J. Phys. Chem. B 2007, 111, 9437-9443
9437
ARTICLES Ionic Cluster Size Distributions of Swollen Nafion/Sulfated β-Cyclodextrin Membranes Characterized by Nuclear Magnetic Resonance Cryoporometry Jae-Deok Jeon and Seung-Yeop Kwak* Hyperstructured Organic Materials Research Center (HOMRC), and Department of Materials Science and Engineering, Seoul National UniVersity, San 56-1, Sillim-dong, Gwanak-gu, Seoul 151-744, Korea ReceiVed: February 5, 2007; In Final Form: April 30, 2007
Nafion/sb-CD membranes were prepared by mixing 5 wt% Nafion solution with H+-form sulfated β-cyclodextrin (sb-CD), and their water uptakes, ion exchange capacities (IECs), and ionic cluster size distributions were measured. Gravimetric and thermogravimetric measurements showed that the water uptake of the membranes increased with increases in their sb-CD content. The IECs of the membrane were measured with acid-base titration and found to increase with increases in the sb-CD content, reaching 0.96 mequiv/g for NC5 (“NCx” denotes a Nafion/sb-CD composite membrane containing x wt% of sb-CD). The clustercorrelation peaks and ionic cluster size distributions of the water-swollen membranes were determined using small-angle X-ray scattering (SAXS) and 1H nuclear magnetic resonance (NMR) cryoporometry, respectively. The SAXS experiments confirmed that increases in the sb-CD content of the membranes shifted the maximum SAXS peaks to lower angles, indicating an increase in the cluster correlation peak. NMR cryoporometry is based on the theory of the melting point depression, ∆Tm, of a liquid confined within a pore, which is dependent on the pore diameter. The melting point depression was determined by analyzing the variation of the NMR signal intensity with temperature. Our analysis of the intensity-temperature (IT) curves showed that the ionic cluster size distribution gradually became broader with increases in the membrane sb-CD content due to the increased water content, indicating an increase in the ionic cluster size. This result indicates that the presence of sb-CD with its many sulfonic acid sites in the Nafion membranes results in increases in the ionic cluster size as well as in the water uptake and the IEC. We conclude that NMR cryoporometry provides a method for determining the ionic cluster size on the nanometer scale in an aqueous environment, which cannot be obtained using other methods.
Introduction Fuel cells are considered a promising alternative for future energy needs because of their high efficiency and low environmental impact.1,2 Polymer electrolyte membrane fuel cells (PEMFCs) and direct methanol fuel cells (DMFCs) are fuel cells that utilize proton-conducting polymer membranes. Nafion is a commercially representative perfluorosulfonate ionomer that has been used in both types of fuel cells because it has high proton conductivity and excellent chemical/electrochemical stability. It has been widely reported that the material in Nafion membranes self-organizes to form hydrophobic and hydrophilic domains. The hydrophobic region consists primarily of poly(tetrafluoroethylene) (PTFE), which ensures the membrane’s long-term chemical stability in both reducing and oxidizing environments, and that the hydrophilic region contains a mixture of sulfonic acid groups, protons, water, and methanol.3 The microstructures of the hydrophilic domains of Nafion membranes can be understood in terms of “reverse-micelle-like” ionic clusters.4 The proton conductivity of these membranes is due to these networked ionic clusters. * Corresponding author. Seung-Yeop Kwak. Tel: +82-2-880-8365. Fax: +82-2-885-1748. E-mail:
[email protected].
