A novel sulfonated polyimide having pendant sulfophenoxypropoxy groups was synthesized, and its electrolyte properties were investigated for fuel-cell applications. A diamine monomer was synthesized from 3,3′-dihydroxybenzidine in four steps, and its polymerization was performed with 1,4,5,8-naphthalenetetracarboxylic dianhydride in the presence of triethylamine and benzoic acid in m-cresol. A flexible, ductile, and self-standing membrane was obtained via casting from an m-cresol solution. Good thermal, hydrolytic, and oxidative stability was confirmed. The ionomer membrane showed very high proton conductivity of 1.0 S cm−1 at 120 °C and 100% relative humidity, which is higher than that of a Nafion membrane.
Proton-conductive polymers are some of the key materials in polymer electrolyte fuel cells and direct methanol fuel cells. Because of their chemical and physical stability and high proton conductivity, perfluorosulfonic acid polymers such as Nafion are the state-of-art materials and have been most studied. There has been a great demand for alternative nonfluorinated polymeric materials that are more conductive, operable at high temperatures (>120 °C), environmentally benign, and inexpensive. One of the challenges is the acid functionalization of aromatic hydrocarbon polymers.1–5 Sulfonated polyimide copolymers have been proposed by several research groups.6–17 The advantage of polyimide ionomers is that a high-molecular-weight polymer with good film-forming capability and high mechanical strength can be easily synthesized by the polycondensation of sulfonated diamine monomers. The hydrophilic/hydrophobic balance or ion-exchange capacity (IEC) can be controlled by changes in the copolymer composition. Some of them are claimed to display higher proton conductivity than the perfluorinated materials. We have found that a sulfonated polyimide containing fluorenyl groups displays quite high proton conductivity, 1.67 S cm−1 at 120 °C and 100% relative humidity (RH), because of the high water-holding capability.10, 11 However, hydrolytic stability is an issue to be addressed. Okamoto and coworkers12–14 reported that introducing propoxy groups between the main chain and sulfonic acid groups improves the hydrolytic stability. We have also investigated the effect of aliphatic groups in both the main and side chains to achieve excellent hydrolytic and oxidative stability.18–20 However, because of the synthetic difficulties, the effect of the side-chain length has not been well examined.21 In this article, we report the synthesis and properties of a novel polyimide electrolyte having longer pendant sulfophenoxypropoxy groups. Its properties are compared with those of previous side-chain-sulfonated polyimide and perfluorinated Nafion membranes.
3,3′-Dihydroxybenzidine (98%) was purchased from TCI Co., Inc. Glacial acetic acid (99.7%), acetic anhydride (97.0%), 1,3-dibromopropane (98.0%), potassium carbonate (99.5%), anhydrous acetonitrile (99.5%), hydrochloric acid (35–37%), benzoic acid (99.5%), sodium hydroxide (97%), potassium iodide (99.5%), sodium ethoxide (95%), 1-butanol (99.0%), ethanol (99.5%), toluene (99.5%), dimethyl sulfoxide (DMSO; 99.0%), and m-cresol (98.0%) were purchased from Kanto Chemical Co., Inc. 4-Hydroxybenzene sulfonic acid sodium salt dihydrate (98%), 1,4,5,8-naphthalenetetracarboxylic dianhydride (TCND), and triethylamine (99.5%) were purchased from Aldrich Co., Inc. m-Cresol was dried over 4-Å molecular sieves before use. Other chemicals were used as received.
Synthesis of 4,4′-Diacetamido-3,3′-bis(3-bromopropoxy)biphenyl (2)
4,4′-Diacetamido-3,3′-dihydroxybiphenyl (1) was prepared by the acetylation of 3,3′-dihydroxybenzidine as described in the literature.22 To a suspension of potassium iodide (catalytic amount) and 1,3-dibromopropane (52.0 g, 256 mmol) in 250 mL of anhydrous acetonitrile, 1 (0.6 g × 8, 16 mmol) and potassium carbonate (1.40 g, 10 mmol) was added in eight portions over 1 h with stirring at 80 °C. The reaction was continued for another 8 h. The mixture was filtered. The solid was washed with hot acetonitrile. The combined filtrate was cooled in a refrigerator to give a white precipitate. The precipitate was washed with cold acetonitrile to give pure 2 in a 45% yield.
