Gas transport properties in chlorosulfonated polyethylene-acrylate based adhesives.
Gas separation by means of polymer membranes has been for the past two decades among the most dynamic fields of applied science, as research has been able to greatly improve the performance of polymeric materials (1-5). For the last 30 years, development has followed the trend of seeking polymeric structures that in themselves show good permeability and permselectivity properties. Thus, several correlations between structure and gas transport properties have been described (3, 6-10), and it is nowadays well known that chain rigidity enhances selectivity and diminishes permeability, whereas interchain spacing increases the permeability of polymers. In this way, open structures with rigid backbones and bearing polar groups, which selectively enhance the solubility of the gases, have proved to be the most successful polymeric structures for gas separation. In particular, polyimides, polyetherimides, polysulfones and related polymers (11-15) present the best combination of properties, and are used in the formulat ion of numerous commercial polymeric structures.
Some new polymer families with increased spacing and rigidity are being introduced as high performance gas separation materials (16-18), though these structural features cannot be augmented without limit. On the other hand, the polymeric materials that have proved to be the best for gas separation present a series of drawbacks, the most important of which are their price and occasional difficulty in film preparation, as they are often difficult to dissolve and have high glass transition temperatures. Hence, alternative ways to produce materials with good transport properties and that are cheaper and easier to handle are beginning to be studied. Blends of inorganic compounds and polymeric matrices have lately been tested, showing interesting combination of properties (12, 19-21). The effect of miscibility and inmiscibility of polymeric blends is also being explored as a means of modifying the transport coefficients (22, 23). For example, it has been shown that in polycarbonate/polymethyl methacrylate blends, p ermeability and diffusivity increase as the mixture becomes immiscible. This behavior has been related to the effect of the mixing at a molecular level on the local chain motions (22). Carriers are also being extensively used to improve the properties of otherwise "dull" but inexpensive polymers (24). Crosslinking also appears to be a means of improving the properties of polymer membranes (25, 26). An interesting approach to the improvement of the end-properties of materials as gas separators is the use of polymers that together with reasonable selectivity to a particular pair of gases can be prepared in very thin films with good mechanical properties. In this way, the small thickness can compensate for low permeabilities, as the gas flow depends inversely upon thickness. In this connection, photocuring offers several advantages for fabricating membranes. First, photocurable resins can be formulated solvent-free, eliminating voids due to solvent evaporation and other shortcomings regarding the well-known lack of reproducibility caused by the use of solvent cast films for gas separation. As a matter of fact, the type of solvent and the drying process are among the factors of greatest influence on the performance of a gas separation membrane (3). Second, photocuring enables the preparation of layers polymerized onto an adequate porous substrate. UV-photoinduced polymerization has been rarely used to produce membranes for gas separation although it is widely used in other applications, for instance, coatings, adhesives, and inks (27, 28).
In a previous work (29) we presented results on the gas transport properties of a Loctite formulation (Loctite 350[R]) based on aliphatic polyurethanes and a mixture of acrylic and methacrylic monomers (called [L.sub.PU] hereafter), which showed interesting gas separation properties, especially as regards the pair of gases [O.sub.2]/[N.sub.2]. The permselectivity of membranes prepared from this formulation to this pair of gases is about 4.6 at 25[degrees]C, and it increases at lower temperatures. Because of their other special properties, adhesivity and the possibility to coat a variety of supports with fine layers, survey through these polymers appears interesting.
In this work, the results obtained with a crosslinked system based on Loctite 329[R] are presented. This is an acrylic adhesive characterized by its high toughness, which makes it very suitable for use in strength structural joints or in sheet steel where continuous or repeated loads are generated. The rough material is composed of a chlorosulfonated polyethylene binder (CSPE) and a mixture of different acrylic and methacrylic monomers, which, upon polymerization, lead to a crosslinked network. The chlorine atoms and the chorosulfonic groups are reactive centers that generate, under certain conditions, crosslinking points between the binder chain and the polymerizing acrylic macroradical. This contributes to its mechanical strength and elastic properties when fully cured. This adhesive is usually used as a two-component thermally curable adhesive, by mixing with room-temperature thermal activators. In this work we have introduced in the formulation a photochemical initiator, which permits better control of th e polymerization reaction, through control of the incident light intensity. The gas transport properties of the membranes thus prepared, which will be called [L.sub.CSPE] hereafter, have been measured and the results compared to those previously obtained with [L.sub.PU].