DMFCs operate without fuel reformers and have significant advantages over PEMFCs: a simple system, quick start-up, easy maintenance, improved safety, portability, etc. Despite these advantages, there are several obstacles to the commercialization of DMFCs, such as their high cost, reduced proton conductivity, and, most significantly, the so-called “methanol crossover”. Methanol crossover through the membrane as a result of electroosmotic drag and the concentration gradient decreases the performance of the cell because of the resulting mixed potential.5 Methanol permeates into the membranes primarily through the ionic clusters, and thus the size and distribution of these clusters determines the methanol permeability. Therefore, an improved understanding of the size and distribution of these clusters might help improve the performances of fuel cell membranes. There have been employed various methods6-11 to determine the pore size distribution of porous membranes such as a microscopic observation method, a bubble pressure method, mercury intrusion porosimetry, permporometry, gas adsorptiondesorption, and differential scanning calorimetry (DSC) thermoporometry. These methods vary widely in applicability, sensitivity, and information that they yield. However, some methods have their specific disadvantages such as irreversible damage of the samples and time-consuming, which limited their
10.1021/jp070980a CCC: $37.00 © 2007 American Chemical Society Published on Web 07/24/2007
9438 J. Phys. Chem. B, Vol. 111, No. 32, 2007 applications for porous materials having the small pores. In the case of mercury intrusion porosimetry and gas adsorptiondesorption, it is necessary to suppose a structural model for the pores, making the interpretation of the results quite complex. DSC thermoporometry observes heat transfer in a measurement consisting of dynamic and isothermal steps, from which the amount of a liquid molten within given temperature ranges can be calculated with the help of the known enthalpy of fusion. However, DSC thermoporometry is somewhat limited in the pore size range detectable,11 implying that it is difficult to obtain detailed information about the size distribution of ionic clusters in the a few nanometer size range for fuel cell membranes. The work presented here uses two different methods, smallangle X-ray scattering (SAXS) and 1H nuclear magnetic resonance (NMR) cryoporometry, to determine the cluster correlation peaks and ionic cluster size distributions of Nafionbased composite membranes, respectively. A large number of SAXS investigations of Nafion membranes have previously been reported,12-15 and in general have provided estimates for the cluster diameter of approximately 4 nm and for the channel diameter of approximately 1 nm. The cluster and channel sizes are dependent on water content, so they change during water swelling and dehydration processes.14,16 Although there is some debate as to whether the observed scattering is produced by the interference between the approximately spherical ionclustered domains or by individual aggregates of various types, here this phenomenon will be referred to as “cluster” correlation. During the past few years, 1H NMR cryoporometry has been used by Strange et al.17 and Hansen et al.18 to determine the pore size distributions of porous materials. The theoretical basis for NMR cryoporometry is the well-known Gibbs-Thompson equation,19-21 which relates the melting point depression, ∆Tm, of a confined liquid to the pore diameter, D. This equation indicates that the difference between the normal and the depressed melting temperature is inversely proportional to the linear dimension of a liquid confined within the pores of a material. NMR cryoporometry makes use of the fact that the nuclear magnetic transverse or spin-spin relaxation time, T2, is very short for a solid but long for a liquid. By determining the amount of liquid that is present at different temperatures, the pore size distributions of materials can be calculated. Water is used as a probe molecule in these NMR measurements because water is the proton conductor in the membranes of PEMFCs and DMFCs. Therefore, it is possible to determine the cluster size distributions of Nafion-based membranes in the presence of water, and also therefore to monitor the variation in the cluster size distribution with water content. The water on the outside of the membranes acts as bulk water and provides a reference point for the change in melting temperature arising from the confinement. Based on these considerations, NMR cryoporometry has been successfully applied to porous nonmembrane materials, for example, uniform and nanoporous materials such as glasses, silica gels, and polymer particles.17,22-24 To our knowledge, there has been no previous direct and precise characterization of the cluster size distributions of Nafion-based membranes using 1H NMR cryoporometry. In this paper, we report the ionic cluster size distributions in Nafionbased membranes, which are determined using NMR cryoporometry. In order to reduce the methanol permeability of the Nafion membranes while minimizing the loss of proton conductivity, we add sulfated β-cyclodextrin into the membranes. Cyclodextrins (CDs) are cyclic oligosaccharides with six (R-), seven (β-), or eight (γ-) glucose units linked by 1,4R-glucosidic bonds.25 CDs have a shallow truncated cone shape
