J O U R N A L O F C H E M I S T R Y Materials Thermal stability of styrene grafted and sulfonated proton conducting membranes based on poly(vinylidene fluoride) Sami Hietala,a Mihkel Koel,b Eivind Skou,c Matti Elomaa,a and Franciska Sundholm*a aL aboratory of Polymer Chemistry, University of Helsinki, PB 55, FIN-00014 Helsinki, Finland bInstitute of Chemistry, Estonian Academy of Science, Akadeemia tee 15, EE-0026 T allinn, Estonia cDepartment of Chemistry, University of Odense, DK-2530 Odense M, Denmark The thermal stability of styrene grafted and sulfonated poly(vinylidene fluoride), PVDF-g-PSSA, proton conducting membranes has been studied using thermal gravimetric analysis in combination with mass spectrometry and thermochromatography.The matrix polymer, PVDF, and the non-sulfonated counterpart, PVDF-g-PS, were studied as reference materials. It was found that the degradation of the PVDF-g-PS membrane proceeds in two steps starting at ca. 340 °C with the evolution of degradation products typical of polystyrene. The PVDF-g-PSSA membranes are stable to around 270 °C even in a strongly oxidising atmosphere. The degradation starts with the simultaneous evolution of water and sulfur dioxide.The polystyrene grafts start decomposing at 340 oC in the PVDF-g-PSSA membranes. Thus the membranes are suitable for tests in electrochemical applications at elevated temperatures. Polymeric separator materials for use in electrochemical cells in the temperature range 20–650 °C in an inert gas atmosphere, and fuel cells have to meet a combination of conditions: high and in oxygen atmosphere of the thermal degradation of ion conductivity, excellent electrochemical and chemical long PVDF based membranes. In addition the influence of crosslinkterm compatibility with the reducing and oxidative reagents ing the styrene grafts with two diVerent crosslinkers, diviat the electrocatalysts, reasonable mechanical stability includ- nylbenzene, DVB, and bis(vinylphenyl)ethane, BVPE, on the ing a defined swelling behaviour in the presence of water are thermal properties of the membranes was studied. The evolved among the most important requirements.1 Thermal stability is gaseous products were analysed by mass spectrometry and of crucial importance for membrane materials.Polyanskii and with thermochromatography to correlate mass losses at diVer- Tulupov2 have published a detailed review on the thermal ent temperatures with the formation of low molar mass degraproperties of polymer electrolytes.dation products related to the structure of the membranes. In the development of new polymer electrolyte materials Nafion 117 membranes were used as reference materials in the proton exchange membranes have been made by radiation- thermal degradation experiments.induced graft polymerisation.3–9 The process involves the polymerisation of a monomer in the presence of a preformed polymer film. The preparation of cation exchange membranes Experimental by the simultaneous radiation grafting of styrene monomer onto poly(tetrafluoroethylene-co-hexafluoropropylene) (FEP) Materials films has been reported by Rouilly et al.10 The grafted films The matrix material was PVDF film supplied by Goodfellow were sulfonated to introduce the ion exchange groups.A three (Cambridge) as melt processed 80 mm sheets. Pre-irradiated step pattern in the thermal degradation of these membranes films (electron beam under nitrogen gas, 100 kGy) were grafted was attributed to dehydration, desulfonation and degradation in styrene (Fluka) solution and subsequently sulfonated with of the FEP backbone.11 A detailed study of the thermal chlorosulfonic acid (Merck) in a three step procedure which properties of FEP based membranes has been done with has been described in detail previously.13,14 Membranes with thermal gravimetric analysis11 and with a