The transport properties have been tested using four gases supplied by Praxair. These are methane of purity 5.0, nitrogen and oxygen of purity 6.0, and carbon dioxide of purity 4.8. Loctite 329 and phenyl-bis- (2,4,6-trimethylbenzoyl) phosphine oxide (Irg819) were generous gifts from Loctite and Ciba SC, respectively.
The composition of Loctite 329 was found to include 36% w/w of a chlorosulfonated polyethylene polymer (Scheme 1), which acts as a binder, and a mixture of different methacrylic monomers, including methyl methacrylate as a main constituent (52% w/w of the adhesive composition).
Sample Preparation and Photocrosslinking
Samples containing 1% w/w phenyl-bis-(2,4,6trimethylbenzoyl) phosphine oxide (Irg819) and Loctite 329 were prepared by mixing both components and stirring until a homogeneous solution was obtained. They were placed between two transparent polyethylene films (LDPE, 80 [mu]m thick) and pressed with 2 X [10.sup.3] kg/c[m.sup.2] during 1 min using different spacers, which permits us to obtain membranes of different and reproducible thickness. The membranes used in this work were about 70 [mu]m thick. The samples were placed inside an ATLAS SUNTEST XSL equipment, provided with Xenon lamps with a power output of 550W/[m.sup.2] and irradiated with polychromatic light for 6 min, which allowed the system to photopolymerize until limiting conversion was obtained. The residual monomer was removed by immersion of the membrane in chloroform, for 24 h and vacuum drying until a constant weight was reached. These amorphous membranes will be called [L.sub.CSPE] hereafter.
Dynamic Mechanical Analysis, DMA
An attempt to determine the transitions and relaxations of the material by differential scanning calorimetry has been made, but no clear variations of the specific heat could be detected. DMTA measurements have been performed with this same scope. The tensile dynamic mechanical spectra were obtained in a Polymer Laboratories Thermal Sciences Dynamic Mechanical Analyzer. The measurements were carried out in tension mode between -90[degrees]C and 100[degrees] using a heating rate of 1.5[degrees]C/min at four frequencies. 1, 3, 10 and 30 Hz. The probes were parallelepipeclic bars of dimensions 0.10 x 2.2 x 15.3 [mm.sup.3].
A lab-made permeator was used, consisting of a gas cell in the middle of which the polymer membrane was placed. This membrane separates the upstream and downstream chambers. For the simplified equations relating the gas flow through the membrane to the transport coefficients to be valid, the downstream pressure must be kept very low, negligible as compared to the upstream. This is accomplished by thorough evacuation of the downstream, prior to any measurement, by means of an Edwards turbomolecular pump. At the low-pressure side, an MKS Baratron type 627B absolute pressure transducer measures the pressure increase, while at the upstream a Gometrics pressure detector is used to control the gas pressure at which the experiment is performed. The Baratron 627B can be used in the pressure range 1 torr to [10.sup.-4] torr. The Baratron is connected via an MKS power supply/readout unit to the PC, which records the pressure increase at given time intervals. The whole setup is temperature controlled in the range 20[degrees]C to 80[degrees]C by means of a water bath.
Prior to any measurement, vacuum is maintained overnight in order to remove any residual solvent from the membrane and to attain a downstream pressure that is as low as possible. Before any permeation experiment is performed, a measurement of the pressure increase due to imperfect vacuum isolation of the downstream chamber is recorded. This blank experiment is then subtracted from the permeation experiment performed immediately after in order to calculate the gas transport coefficients from the corrected pressure curves. In that way the pressure increase is related solely to the gas diffusing across the membrane. Pressure data are recorded every 5 s.