4,4′-Di(acetamido)-3,3′-bis[3-(4-sulfophenoxy)propoxy]biphenyl Disodium Salt (3)
To a mixture of dehydrated 4-hydroxybenzensulfonic acid sodium salt (6.68 g, 28.8 mmol) and 40 mL of ethanol was slowly added sodium ethoxide (2.06 g, 30.3 mmol). After the addition of 30 mL of toluene, the mixture was evaporated to dryness. DMSO (60 mL) and 30 mL of toluene were added to the residue, and toluene was evaporated from the mixture. This drying process was performed to remove ethanol completely. To the mixture was added 2 (3.90 g, 7.2 mmol), and the resulting mixture was stirred at 80 °C for 1 day. After the reaction, a 25% NaCl aqueous solution was added to the mixture to give a yellow precipitate. The precipitate was collected by filtration and recrystallized from acetonitrile to give pure 3 in a 76% yield.
A mixture of 3 (4.20 g, 5.44 mmol), 50 mL of water, 50 mL of hydrochloric acid, and 70 mL of 1-butanol was stirred at reflux under an N2 atmosphere for 4 h and at room temperature for 1 night. The resulting precipitate was collected by filtration, washed with 80% aqueous ethanol, and dissolved in a mixture of 30 mL of a 5% sodium hydroxide aqueous solution and 30 mL of ethanol. After filtration, the filtrate was acidified with concentrated hydrochloric acid and cooled in an ice bath for 30 min to give a precipitate. The precipitate was collected by filtration and washed with water to obtain pure 4 in a 37% yield.
Monomer 4 (0.30 g, 0.62 mmol), triethylamine (0.22 mL, 1.6 mmol), TCND (0.17 g, 0.62 mmol), benzoic acid (0.23 g, 0.19 mmol), and 5 mL of m-cresol were placed in a sealed glass tube. The suspension was heated at 150 °C until a clear solution was obtained and then stirred at 175 °C for 15 h and at 195 °C for another 3 h. The polymer was recovered by precipitation from acetone and dried in vacuo.
An m-cresol solution of polymer 5 as the polymerization mixture was cast onto a flat glass plate and dried at 60 °C for 1 day. The crude membrane in its triethylammonium salt form was soaked in ethanol containing 1 N HNO3 for 12 h. This acidification procedure was repeated three times. The 5 membrane in the acid form was washed thoroughly with ethanol at 60 °C for 30 min three times and dried at 60 °C under reduced pressure.
Fourier transform infrared (FTIR) spectra were measured as KBr pellets on a Jasco FT/IR-500 FTIR spectrometer. 1H and 13C NMR spectra were recorded on a Bruker Avance 400S spectrometer with DMSO-d6 or chloroform-d (CDCl3) as a solvent and tetramethylsilane as an internal reference. Gel permeation chromatography (GPC) measurements were performed with dimethylformamide (DMF) containing 0.01 M LiBr at a flow rate of 1.0 mL/min and an injection volume of 20 μL with a polymer concentration of approximately 4 mg/mL. A Jasco 875 UV detector (set at 300 nm) and two Shodex KF-805 columns were used. The weight-average molecular weight and number-average molecular weight were calibrated with standard polystyrene samples. The thermal properties were studied by thermogravimetry/differential thermal analysis and mass spectrometry (TG/DTA–MS) with a Mac Science TG-DTA 2000 and an MS9600 quadrupole mass spectrometer. Approximately 10 mg of a sample was heated at a heating rate of 10 °C/min from 30 to 400 °C in an argon atmosphere.
Oxidative and Hydrolytic Stability
As an accelerated test, a small piece of a membrane sample with a thickness of 30 μm was soaked in Fenton's reagent (3% H2O2 containing 2 ppm FeSO4) at 80 °C for 1 h or in pressurized water at 140 °C for 24 h for oxidative or hydrolytic stability testing, respectively. The stability was evaluated by weight changes after exposure to Fenton's reagent or pressurized water.
Small-Angle X-Ray Scattering (SAXS) Measurements
SAXS measurements were collected with a MAC Science MXP3 X-ray diffractometer with Ni-filtered Cu Kα radiation and a scintillation counter detector. The beam was collimated by three slits 0.1, 0.04, and 0.05 mm wide. The sample-to-detector distance was 200 mm, and the scanning rate was 0.06 °/min. To obtain small-angle X-ray diffraction data for fully hydrated samples, the polymer films were sealed in a Mylar bag with water.