combination of degrees of grafting, d.o.g., of 18, 32, 48, 60, 73 and 100%, thermogravimetry, FT IR and mass spectrometry.12 respectively, (PVDF-g-PS membranes), were fully sulfonated The preparation of proton exchange membranes by pre- (PVDF-g-PSSA membranes).Part of the membranes were irradiation induced styrene grafting onto poly(vinylidene flucrosslinked in the grafting step with either DVB (Fluka, oride) (PVDF) films followed by sulfonation has been reported isomeric mixture in ethylvinylbenzene) or BVPE (isomeric previously.13,14 It was found that the grafting occurs mainly in mixture) as has been previously described.16 the amorphous regions of the PVDF in which polystyrene The original PVDF film and the irradiated film were meas- domains are formed.14 In the sulfonation sulfonic acid groups ured without further treatment.The PVDF-g-PS samples were are introduced in the polystyrene chains thus forming hydromeasured as received, after drying in vacuo at 80 °C for 20 h, philic domains in the hydrophobic matrix polymer. The crystaland after a second drying period in vacuo at 120 °C for 20 h. linity of the matrix polymer changes only slightly during the The PVDF-g-PSSA membranes were treated for 1 h in boiling grafting and sulfonation.14 Good ion conductivities were found water and dried in vacuo at 80 °C for 20 h before measurement.for membranes based on PVDF.15 High conductivity, although The Nafion 117 membrane was boiled for 0.5 h in 3% aqueous an important factor, is not suYcient to make the polymer H2O2, 0.5 h in water, 0.5 h in 1 M aqueous H2SO4, 0.5 h in electrolyte suitable for demanding applications, such as fuel water and finally dried in vacuo at 80 °C for 20 h before cells.The clear identification of the degradation products measurement. The samples were stored in plastic bags in formed at various temperatures oVers the possibility of underambient conditions. 5–10 mg pieces of membrane were cut for standing the mechanisms of the thermal degradation of the membranes. This paper contains the results of an investigation the measurements. J. Mater. Chem., 1998, 8(5), 1127–1132 1127Measurements The thermal analyses were done with a Setaram 92-12 thermobalance connected to a Varian CH 7A mass spectrometer through a flow meter and a heated tube, approximately 40 cm in length.The samples were typically 4–6 mg. The inlet to the mass spectrometer was through a membrane inlet using a silicone rubber membrane, SR 606 by Radiometer Copenhagen, thickness 25 mm. Excess gases were directed out through holes in the inlet. HO3S F F F F F F F SO3H m n y x Tentative structure of the styrene grafted and sulfonated poly(vinylid- The samples were weighed on the thermobalance without ene fluoride), PVDF-g-PSSA any gas flow.Before the measurement the furnace was purged with a nitrogen flow of 150 ml min-1 (high flow) for 10 min. The gas flow was then changed to 25 ml min-1 (low flow) for The supermolecular structure of the system forms a very 1 min to stabilise the balance, after which the balance was complex system of crystalline and amorphous domains of tared.For measurements in an O2–N2 atmosphere this stabilis- PVDF, in which the sulfonated polystyrene grafts form hydroation was made accordingly with a 50550 mixture flowing at philic domains within the amorphous parts of the hydrophobic 50 ml min-1. The samples were heated from 20–650 °C with a PVDF.14 It was found that the ion conductivity at 20 °C in heating rate of 10 K min-1 under a nitrogen flow of the PVDF-g-PSSA membranes is of the same order of magni- 25 ml min-1, and the thermograms were recorded.The samples tude as measured for Nafion 117. The ion conductivities were were kept at 650 °C for 10 min, then the residue was burned measured with impedance spectroscopy. The values varied by flowing the O2–N2 mixture through for 2 min.During with d.o.g. and were typically around 100 mS cm-1.15 The measurements in an O2–N2 atmosphere the heating procedure hydrogen and helium gas permeabilities through the mem- was the same, but the gas atmosphere was kept constant. The branes were measured with a mass spectrometric leak detector, results were corrected by subtracting a baseline measurement and gas permeabilities equal to or lower than those for Nafion with an empty crucible. 