoliated (TEM), intercalated (XRD) and confined-intercalated (XRD) forms, as well. In view of the above findings, the octadecylamine (primary amine groups) modified clay (MMT-ODA) is far more efficient in ENR stocks than that of clay modified by a methyl-tallow-dihydroxyethyl ammonium compound (quaternary ammonium; MMT-MTH).

Future issues

The development of organoclay/rubber composites is still in its embryonic stage. Thus, the methods that have been practiced are related to the solution (latex) and melt intercalation methods. No direct (in-situ) intercalation has been reported in the open literature, either during rubber synthesis or compounding. Nevertheless, the development in the near future will likely be focused on the melt intercalation techniques. The target issues are: Role of polar rubbers as additives; effects of curing components: sulfurless vulcanization; effects of processing oils; and in-situ melt intercalation (rendering the clay organophilic and achieving its intercalation/exfoliation during mixing/curing). As the research and development activity in the field of nano-reinforcement is far more advanced with thermoplastic resins than with rubber, attention should be paid to the related achievements, which may serve as guidance for the rubber-based nanocomposite development.

Table 1--cure characteristics of ENR mixes
containing 10 phr fillers

MDR 2000 at 150[degrees]C               Silica    Fluorohec-
Parameters                               (S)      torite (F)

Minimum torque, dNm                      0.25        0.27
Maximum torque, dNm                      8.16        8.46
Max.-min. torque, dNm                    7.91        8.19
Scorch time [t.sub.2], min.              9.20        4.38
Cure time [t.sub.90], min.              12.10        7.08
Cure rate [t.sub.90]-[t.sub.2], min.     2.90        2.70

MDR 2000 at 150[degrees]C               Bentonite    MMT-    MMT-
Parameters                                 (B)       ODA     MTH

Minimum torque, dNm                       0.33       0.48    0.52
Maximum torque, dNm                       9.18       9.41    9.39
Max.-min. torque, dNm                     8.85       8.93    8.87
Scorch time [t.sub.2], min.               4.09       2.00    2.29
Cure time [t.sub.90], min.                6.65       3.85    4.94
Cure rate [t.sub.90]-[t.sub.2], min.      2.56       1.85    2.65

ENR(50)--100, filler--10, ZnO--5, stearic acid--2, sodium
carbonate--0.3, sulfur--1.5, N-cyclohexylthiophthalimide--0.2,
N-cyclohexylbenzothiazole-2-sulfenamide--1.5, TMQ antioxidant--1.5

Table 2--mechanical properties of the ENR mixes
containing various fillers (10 phr)

                                  S           B

Properties              Units     Silica     Ben-
                                            tonite

Tensile strength         MPa        16.7      20.6
Tensile modulus at       MPa
100% elongation                      0.9       1.5
200% elongation                      1.4       2.0
300% elongation                      2.0       3.1
Elongation at break       %        946       857
Tear strength            KN/m       27.9      28.0
Resilience                %         38.3      43.9
Hardness                Dur. A      34        40

                                F MMT-ODA MMT-MTH

Properties               Fluoro-     Organo-    Organo-
                        hectorite     clay       clay

Tensile strength             17.0       21.2       18.4
Tensile modulus at
100% elongation               1.2        1.7        1.2
200% elongation               1.9        3.1        1.9
300% elongation               2.8        4.6        2.7
Elongation at break         785        767        888
Tear strength                33.0       41.2       28.7
Resilience                   45.1       45.1       48.8
Hardness                     40         44         40

References

(1.) Y. Kojima, A. Usuki, Y. Kawasumi, A. Okada, T. Kurauchi and O.J. Kamigaito, J. Polym. Sci., Part A: Polym. Chem. 31, 983 (1993).

(2.) K. Yano, A. Usuki, A. Okada, T. Kurauchi and O.J. Kamigaito, J. Polym. Sci., Part A: Polym. Chem. 31, 2,493 (1993).

(3.) E.P. Giannelis, Adv. Mater. 8, 29 (1996).

(4.) Z. Wang and T. Piannavaia, J. Chem. Mater: 10, 3,769 (1998).

(5.) P.B. Messersmith and E.P. Giannelis, J. Polym. Sci., Part A: Polym. Chem. 33, 1,047 (1995).

(6.) P.C. LeBaron, Z. Wang and T. Pinnavaia, J. Appl. Clay Sci., 15, 11 (1999).

(7.) T. Lan, P.D. Kaviratna and T. Pinnavaia, J. Chem. Mater. 7, 2,144 (1995).

(8.) C. Zigl, R. Mulhaupt and J. Finter, Macromol. Chem. Phys. 200, 661 (1999).

(9.) R.A. Vaia and E.P. Giannelis, Macromolecules 30, 7,990 (1997).

(10.) V.V. Ginzbury, C. Singh and A.C. Balazs, Macromolecules 33, 1,089 (2000).

(11.) R.A. Vaia, K.D. Jandt, E.J. Kramer and E.P. Giannelis, Macromolecules 28, 154 (1995).

(12.) L. Zhang, Y. Wang, Y. Wang, Y. Sui and Y. Dingsheng, J. Appl. Polym. Sci. 78, 1,873 (2000).

(13.) R.A. Vaia, S. Vasudevan, W. Krawiec and L.G. Scanlon, E.P. Giannelis, Adv. Mater. 7, 154 (1995).

(14.) A.C. Balazs, C. Singh and E. Zhulina, Macromolecules 31, 8,370 (1998).

(15.) Y. Komori and K. Kuroda, Polymer-Clay Nanocomposites, Ed. by T.J. Pinnavaia and G. W. Beall, John Wiley & Sons, p. 4 (2001).

(16.) M. Ganter, W. Gronski, P. Reichert and R. Mulhaupt, Rubber Chem. Tech. 74, 221 (2001).

(17.) Y. Kojima, K. Fukumori, A. Ususki, A. Okada and T. Kuraychi, J. Mater: Sci. Letters 12, 884 (1993).

(18.) S.D. Burnside and E.P. Giannelis, J. Appl. Polym. Sci., 38, 1,595 (2000).

(19.) T. Yen, J.E. Mark, L.H. Pham and M. Engelhardt, J. Appl. Polym. Sci., 82, 1,391 (2000).

(20.) Y. Wang, L. Zhang, C. Tang and D. Yu, J. Appl. Polym. Sci,. 78, 1,879 (2000).

(21.) S. Varghese and J. Karger-Kocsis, Polymer 44, 4,921 (2003).

(22.) H. Ususki, A. Tukigase and M. Kato, Polymer 43, 2,185 (2002).

(23.) A. Mousa and J. Karger-Kocsis, Macromol. Mater. Eng. 286, 260 (2001).

(24.) S. Varghese and J. Karger-Kocsis, J. Appl. Polym. Sci. (in publication).

(25.) M. Arroyo, M.A. Lopez-Manchado and B. Herrero, Polymer 44, 2,447 (2003).

(26.) C. Nah, H.J. Ryu, S.H. Han, J.M. Rhee and M.H. Lee, Polym. Int. 50, 1,265 (2001).