As there is a certain scatter of the experimental data, a large number of experiments were performed, and permeation measurements were done roughly every two degrees. Calculation of the transport coefficients was performed from the straight line characteristic of the steady state of gas diffusion through the membranes.
RESULTS AND DISCUSSION
Figure 1 shows the elastic modulus of [L.sub.CSPE] between -90[degrees]C and 100[degrees]C. A strong relaxation appears at roughly -20[degrees]C, corresponding to the soft chlorosulfonated polyethylene segments. After the relaxation, there is a continuous decrease of the elastic modulus, which makes it difficult to measure transport properties over 45[degrees]C, but no other relaxations up to 100[degrees]C can be seen.
P, D and S in the Temperature Range 20[degrees]C-45[degrees]C
The values of the permeability, diffusion and solubility (S=P/D) coefficients at 25[degrees]C are collected in Table 1. Figures 2 and 3 show the temperature evolution of the permeability and diffusivity coefficients in Arrhenius coordinates. At this moment, equipment shortcomings make it impossible to measure under 20[degrees]C, and owing to the material's mechanical properties, no measurements over 50[degrees]C were performed. Experimental measurements were thus limited to a small range, from 20[degrees]C to 45[degrees]C. In that range, Arrhenius behavior is followed by both permeation and diffusion coefficients.
As shown in Fig. 3, the activation energy of permeation for carbon dioxide is much smaller than for the other three gases because of the low value of the heat of solution [DELTA][H.sub.s]. The heat of solution can be either positive or negative, depending on the gas condensability and its ability to interact with the polymer. Among the gases tested, carbon dioxide has the lowext [DELTA][H.sub.s], and thus, as [E.sub.p] = [E.sub.d] + [DELTA][H.sub.s], the activation energy for this gas is the lowest. On the other hand, the solubility of [O.sub.2] and C[H.sub.4] varies little in that temperature range, whereas that of C[O.sub.2] diminishes clearly. This last behavior is characteristic of gases that interact with the polymer matrix, those bearing negative values of the heat of solution.
While oxygen, carbon dioxide, and methane measurements were quite straightforward in that little experimental dispersion of the data was found, nitrogen data are much more scattered and a great number of measurements were necessary to obtain reliable variations of the transport coefficients with temperature. The reason for this peculiar behavior of nitrogen is still unclear.
Table 2 shows the results on permselectivity at 25[degrees]C and 1 atm. The experimental data are obtained by simply calculating the ratio of experimental permeabilities ([O.sub.2]/[N.sub.2] and [CO.sub.2]/[CH.sub.4]) measured at the same temperature. On the other, hand it is possible to define an experimental permselectivity equation [alpha]([O.sub.2]/[N.sub.2]) and [alpha]([CO.sub.2]/[CH.sub.4]), by dividing the permeation Arrhenius equations derived from the representations in Fig. 3:
[alpha](A/B) = [P.sub.0](A)/[P.sub.0](B) [e.sup.[-([E.sub.A] - [E.sub.B])/RT]] (1)
As in the case of [L.sub.PU], the negative temperature coefficients in the Arrhenius equations and the activation energy balance lead to a diminishing of both [alpha]([O.sub.2]/[N.sub.2]) [alpha]([CO.sub.2]/[CH.sub.4]) with temperature.
Minimum and maximum values of solubility selectivity reported in the literature (30) for the gas pair [O.sub.2]/[N.sub.2] are 0.1 and 3.7 respectively, while for diffusivity selectivity the values range from 1.1 to 8.6. [L.sub.CSPE] has a solubility selectivity at 25[degrees]C of 1.9 and a diffusivity selectivity of 2.3, i.e., solubility selectivity is slightly over the average but the diffusivity selectivity of this polymer is not very high, which is not surprising if it is considered that the rigidity of the chain segments of the binder must be very low. On the other hand, permeability is small because of crosslinking, and surprisingly, [L.sub.CSPE] and [L.sub.PU] show very similar diffusivity, solubility and permeability selectivities. The greatest differences are found in the values of permeability, which are about three times lower for the former than for the latter. This is due to the smaller diffusion coefficients, i.e., to the lower fractional free volume existing in the [L.sub.CSPE] system as compare d to that of [L.sub.PU].