117 were found.20 The detailed structural analysis of the Mass spectra were recorded every 60 s. The scanning was PVDF-g-PSSA membranes is still in progress in our started 5 min before the gas purging and the actual measurelaboratory. ment to check the mass spectrometer stability and record the The thermal degradation of the matrix polymer PVDF, the intensity of atmospheric air for comparison.The mass specgrafted PVDF-g-PS membranes, and the sulfonated PVDF-g- trometer responded quickly to weight losses and changes in PSSA membranes were studied in a nitrogen atmosphere, and the gas composition. However, due to the large volume of the in a nitrogen–oxygen (15:1) atmosphere.oven and condensation of high molar mass products in the It was found that the PVDF is very stable to around 420 °C tubing the decay of some peaks was slow and occasionally in a nitrogen atmosphere, and to around 410 °C in the presence produced a constant background for several mass peaks. The of oxygen which is to be expected for a fluorinated polymer accuracy of the mass data is one mass unit for masses below backbone.The thermal stability is slightly lower than for fully m/z 100, and two units for masses over m/z 100. The accuracy fluorinated vinyl polymers like FEP in which the degradation is adequate for this kind of mass trace analysis. starts at around 450 °C in a helium atmosphere, according to The thermochromatographic (ThGC)17,18 analysis was done a report by Gupta et al.12 In another report by the same group with a gas chromatograph (Carlo Erba 4200) equipped with a the ungrafted FEP was shown to be stable up to 490 °C in a pyrolysis oven, a sampling valve, a capillary column [NSWnitrogen atmosphere.21 The degradation products of FEP were Plot (HNU Nordion), inner diameter 0.53 mm and length determined using mass spectrometry.The main products in 25 m] and a thermal conductivity detector (Model 430). The the decomposition correspond to the splitting of fragments of pyrolysis oven was a quartz tube with an inner diameter of the main monomer, C2F4, and the comonomer, C3F6.12 In 4 mm and length of 250 mm having a centrally located 25 mm PVDF, the degradation caused formation of fragments corre- long quartz sample vessel.The sample size was around 5 mg. sponding to dimers, monomers and oligomers of vinylidene The samples were conditioned to constant mass in a constant fluoride (m/z=129, 130, 63, 64, 65 and larger fragments). Some relative humidity of 75% over a saturated aqueous solution of typical mass traces from degradation products of PVDF are sodium nitrate at room temperature before the analysis.The shown in Fig. 1. The irradiated PVDF film (radiation dose heater for the pyrolysis consisted of a copper block with 100 kGy) showed the same fragmentation pattern as the resistor heating elements surrounding the quartz tube. untreated film, but the onset of the degradation was about Temperature calibration was achieved by placing a thermo- 5 °C lower. When heated in an inert atmosphere only a charred couple in the sample vessel in place of a sample.The heating residue of almost 30% was left of the PVDF at 650 oC. In the ramp could be repeated with a measured precision of ±1 °C. presence of oxygen the whole sample was burned to gaseous The sampling valve consisted of a Deans’ type valve19 inside products at around 530 °C, see Fig. 2. the column oven