Proton Conductivity and Water Uptake
A four-point-probe conductivity cell with two gold-wire outer current-carrying electrodes and two gold-wire inner potential-carrying electrodes was fabricated. Membrane samples were cut into strips that were 1.0 cm wide, 3.0 cm long, and 30 μm thick before being mounted in the cell. The cell was placed in a closed stainless steel chamber, in which the temperature and the humidity were controlled by flowing humidified nitrogen. Above 100 °C, the system was pressurized. Impedance measurements were taken with a Solartron SI1280 electrochemical impedance analyzer. The instrument was used in the galvanostatic mode with a current amplitude of 0.005 mA over a frequency range from 10 to 100,000 Hz. The temperature dependence of the conductivity was measured by the heating of the membrane samples from 40 to 120 °C. Impedance data were taken 2 h after the constant temperature was obtained.
The water uptake was simultaneously measured in a conductivity measurement chamber with a magnetic suspension balance. The membrane samples (50–70 mg) were set in a chamber and dried at 80 °C in vacuo for 3 h until a constant weight as a dry material was obtained. The membrane was then equilibrated with the gases at the given temperature and humidify for at least 1 h before the gravimetry was performed.
RESULTS AND DISCUSSION
The synthetic route of diamine monomer 4 having pendant sulfophenoxypropoxy groups is shown in Scheme 1. First, the amino groups of 3,3′-dihydroxybendizine were protected by acetylation. Protected compound 1 was reacted with an excess of dibromopropane under basic conditions to produce compound 2 having bromide groups at the end of the pendant groups. Using a large excess (16 times) of dibromopropane could efficiently prevent intermolecular etherification reactions. Compound 2 was then reacted with 4-hydroxybenzene sulfonic acid sodium salt to give compound 3. Deacetylation of the amines and acidification of the sulfonic acid groups were carried out simultaneously in HCl/n-BuOH to give pure compound 4 as a pale yellow powder. Compound 4 was obtained in a 12% total yield. Each compound (1–4) was well characterized with 1H NMR and FTIR spectra, as described in the Experimental section.
The polymerization reaction of monomer 4 with TCND was carried out in an m-cresol solution to give the title polymer 5 having pendant acidic groups, as shown in Scheme 2. During the polymerization, the viscosity of the mixture increased too much for stirring with a magnetic stirrer, and this was indicative of the formation of a high-molecular-weight product. The resulting polymer became less soluble once isolated; therefore, the polymerization mixture was directly cast onto a glass plate to fabricate a membrane. Although polymer 5 did not dissolve completely in DMF as a GPC solvent, GPC analysis of the dissolved part proved its high molecular weight (number-average molecular weight = 1.14 × 105, weight-average molecular weight = 1.75 × 105). In the IR spectrum of 5 (Fig. 1), three strong absorptions typical for imide groups were observed at 1349, 1674, and 1713 cm−1, which were assigned to the CN stretching vibration, asymmetric CO stretching vibration, and symmetric CO stretching vibration, respectively. The medium strong and broad absorptions at 1197 and 1253 cm−1 were assigned to symmetric and asymmetric SO stretching vibrations of the sulfonic acid groups. No peaks attributable to the aromatic amide and carboxylic acid groups were observed, and this suggested complete imidization.
The 1H NMR spectrum of 5 in DMSO-d6 showed peaks of aromatic protons at 6.4–9.0 ppm and aliphatic protons at 1.2–4.6 ppm, all of which were well assigned to the supposed chemical structure (Fig. 2). The integration ratio of these peaks also agreed with the structure. No peaks attributable to amic acids were detected.
Figure 3 shows the TG/DTA–MS curves of 5. In a typical TG curve, a two-step weight loss was observed from 30 to 130 °C and above 200 °C. Both weight losses were endothermic, as shown in the DTA curve. The mass number corresponding to the water molecule (m/z = 18) was observed in both steps, whereas the mass number corresponding to SO2 (m/z = 64) was observed only in the second step. The results suggest that the first weight loss was attributable to the evaporation of hydrated water, and the second weight loss was attributable to the decomposition of the pendant sulfonic acid groups. The CS bonds are more likely to be thermally decomposed than the other side-chain bonds (e.g., CO bonds). These thermal properties are comparable to those of other sulfonated polyimides with shorter or longer aliphatic side chains,18–21 and so we conclude that the phenylene rings introduced into the side chains do not practically affect the thermal stability.