Transport Coefficients in Relation to the Gas Nature and to the Polymer Structure
Correlations between the gas transport properties and the structure of the polymeric membranes have been described since the first half of the past century (6, 9, 10, 30-32). For example, it has been stated that the activation energy of diffusion frequently varies linearly with the square of the kinetic diameter of the gas:
[E.sub.d] = c[d.sup.2.sub.c] - f (2)
where c is a measure of flexibility and f of the separation between chains. This correlations holds for the data presented in this work, as shown in Fig. 4, and a value of c = 655 [calmol.sup.-1] [[Angstrom].sup.2] is found, which is characteristic of flexible polymers. [L.sub.PU], on the other hand, had a c = 1270 [calmol.sup.-1] [[Angstrom].sup.2], characteristic of a more rigid medium (9).
These parameters are in agreement with the chemical composition of both crosslinked systems. The different nature of the binder in both adhesives introduces important structural differences, as shown for example by the different mechanodynamical behavior of both samples. In [L.sub.CSPE], which can be considered an elastomer, a strong relaxation related to the chlorosulfonated polyethylene appears at roughly -20[degrees]C. [L.sub.PU] has, on the other hand, a strong relaxation peaking at 70[degrees]C.
The higher density of [L.sub.CSPE] as compared to [L.sub.PU] is easy to explain, as the reactivity of the CSPE binder is greater than that of the aliphatic polyurethane. Both formulations have been polymerized up to limiting double bond conversion (around 65%-70%), which means that both possess the same degree of polymerization of the acrylic monomer. But given that both formulations include only monofunctional monomers, crosslinking reactions can only take place through a coupling reaction of a growing acrylic macroradical with a reactive site of the binder. As shown in Scheme 2 in the case of [L.sub.CSPE], three different reactions can lead to reactive sites in the binder structure: (a) direct photolysis of the C-C1 bond of the binder, given its high photochemical lability, (b) direct photolysis of the chlorosulfonic group and (c) direct hydrogen atom abstraction in the binder chain by the photoinitiator radicals.
In all cases, radical centers are efficiently formed in the binder chain, which can couple with the growing methacrylic radical, giving a certain degree of crosslinking for the same degree of polymerization. Full analysis of the photocuring reaction has been recently described (33) and the role played by the photoinitiator and the binder is discussed therein.
As mentioned earlier, the binder of [L.sub.CSPE] is a chlorosulfonated polyethylene similar to Hypalon 40. A comparison of this elastomer and our polymeric system, which is crosslinked using a mixture of acrylic and methacrylic monomers, appears interesting. In the last decade, various studies on the properties of special rubbers have appeared (34-37), and among them studies including Hypalon 40. As an example, Table 3 shows the permselectivities and the solubility and diffusivity selectivities for a series of commercial elastomers, together with the permeability of oxygen. Butyl rubber and Hypalon 40 are very similar in all three selectivities and differ clearly from the rest of elastomers, and in fact, the permselectivity data for our sample is very similar to that of Hypalon 40. While at 35[degrees]C most of the rubbers show permselectivites in the range 2-3, those of butyl rubber, Hypalon 40 and [L.sub.CSPE] are near 4. In this adhesive, up to 52% by weight is methyl methacxylate. However, comparison to P MMA gas transport coefficients (Table 3) shows that our sample's properties have very little to do with those of PMMA. It seems that even if a 52% by weight is PMMA, this does not strongly affect the diffusion properties. The methacrylic monomers initiate their polymerization at the radicals generated on the chlorosulfonated polyethylene chain. As they are monofunctional monomers, they grow linearly and terminate when they find other radicals, i.e., very probably on the chiorosulfonated polyethylene. Apparently, while the binder retains some of its properties in the final crosslinked network this is not the case for the growing methacrylic polymer. This is in agreement with the absence of a PMMA glass transition in the DMTA spectrum (Fig. 1), and the coincidence of the relaxation observed at -20[degrees]C with that of Hypalon 40 (37). In fact, the effect of the polymerized methacrylic monomers on the transport properties seems to be rather that of a filler. As a comparison, for example, Hydrin (one of the ela stomers in Table 3) has on its own a permeability to oxygen of 1.05 barrer, and with a filler It decreases to 0.31 (37).