Jeon and Kwak and a hydrophobic cavity that is apolar relative to the outer surface. β-CD is the most accessible, lowest-priced and generally the most useful CD, and is used in this study. Sulfated β-CD is very hydrophilic because its external surface has many reactive sites, i.e., sulfonic acid groups. Therefore, the addition of hydrophilic sulfated β-CD into the Nafion membranes assists the transport of protons, since the number of reactive ionic cluster sites is increased. In addition, the presence of sulfated β-CD nanoparticles inside the ionic clusters in the membranes means that the methanol transport pathway is tortuous, resulting in a decrease in the methanol permeability. The objective of this study was to prepare Nafion/sulfated β-CD composite membranes with various H+-form sulfated β-CD content and to probe the ionic cluster size distributions for the swollen membranes with NMR cryoporometry. The detailed results of this experimental approach are presented and discussed. In the near future, we plan to report the performance, i.e., the proton conductivity and methanol permeability, of these composite membranes in DMFCs. Experimental Section Materials. Nafion-perfluorinated ion-exchange resin (5 wt % solution in a mixture of lower aliphatic alcohols and water) was purchased from Aldrich Chemicals and used as the membrane material; it has an equivalent weight (EW) of 1100 g for each sulfonic acid group. The Na+-form sulfated β-CD (typical substitution: 7-11 mol/mol β-CD) was purchased from Aldrich Chemicals. The H+-form sulfated β-CD (denoted hereafter as sb-CD) was obtained by recrystallization after adjusting the pH of the Na+-form sulfated β-CD solution. Fabrication of Composite Membranes. The composite membranes were prepared by the solution casting method.26-28 The desired amount of sb-CD was added into a 5 wt% Nafion solution, and then stirred at room temperature and degassed by ultrasonication. The sb-CD content of the mixture was 1, 3, or 5 wt% with respect to Nafion. The resulting mixture was slowly poured into a glass dish in an amount that would produce a formed membrane thickness of approximately 80 µm. The filled glass dish was evaporated at 40 °C for 2 days and then annealed at 120 °C in a convection oven for 2 h. After cooling, the membrane was peeled off the glass dish by the addition of water. The membranes were stored in deionized water so that they were water-saturated. In this study, NCx denotes a Nafion/sbCD composite membrane containing x wt% of sb-CD. Characterization. The swelling characteristics of the membranes were determined with water uptake measurements. The samples were completely dried under vacuum for 3 days at 30 °C and then weighed. They were then placed in deionized water for a week at 25 °C. Water on the surfaces of the wet samples was removed with filter paper, and then the samples were immediately transferred to a weighing dish and weighed. The water uptake was calculated according to the equation
water uptake (%) )
Wwet - Wdry × 100 Wdry
(1)
where Wwet and Wdry are the weights of the wet and dried membranes, respectively. To obtain further information about the water uptake of the samples, the swollen samples were analyzed with a TA model 2050 thermogravimetric analysis (TGA) instrument under a nitrogen atmosphere in the temperature range 30-600 °C at a heating rate of 10 °C/min. The ion exchange capacity (IEC) indicates the number of milliequivalents of ions in 1 g of the dry polymer. H+-form
Swollen Nafion/Sulfated β-Cyclodextrin Membranes
J. Phys. Chem. B, Vol. 111, No. 32, 2007 9439
samples of similar weight were soaked in 50 mL of 0.1 N NaCl solution for 24 h at 25 °C in order to achieve complete H+ to Na+ ion exchange. 10 mL of the released H+ was then titrated with 0.1 N NaOH solution, in which phenolphthalein was used as an indicator. The IEC was calculated from the titration data with the following equation:
IEC (mequiv/g) )
VNaOH × NNaOH × 5 Wdry
(2)
where VNaOH is the amount of NaOH required to neutralize a blank sample, NNaOH is the normality of the NaOH solution, 5 is the ratio of the amount of NaOH required to dissolve the sample to the amount used for titration, and Wdry is the weight of the dried sample. Small-angle X-ray scattering was carried out on a Bruker AXS Nanostar. In the experiments, the X-ray beam was produced with a rotating-anode X-ray generator operated at 40 kV and 35 mA, of which X-ray source was a monochromatized Cu KR (λ ) 1.54 Å) radiation. The scattering curves were corrected for background scattering, slit-length, and slit-width smearings. The samples were immersed in deionized water for a week and then three swollen pieces were placed within a Mylar bag under wet conditions. In order to carry out the 1H NMR cryoporometry, each sample was immersed in deionized water for a week, and its surface was wiped to remove excess water before the sample was packed into a 10 mm outer diameter NMR tube and sealed. The sample height was restricted to approximately 15 mm to ensure that the sample located within the transmitter/receiver coil was homogeneously irradiated. In order to prevent the evaporation and desorption of water from the samples, the spaces above the samples in the NMR tube were filled with hydrophobic perfluorooctane (C8F18, Aldrich Chemicals). The NMR cryoporometry measurements were carried out with a Bruker mq20 spectrometer at 0.47 T and a resonance frequency of 19.95 MHz. The spin-echo amplitude from 90°-τ-180°-τ-echo pulse sequence was measured with a pulse separation time, τ, of 10 ms to ensure that the signal was entirely from the liquid present. Since the spin-spin relaxation time of solid ice confined within pores of NCx membranes was very short, whereas the corresponding relaxation time of mobile water was long, the mobile water could be detected during 20 ms of total echo time. A 90° pulse length of 2 µs was applied with recycle delay, T > 5T1. The signal amplitude was measured as a function of temperature ranging from 210 to 277 K with an interval of 1 K using a Bruker BVT-3000 temperature control unit. During an experiment, the monitored temperature usually varied only by ( 0.1 K. All measurements were obtained by increasing the temperature after initially cooling the samples to a low temperature (i.e., 190 K) in order to prevent the complications of supercooling or hysteresis. The warming rate was low enough to achieve equilibrium, which was controlled by waiting above 10 min at each temperature step for NMR signal to stop changing. The signal intensity was corrected for temperature by implementing the Curie law; i.e., the observed signal intensity was multiplied by the factor T/To (To ) 273 K and T defining actual temperature). Results and Discussion Water Uptake and Ion Exchange Capacity (IEC). Figure 1 shows the equilibrium percentage sorption of water, obtained by soaking the membranes in water at 25 °C. The water uptake of the membranes was found to increase from 21.4 for NC0 to
Figure 1. Water uptake (obtained from gravimetric measurements) and IEC for the composite membranes as functions of sb-CD content.