Hydrolytic and Oxidative Stability
In accelerated tests, the hydrolytic and oxidative stability was investigated for polymer 5 under very harsh conditions (see the Experimental section). The weight loss of 5 was 19 wt % in the hydrolytic test and 16 wt % in the oxidative test. Although the stability is not high enough for practical applications, these weight-loss values are comparable to those of other side-chain-sulfonated polyimides.18–20 After these tests, 5 still remained a self-supporting and flexible membrane but was somewhat more brittle when bent. There are two possible degradation modes for the side-chain-sulfonated polyimides: imide linkage scission (main-chain degradation) and alkyl/aryl ether linkage scission (side-chain degradation), as clearly reported by Okamoto et al.14 Under operating fuel-cell conditions, however, the main-chain degradation seemed mainly to cause membrane brittleness for a similar side-chain-sulfonated polyimide membrane.23
Proton Conductivity and Water Uptake
Figure 4 shows the proton conductivity of 5 at 100% RH as a function of temperature. For comparison, another sulfonated polyimide having the same IEC (2.19 mequiv/g) but shorter side chains (6 in Chart 1) and Nafion 112 are also included. At a low temperature of 40 °C, both 5 and 6 membranes showed 2 times higher proton conductivity (ca. 0.2 S cm−1) than that of Nafion. With increasing temperature, the proton conductivity of 5 increased and reached 1.0 S cm−1 at 120 °C, which is 5 times higher than that at 40 °C, whereas the conductivity of 6 and Nafion increased to a lesser extent. The apparent activation energy for the proton conduction, which was roughly estimated from the slope, was about 29 kJ/mol for 6 and higher than that of 5 and Nafion 112 (<20 kJ/mol), probably because of the lower acidity of the aromatic sulfonic acid groups versus that of the aliphatic ones (e.g., pKa is ca. 0.7 for benzenesulfonic acid).
The water uptake and proton conductivity are plotted as a function of the humidity in Figure 5. As expected, the membranes of nonfluorinated polyimides 5 and 6 showed higher water uptake but lower proton conductivity than Nafion from 40 to 90% RH. This is attributed to the lower acidity of the nonfluorinated sulfonic acid groups and less developed hydrophilic domain as proton-transporting channels should be responsible. Membrane 5, having longer side chains, showed less water uptake and lower conductivity than 6. Similar trends were observed for the even longer side chain polyimides.21 Therefore, the most effective side-chain length seems to be three or four methylene units from the viewpoint of the proton-conducting properties.
To investigate the hydrophilic cluster in the membrane structure, SAXS was performed for 5, 6, and Nafion 117 membranes under fully hydrated conditions. As reported in the literature,11 the sulfonated polyimides displayed much lower scattering intensity than Nafion because of the less developed hydrophilic cluster. To compare the SAXS data, the scattering intensity was normalized with the following equation:
where I and Inorm are the scattering intensity and normalized scattering intensity, respectively; q is the scattering vector; and [INV] is the area of the q–Iq2 plot. Figure 6 shows the q–Inorm plot for these membranes. Nafion 117 showed a cluster peak at q = 1.26 nm−1 corresponding to d = 5.0 nm, which is consistent with the volume reported in the literature.24 On the other hand, both the 5 and 6 membranes also displayed a scattering maximum peak at the same position as that of Nafion but with much less normalized scattering intensity. The peaks were somewhat broader. These results indicate that the hydrophilic cluster in the sulfonated polyimides is similar in size to that of Nafion but not as uniform. A considerable amount of a smaller hydrophilic domain was also present for 5 and 6. The side-chain length did not affect any observable effect in the phase-separation behavior.
We synthesized a novel sulfonated polyimide having pendant sulfophenoxypropoxy groups. From 3,3′-dihydroxybenzidine, monomer 4 was successfully synthesized in four steps (acetylation, bromoalkoxylation, sulfophenoxylation, and acidification). Although polymer 5 gave very viscous solutions in organic solvents, a membrane could be fabricated by direct casting from the polymerization mixture. Thermal analyses revealed that 5 displayed a two-step weight loss, which is typical for a sulfonated polyimide. A membrane of 5 showed very high proton conductivity of 1 S/cm at 120 °C and 100% RH. On the other hand, the conductivity dropped at a low humidity and could not be improved by an increase in the side-chain length. SAXS measurements revealed that the hydrophilic cluster in polymer 5 was similar in size to that of Nafion. However, the weak intensity and wider distribution of the hydrophilic cluster in a smaller region indicated less pronounced phase separation of 5 versus Nafion. Through these studies, the optimum side-chain length for the polyimide ionomers was found to be three or four methylene units. Longer side chains and/or phenoxy groups resulted in lower proton conductivity, probably because of less connected proton channels and lower acidity.
This work was partly supported by the Japanese Ministry of Education, Culture, Sports, Science, and Technology through a Grant-in-Aid for Scientific Research (18750167) and the Fund for Leading Projects and by the New Energy and Industrial Technology Development Organisation through the Industrial Technology Research Grant Program (02B70007c).