Table 4 lists the activation energies for diffusion and permeation, and the solution enthalpy as obtained from the Van't Hoff relation. Figure 5 is a plot of these three magnitudes as a function of the critical temperature of each gas for both crosslinked systems. In Fig. 6 the heats of solution for [L.sub.CSPE] and [L.sub.PU] appear, together with the data for a polyimide from the literature (32). All three magnitudes follow similar trends in both adhesives, and the diffusion data of the two polymers obey the well-known linear free energy relation (10), as shown in Fig. 7.
The greatest difference in the results depicted in Fig. 5 is perhaps seen in the values of the heat of solution of the four gases in [L.sub.CSPE], which are lower than in [L.sub.PU], implying stronger interactions and lower heat of mixing between the gases and the polymer matrix. On the other hand, the smaller diffusion coefficients in [L.sub.CSPE] and the higher [E.sub.d], especially for oxygen and nitrogen, imply a denser structure in [L.sub.CSPE] than in [L.sub.PU]. In this connection, other authors (37) have pointed out that Hypalon 40 is, among the elastomers, one with a closely packed structure, which may account for the observations on the diffusion data (actual values and activation energies) found for [L.sub.CSPE].
The values of the transport coefficients in both adhesives indicate i) that specific chemical interactions are taking place in [L.sub.CSPE] that do not exist in [L.sub.PU], ii) that the free volume is higher in the latter and iii) that the chlorosulfonated polyethylene binder introduces flexibility and a higher density as compared with the aliphatic polyurethane. All these observations are in accordance with what we know about the chemical composition, microstructure and morphology of these materials.
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
[FIGURE 4 OMITTED]
[FIGURE 5 OMITTED]
[FIGURE 6 OMITTED]
[FIGURE 7 OMITTED]
Table 1 Solubility, Permeability and Diffusivity Coefficients at 25[degrees]C and 1 atm. [O.sub.2] [N.sub.2] [CH.sub.4] [CO.sub.2] P (barrer) 0.33 0.07 0.21 2.56 D*[10.sup.8] ([cm.sup.2]* 1.42 0.61 0.52 0.65 [s.sup.-1] S*[10.sup.4] [cm.sup.3] (STP)*[cm.sup.-3]*cm H[g.sup.-1] 24 13 40 394 1 barrer = [cm.sup.3] (STP) cm/[cm.sup.2] s cmHg X [10.sup.