run by a three way solenoid valve outside Thermograms measured in an inert atmosphere of an the oven. A microcomputer was used to control both the ungrafted and several grafted PVDF-g-PS films are shown in heating of the pyrolysis oven and the timed sampling of Fig. 3. The grafted films show mainly a two step degradation. the evolved gases in the pyrolyser tube head space. The sample The first degradation occurs at 390 °C in the non-crosslinked was heated from 70–400 °C with a heating rate of 5 K min-1.grafted films. In the crosslinked films the first degradation step Helium gas (99.99%) was used to purge the pyrolyser at a rate occurs at 10–15 °C lower temperatures than in the non- of 10 ml min-1 and through the column at a rate of 4 ml min-1.crosslinked films. Thermograms of samples crosslinked with An injection period of 1 s was repeated at 110 s intervals. The 5% BVPE and 5% DVB, respectively, are shown in Fig. 4. design of the device permitted eYcient, rapid separation of low The degradation in the PVDF-g-PS film crosslinked with DVB boiling point components of the pyrolysate at a constant starts at the lowest temperature.The second degradation step column-oven temperaure of 70 °C. occurs in all these samples at around 420 °C and corresponds to the degradation of PVDF. This is clearly illustrated in Fig. 5 Results and Discussion which shows the mass trace for typical degradation products from polystyrene (m/z 103, 104) and from PVDF (m/z 130) as The tentative chemical structure of the PVDF-g-PSSA membranes is shown: a function of temperature, recorded during the thermal analysis. 1128 J. Mater. Chem., 1998, 8(5), 1127–1132Fig. 4 The thermal degradation of non-crosslinked PVDF-g-PS (—), Fig. 1 The thermal degradation of poly(vinylidene fluoride), PVDF. Mass trace of the evolution of fragments with m/z 63–65 (%), 80 (o), and of PVDF-g-PS crosslinked with 5% BVPE (- - -) and with 5% DVB (....) in a nitrogen atmosphere.The d.o.g. is 32%, 30% and 31%, 115 (+) and 130 (×). respectively. Fig. 2 The thermal degradation of PVDF in a nitrogen atmosphere Fig. 5 The thermal degradation of PVDF-g-PS crosslinked with 5% (—) and in an O2–N2 atmosphere (- - -), and of styrene grafted PVDF, DVB. Mass trace of the evolution of fragments with m/z 104–105 (%) PVDF-g-PS in a nitrogen atmosphere (....) and in an O2–N2 atmosphere from polystyrene grafts and 130 (o) from the PVDF matrix.(-.-.-). Degree of grafting, d.o.g., 48%. Thus the conclusion is drawn that the matrix polymer remains unchanged in the grafted samples and the polystyrene grafts do not alter the inherent decomposition of the PVDF. The two components, PVDF and PS, undergo fragmentation separately.Once the decomposition of the polystyrene grafts is completed, it leaves behind the more stable PVDF backbone which decomposes above 430 °C. This is in accordance with the view that the polystyrene grafts are incompatible with the PVDF matrix and form phase separated microdomains in the grafted polymer,13 and behave as a distinct two phase system on thermal degradation.Similar observations have been made by Momose et al.22 for the decomposition of a,b,b-trifluoroethylenesulfonyl fluoride grafted onto polyethylene film. Furthermore, Gupta et al.12,21 conclude that the introduction of polystyrene onto FEP films introduces a two step degradation behaviour in the thermograms of copolymer films where both the polystyrene and FEP components undergo degradation in separate steps.The decreased thermal stability of the crosslinked PVDF-g- Fig. 3 The thermal degradation of PVDF (—) and PVDF-g-PS mem- PS membranes with respect to the non-crosslinked is somewhat branes with d.o.g. 18% (- - -), 32% (....), 48% (-.-.), 68% (-..-) and surprising, since it has been shown that crosslinks stabilise the 73% (-.-.) in a nitrogen atmosphere polymer structure.23,24 These studies were concerned with styrene grafted and sulfonated FEP membranes crosslinked with DVB or triallyl cyanurate.It was shown that the grafting of FEP was considerably reduced by the presence of the J. Mater. Chem., 1998, 8(5), 1127–1132 1129crosslinker, and that the rate and the final d.o.g. were decreased. 20% in PVDF in films with d.o.g.=75%, which further points to some synchronous degradation mechanism of PVDF and In the present case it was found that both crosslinkers, DVB and BVPE, increase the final d.o.g. dramatically.16 Thus the short residual polystyrene grafts in the films. If all the polystyrene were to decompose, the residue would be about 17% grafting reactions in PVDF and in FEP diVer in mechanism.Higher d.o.g. can be achieved in PVDF than in FEP under for the sample with d.o.g.=75% at 650 °C. The thermal degradation of the PVDF-g-PS films in the similar reaction conditions. One possible explanation is the diVerence in glass transition temperatures, Tg, which for FEP presence of oxygen is illustrated in Fig. 2. The mass loss curve shows the onset of the degradation of the PVDF part at is 55 °C21 and for PVDF is -40 °C.25 Thus the FEP is close to its glassy state during the grafting, whereas the grafting of slightly lower temperatures than in the ungrafted film.The degradation of the styrene grafts starts at around 270 °C, the the PVDF takes place in the rubbery state. The diVusion of the styrene and the crosslinkers into the irradiated matrix is mass loss in the interval 270–420 °C is of the same order of magnitude as for the PVDF-g-PS sample in a nitrogen atmos- influenced by the diVerence in mobility of the polymer matrices.The diVerence between the two crosslinkers, DVB and BVPE, phere in the interval 390–410 °C. The degradation of the sample in the presence of oxygen is complete at around 530 °C.on the other hand is explained by the very large diVerence in their reactivities compared with the reactivity of styrene.16,26 The sulfonation of the polystyrene grafts in PVDF-g-PS produces a strongly acidic polyelectrolyte membrane, PVDF- DVB has a much higher reactivity than styrene which results in the formation of highly crosslinked stiV areas close to the g-PSSA. The thermograms of the degradation of the PVDFg- PSSA membranes in an inert atmosphere are shown in grafting points. The product of the reactivity ratios r1/r2 of BVPE and styrene is close to 1, hence the reaction results in Fig. 7. The decomposition reaction diVers from the one in the PVDF-g-PS membranes since the residue at 650 °C increases a more homogeneously and randomly crosslinked membrane.16 The degradation pattern of the PVDF-g-PS films is slightly with increasing d.o.g.This is probably due to an increase in char formation of the polystyrene grafts in the presence of dependent on the preceding drying procedure. In samples dried at 120 oC the thermograms show only the two degradation sulfur dioxide (and other acidic fragmentation products originating from the sulfonic acid groups) and water during the steps of the polystyrene grafts and the matrix polymer, see Fig. 2 and 3. Samples which were measured as received, or thermal degradation. Similar eVects have been reported in the thermal degradation of polystyrene in the presence of sulfuric had been dried in vacuo at 80 °C, showed a small mass loss around 120 °C. The mass loss is about 1% at d.o.g.= 18%, acid or Lewis acids.27,28 The first stage in the thermal degradation of the PVDF-g- and it increases to about 4% at d.o.g.