-10] Table 2 Diffusivity, Solubility and Permeability Selectivity of [L.sub.CSPE]. Selectivity [O.sub.2]/ [CO.sub.2]/ [N.sub.2]/ [N.sub.2] [CH.sub.4] [CH.sub.4] [P.sub.1]/[P.sub.2] 4, 6 12.2 0.5 [D.sub.1]/[D.sub.2] 2.3 1.2 1.6 [S.sub.1]/[S.sub.2] 1.9 9.8 0.3 Table 3 Permselectivity, Solubility and Diffusivity of Commercial Elastomers (37) and Commercial PMMA (3) at 35[degrees]C. Polymer P([O.sub.2])/ P([O.sub.2]) D([O.sub.2])/ S([O.sub.2])/ P([N.sub.2]) (barrer) D([N.sub.2]) S([N.sub.2]) Butyl rubber 3.80 2.89 1.90 2.01 Hypalon 40 3.80 2.37 2.08 1.81 Hypalon 45 3.30 4.20 1.41 2.30 Hydrin 100 2.30 1.05 1.80 1.33 Neoprene 2.97 8.83 1.90 1.57 ENR 76 2,22 10.90 1.29 1.72 EPDM 2.54 21.14 1.60 1.56 Kraton G1652 2.7 26.60 1.63 1.67 PMMA 7.8 0.06 5.01 1.55 [L.sub.CSPE] 3.8 0.33 1.9 2.0 Table 4 Activation Energy and Pre-exponential Factors of the Arrhenius Temperature Dependence of the Permeability and Diffusivity Coefficients. [O.sub.2] [N.sub.2] [CH.sub.4] [E.sub.d] [kJmol.sup.-1] 46.6 55 52 [E.sub.p] [kJmol.sup.-1] 46 57.7 49 [DELTA][H.sub.s] [kJmol.sup.-1] -0.6 2.8 -6 [D.sub.0] [cm.sup.2][s.sup.-1] 2.6 37 6.8 [P.sub.0] barrer*[10.sup.-7] 5 147.2 8.6 [CO.sub.2] [E.sub.d] [kJmol.sup.-1] 44 [E.sub.p] [kJmol.sup.-1] 24.3 [DELTA][H.sub.s] [kJmol.sup.-1] -19.6 [D.sub.0] [cm.sup.2][s.sup.-1] 0.33 [P.sub.0] barrer*[10.sup.-7] 0.048
We acknowledge financial support from the Consejeria de Cultura de la Communidad de Madrid (07M/0069/1998) and from the EU (BE97-4472).
(1.) A. K. Fritzsche and R. S. Narayan. Chem. Econ. Eng. Rev., 19, 19-31 (1987).
(2.) L. M. Robeson, J. Membr. Sci., 62, 165-85 (1991).
(3.) R. E. Kesting and A. K. Fritzsche, Polymeric Gas Separation Membranes, Chapters 1 and 2, pp. 1-60, R. E. Kesting and A. K. Fritzsche, eds., John Wiley & Sons, Inc. (1993).
(4.) S. Kagawa, M. Kawakami, K. Nishi, and M. Iwamoto, Synthetic Polymeric Membranes, B. Sedlacek and T. Kahovic, eds., Walter de Gruyter (1987).
(5.) S. A. Stern and H. L. Firsch, Ann. Rev. Mater Sci., 11, 523-27 (1981).
(6.) V. Stannett, Diffusion in Polymers, Chapter 2, pp. 41-73 and references cited therein, J. Crank and G. S. Park, eds., Academic Press (1968).
(7.) A. Thrau, G. Kroll, and F. Faupel, J. Polym. Sci.: Polym. Phys., 37, 3344-58 (1999).
(8.) Y. Maeda and D. R. Paul, J. Polym. Sci.: Polym. Phys., 25, 1005-16 (1987).
(9.) B. D. Freeman, Macromolecules, 32, 375-80 (1999).
(10.) D. W. Van Krevelen. Properties of Polymers: Their Correlation With Chemical Structure; Their Numerical Estimation and Prediction From Additive Group Contributions. 3rd Ed., pp. 535-82, Elsevier (1990).
(11.) A. Shimazu, T. Miyazakl, T. Matsushita. M. Maeda, and K. Ikeda, J. Polym. Sci., Part B: Polym. Phys., 37, 2941-49 (1999).
(12.) A. B. Fuertes, D. M. Nevskaia, and T. A. Centeno, Microporous Mesoporous Mater., 33, 115-25 (1999).
(13.) F. Wang, J. Li, X. Tianlu, and J. Xu. Polymer, 40, 795-99 (1999).
(14.) X. G. Jian, Y. Dai, L. Zeng, and R. X. Xu, J. Appl. Polym. Sci., 71, 2385-90 (1999).