=73%. Mass spectra recorded of the evolving gases at 120 °C do not, however, PSSA membranes is a mass decrease of 1–10% depending on d.o.g. between 100 and 180 °C. This mass decrease is due to indicate evaporation of residual solvents or reagents from the membranes.The fragmentation trace points to chain end loss of bound water in the membranes. The introduction of the hydrophilic sulfonic acid group in the membrane makes it fragmentation in the PVDF matrix; the peaks in the mass spectrum can be attributed to fragments of type CHxCFy. hygroscopic. Part of the water becomes hydrogen bonded to the sulfonic acid groups, and remains in the membrane even There is no indication of fragmentation of the polystyrene grafts at this temperature.Since it was found that the PVDF after drying in vacuo at 80 °C. Similar behaviour has been observed in the FEP based proton exchange membranes12 and and the irradiated PVDF are stable at 120 °C the conclusion is drawn that the styrene grafting reaction has caused some in commercial membranes.29 Massive degradation of the membranes starts at 220 °C.decomposition of the PVDF matrix. The mass losses from PVDF-g-PS films with various d.o.g. Fig. 8 shows the mass trace of the thermal degradation of a PVDF-g-PSSA membrane with d.o.g. 48%. The mass trace at 420 °C are seen in Fig. 6. The polystyrene content of the PVDF-g-PS films was calculated from the d.o.g.{d.o.g. %= shows that there is a simultaneous increase in water formation and onset of formation of sulfur dioxide at 220 °C. The [(W-W0)/W0]×100, polystyrene content=[d.o.g./(d.o.g. +100)]×100% where W is the mass of the grafted membrane evolution of sulfur dioxide is clearly seen from the mass spectra. Evidence of the formation of sulfur oxide and sulfur trioxide and W0 is the mass of the ungrafted membrane, respectively}.It is seen that the mass loss at the temperature of the onset of decomposition of PVDF, 420 °C, increases with increasing d.o.g. However, in the grafted films there is a polystyrene residue of 4–10% left at this temperature. This implies that the mechanism of degradation of the polystyrene changes as the decomposition of the chains approaches the graft points.The polystyrene decomposition becomes linked to the decomposition of the PVDF when small amounts of the grafts are left. The residue at 650 °C decreases with increasing d.o.g. on heating in a nitrogen atmosphere, from around 30 to only Fig. 7 The thermal degradation of PVDF-g-PSSA membranes with d.o.g. 18% (- -), 32% (....), 48% (.-.-) and 73% (-.-.-.) in a nitrogen atmosphere.The thermogram of pure PVDF (—) is included as a Fig. 6 Mass loss at 420 °C from PVDF-g-PS membranes as a function reference, as is the degradation of a PVDF-g-PSSA membrane with d.o.g. 70% and crosslinked with 5% DVB (6) and the degradation of the mass% styrene in the membrane, non-crosslinked (&), crosslinked with 5% BVPE ($) and crosslinked with 5% DVB (+), of a PVDF-g-PSSA membrane, d.o.g. 60%, in an O2–N2 atmosphere (+++). respectively 1130 J. Mater. Chem., 1998, 8(5), 1127–1132Fig. 9 The mass loss from PVDF-g-PSSA membranes between 220 °C and 320 °C ($) and the ion exchange capacity Q (&) as a function Fig. 8 The thermal degradation of PVDF-g-PSSA membrane with of d.o.g. d.o.g. 48%. Mass traces of evolving fragments m/z 18 (water %), 64 (SO2 o), 104 (styrene +) and 117 (×) are shown.was also seen in the mass trace. Desulfonation is at its maximum around 320 °C. Depolymerisation of the polystyrene grafts occurs after this in the interval 390–410 °C, and degradation of the PVDF matrix sets in at 430 °C as in the PVDFg- PS membranes. The desulfonation temperature, the decomposition temperatures of the polystyrene grafts and of the PVDF backbone are unaVected by the degree of grafting. The crosslinkers DVB and BVPE shift the degradation to lower temperatures, DVB more than BVPE.The degradation trace for a PVDF-g-PSSA membrane crosslinked with 5% DVB is included in Fig. 7. The onset of the formation of sulfur dioxide is as low as 200 °C. The degradation pattern of the PVDF-g-PSSA membranes resembles the degradation trace obtained under similar conditions from styrene grafted and sulfonated FEP membranes.11,12 The main diVerence is that the degradation of the PVDF-g-PSSA membranes starts at Fig. 10 Thermochromatogram showing the evolution of water (reten- lower temperatures. Since the chemical composition of the tion time 60–70 s) and sulfur dioxide (retention time 130–140 s) from FEP based and the PVDF based membranes is very similar a PVDF-g-PSSA membrane as a function of temperature.D.o.g. 48%. this marked diVerence could be due to diVerent mechanisms of formation due to the very diVerent Tg values of the matrix materials leading to diVerences in morphologies in the prod- PVDF-g-PSSA membranes; an example with d.o.g. 48% is seen in Fig. 10, with a maximum around 135 s at 300 °C. The ucts. Further studies are in progress. The mass loss due to evolution of sulfur dioxide from the evaporation of the strongly bound water in the range 70–180 °C as a function of d.o.g. is illustrated in Fig. 11. Formation of PVDF-g-PSSA membranes correlates with measured values of the ion exchange capacity, Q, giving further evidence for the water and sulfur dioxide in the range 180–400 °C as a function of d.o.g.is included in the same figure. The conclusion is drawn loss of sulfonic acid groups from the membrane at elevated temperatures. The measurements of Q have been reported that water evaporates in two distinct fractions, and the water formation in the higher temperature range is accompanied by previously.30 The mass loss in the PVDF-g-PSSA membranes in the temperature interval 220–320 °C as a function of d.o.g. the simultaneous formation of sulfur dioxide.Thus the results from the thermal analysis with gas leak detection and from is shown in Fig. 9. The formation of water in the degradation of the PVDF-g- the thermochromatography are in excellent agreement. The degradation of the PVDF-g-PSSA membranes in the PSSA films is not very clearly seen from the mass trace because of the high water background. This is because some atmos- presence of oxygen is very similar to the degradation in nitrogen atmosphere.In the oxidising environment the degra- pheric water is present in the ionisation chamber, and because of the hydrophobic nature of the silicone membrane in the dation of the PVDF component starts at a lower temperature than in the nitrogen atmosphere, see Fig. 7. The membrane is spectrometer inlet; water does not pass the membrane at rates corresponding to the evolution of water in the degradation. completely combusted at 480 °C in the oxidising environment. Therefore the thermal degradation of the PVDF-g-PSSA membranes was analysed with thermochromatography.The two Conclusion dimensional picture of gas evolution as a function of temperature is shown in Fig. 10. The thermogram shows a secondary The matrix polymer, PVDF, is thermally stable to 420 °C in an inert atmosphere, and to 400 °C in the presence of oxygen. background peak seen as a front at retention time 43 s. This peak is due to secondary degradation products formed from In the PVDF-g-PS films the degradation of the polystyrene grafts at lower temperatures than the PVDF is evident. degradation products in previous heating cycles.The evaporation of strongly bound non-freezing water is clearly seen after Crosslinking of the grafts decreases the thermal stability of the grafted polymers. The PVDF-g-PSSA membranes are stable 63 s at around 100 °C, as is water formed as a degradation product from the sulfonate groups at around 63 s above 200 °C.to 370 °C in an inert atmosphere, and to 270 °C in a highly oxidising atmosphere. The degradation starts with the splitting The formation of sulfur dioxide is seen in thermograms of a J. Mater. Chem., 1998, 8(5), 1127–1132 11313 A. Bozzi and A. Chapiro, Eur.Polym. J., 1987, 23, 255. 4 A. Chapiro, Radiat. Phys. Chem., 1979, 9, 55. 