(15.) R. E. Kesting and A. K. Fritzsche, Polymeric Gas Separation Membranes, Chapter 3, pp. 108-85 and references cited therein, R. E. Kesting and A. K. Fritzsche, eds., John Wiley & Sons, Inc. (1993).
(16.) C. Zimmerman and W. J. Koros, J. Polym. Sci, Part B: Polym. Phys., 37, 1235-49 (1999).
(17.) C. Zimmerman and W. J. Koros, Polymer, 40, 5655-64 (1999).
(18.) C. Zimmerman and W. J. Koros, J. Polym. Sci. Part B: Polym. Phys., 37, 1251-65 (1999).
(19.) R. Mahajan, C. Zimmerman, and W. J. Koros, ACS Symp. Ser., 733, 277-86 (1999).
(20.) R. Tamaki, Y. Chujo, K. Kuraoka, and T. Yazawa, J. Mater. Chem., 9, 1741-46 (1999).
(21.) M. L. Sforca, I. V. P. Yoshida, and S. P. Nunes, J. Membr. Sci., 159, 197-207 (1999).
(22.) M. H. Kim, J. H. Kim, C. K. Kim, Y. S. Kang, H. C. Park, and J. O. Won, J. Polym. Sci., Part B: Polym. Phys., 37(21), 2950-59 (1999).
(23.) T. Miyata, S. Obata, and Uragami, Macromolecules, 32, 8465-75 (1999).
(24.) H. Matsuyama, A. Terada, T. Nakagawara, Y. Kitamura, and M. Teramoto. J. Membr. Sci., 163(2), 221-27 (1999).
(25.) Y. Hirayama, Y. Ease, N. Tanihara, Y. Sumiyama, Y. Kusuki, and K. Haraya, J. Membr. Sci., 160, 87-99 (1999).
(26.) B. Bayer, B. Schoberi, K. Nagapudi, M. E. Rezac, and H. W. Beckham, ACS Symp. Ser., 733, 244-55 (1999).
(27.) C. L. Bennett and R. E. Richards, Brit. U.K. Pat Appl. GB 2244997 A1 (1991).
(28.) P. K. T. Oldring, ed., Chemistry and Technology of UV &EB Formulation for Coatings, Inks and Paints, SITA Technology, London (1999).
(29.) P. Tiemblo, J. Guzman, E. Riande, E. F. Salvador, and C. Pelnado, J. Polym. Sci. Part B: Polym. Phys., 39(7), 786-95 (2001).
(30.) A. Y. Alentiev and Y. P. Yampolskii, J. Membr. Sci., 165, 201-16 (2000).
(31.) J. E. Shelby, Handbook of Gas Diffusion in Solids and Melts, Chapter 6. pp. 123-31, ASM international (1996).
(32.) S. Pauly, Polymer Handbook, Permeability and Diffusion Data, Fourth Edition, pp. VI/543-VI/569, J. Brandrup, E. H. Immergut, and A. E. J. Grulke, eds., John Wiley and Sons Inc.
(33.) C. Peinado, A Fernandez-Arizpe, J. L. Mateo, F. Catalina, and P. Bosch, Macromol. Chem. Phys.. in Press (2001).
(34.) I. Pinnau, K. A. Lockandwala, P. Nguyen, L. G. Toy, and M. L. Jacobs, U.S. Patent 5,538,535 A 23 (1996).
(35.) R. S. Khinnavar and T. M. Aminabhavi, Polym-Plast. Technol. Eng., 31, 571-88 (1992).
(36.) T. M. Aminabhavi and R. S. Khinnavar, Polymer, 34, 4280-86 (1993).
(37.) M. W. Fitch, W. J. Koros, R. L. Nolen, and J. R. Carnes, J. Appl. Polym. Sci., 47, 1033-46 (1993).
P. TIEMBLO (*)
(*) Corresponding author. E-mail: firstname.lastname@example.org.