5 A. Bozzi and A. Chapiro, Radiat. Phys. Chem., 1988, 32, 193. 6 B. D. Gupta and A. Chapiro, Eur. Polym. J., 1989, 11, 1137. 7 B. D. Gupta and A. Chapiro, Eur. Polym. J., 1989, 11, 1145. 8 E. A. Hegazy, I. Ishigaki, A. M. Dessouki, A. Rabie and J. Okamoto, J. Appl. Polym. Sci., 1982, 27, 535. 9 A.Chapiro and A. M. Jedrychowska-Bonamour, Eur. Polym. J., 1984, 20, 1079. 10 M. V. Rouilly, R. Ko� tz, O. Haas, G. G. Scherer and A. Chapiro, J.Membr. Sci., 1993, 81, 89. 11 B. Gupta and G. G. Scherer, J. Appl. Polym. Sci., 1993, 50, 2129. 12 B. Gupta, J. G. Highfield and G. G. Scherer, J Appl. Polym. Sci., 1994, 51, 1659. 13 S. Holmberg, T. Lehtinen, J. Na�sman, D. Ostrovskii, M.Paronen, R. Serimaa, F. Sundholm, G. Sundholm, L. Torell and M. Torkkeli, J.Mater. Chem., 1996, 6, 1309. 14 S. Hietala, S. Holmberg, M. Karjalainen, J. Na�sman, M. Paronen, R. Serimaa, F. Sundholm and S. Vahvaselka�, J.Mater. Chem., 1997, 7, 721. 15 T. Lehtinen, F. Sundholm, G. Sundholm, P. Bjo�rnbom and Fig. 11 The evolution of water and sulfur dioxide as a function of M. Bursell, Electrochim.Acta, accepted. d.o.g. Water evolved in the range 100–180 °C from a PVDF-g-PSSA 16 S. Holmberg, J. H. Na�sman and F. Sundholm, Polym. Adv. membrane (o), and from a BVPE crosslinked (5 mol%) PVDF-g- T echnol., accepted. PSSA membrane (×). Water evolved in the range 180–400 °C from a 17 M. Kaljurand and M. Koel, Computerized Multiple Input PVDF-g-PSSA membrane (%), and from a BVPE crosslinked Chromatography, Ellis Horwood, Chichester 1989, p. 139. (5 mol%) PVDF-g-PSSA membrane (6). Sulfur dioxide evolved in 18 M. Elomaa, PhD Thesis, University of Helsinki, Helsinki 1991. the range 180–400 °C from a PVDF-g-PSSA membrane ($), and from 19 D. R. Deans, J. Chromatogr. 1984, 289, 43. a BVPE crosslinked (5 mol%) PVDF-g-PSSA membrane (s) . 20 S. Hietala, E.Skou and F. Sundholm, submitted to Polymer. 21 B. Gupta and G. G. Scherer, Angew. Makromol. Chem., 1993, 210, 151. of the sulfonic acid groups as water and mainly sulfur dioxide. 22 T. Momose, I. Ishigaki and J. Okamoto, J. Appl. Polym. Sci., 1988, The degradation of the polystyrene grafts start at 390 °C. 36, 669. Crosslinking of the grafts decreases the thermal stability of the 23 F. N. Bu� chi, B. Gupta, O. Haas and G. G. Scherer, J. Electrochem. membranes. Thus the PVDF-g-PSSA membranes can be Soc., 1995, 142, 3044. 24 F. N. Bu� chi, B. Gupta, O. Haas and G. G. Scherer, Electrochim. regarded as suitable for tests as polyelectrolyte membranes in Acta, 1995, 40, 345. applications at elevated temperatures up to 200 °C. 25 J. Brandrup and E. H. Immergut, Polymer Handbook, Wiley, New York, 3rd edn., 1989, p. VI-226 and VI-258. S.H. and M.E. are indebted to the Nordic Energy Research 26 R. Wiley and G. Mayberry, J. Polym. Sci. A, 1963, 1, 217. Programme (NEFP) for grants. F.S. acknowledges research 27 C. F. Cullis and M. M. Hirschler, T he Combustion of Organic funding from The Academy of Finland. The authors are well Polymers, Clarendon, Oxford, 1981, pp. 229–230. 28 X. Zhu, M. Elomaa, F. Sundholm and C. H. Lochmu� ller, aware that without the assistance with the syntheses by Svante Macromol. Chem. Phys. 1997, 198, 3137. Holmberg and Jan Na�sman, A ° bo Akademi University, this 29 T. A. Zawodzinski, Jr., M. Neeman, L. O. Sillerud and study would not have been possible. S. Gottesfeld, J. Phys. Chem., 1991, 95, 6040. 30 S. Hietala, S. Holmberg, J. Na�sman, D. Ostrovskii, M. Paronen, R. Serimaa, F. Sundholm, L. Torell and M Torkkeli, Appl. References Macromol. Chem. Phys., 1997, 253 151. 1 Fuel Cell Handbook, ed. A. J. Appleby and R. L. Foulkes, Van Nostrand, New York, 1989, G. G. Scherer, Ber. Bunsenges. Phys. Paper 7/08288F; Received 18th November, 1997 Chem., 1990, 94, 1008. 2 N. G. Polyanskii and P. E. Tulupov, Russ. Chem. Rev., 1971, 40, 1030. 1132 J. Mater. Chem., 1998,