Figure 2. TGA thermograms of the Nafion/sb-CD composite membranes.
TABLE 1: Water Uptake and IEC Data for the Composite Membranes water uptake (%) nafion/sb-CD gravimetric thermogravimetric method IEC method content (at 100 °C) (mequiv/g) membrane (wt %/wt %) (at 25 °C) NC0 NC1 NC3 NC5
100/0 99/1 97/3 95/5
21.4 ( 0.3 22.1 ( 0.2 23.3 ( 0.4 24.4 ( 0.3
16.6 ( 0.4 17.8 ( 0.3 19.6 ( 0.4 21.1 ( 0.3
0.89 ( 0.01 0.91 ( 0.00 0.94 ( 0.01 0.96 ( 0.01
24.4% for NC5 with the increase in the sb-CD content. In order to validate whether or not there was any loss of sb-CD during swelling, the weight of the dried membranes before and after water uptake (i.e., swelling) experiments was measured. As a result, there was no significant difference in membrane weight between before and after the experiments, implying that loss of sb-CD did not occur during swelling. Thermal analysis was performed to gain better information about the swelling behavior. As shown in Figure 2, a mass loss by about 100 °C corresponding to the loss of water was observed, the process of de-sulfonation began at about 300 °C, and above 420 °C, complete decomposition of all the membranes occurred. In addition, the water uptake of the membranes increases with increases in their sb-CD content, which is in good agreement with the results obtained from the gravimetric measurements. This result shows that the water uptake of the membranes can be controlled by varying their sb-CD content. Table 1 shows the water uptake results obtained from the gravimetric and thermogravimetric measurements. There is a discrepancy between the results of these two measurement methods. This difference can be explained in terms of the different conditions under which the two analyses were conducted. The ion exchange capacity (IEC) provides an indication of the number of ion-exchangeable groups present in an ion-
9440 J. Phys. Chem. B, Vol. 111, No. 32, 2007
Jeon and Kwak the cluster correlation peak for Nafion membranes are directly proportional to the volume of absorbed water.12,14 Cluster Size Distribution: NMR Cryoporometry. The theoretical basis for this NMR application is the well-known Gibbs-Thompson equation,19-21 which relates the melting point depression, ∆Tm, of a confined liquid to the pore diameter, D:
∆Tm(D) ) Tm - Tm(D) )
4σslTm D∆Hf Fs
(3)
where Tm is the normal bulk melting point, Tm(D) is the melting point of the confined solid, σsl is the crystal-liquid interfacial energy, ∆Hf is the bulk enthalpy of fusion, and Fs is the density of the frozen liquid (solid). It is assumed that σsl is isotropic, that the crystal is sufficiently large for the confined liquid to retain its bulk properties ∆Hf and Fs, and that the contact angle of the melting crystal is 180°.24,30 For a particular liquid, eq 3 takes the rather simple form:17,24
∆Tm(D) ) K
Figure 3. (a) SAXS curves for a swollen pristine Nafion membrane (NC0) and Nafion/sb-CD composite membranes (NC1, NC3, and NC5) and (b) peak position of the cluster correlation as a function of the sb-CD content.
conducting polymer membrane; these groups are responsible for the conduction of protons and thus the IEC is a reliable measure of the proton conductivity. The IEC of each membrane was determined by using the acid-base titration method. The results are shown in Figure 1 and listed in Table 1, and demonstrate that the IECs of the membranes increase from 0.89 to 0.96 mequiv/g with increases in the water uptake. Thus the introduction of sb-CD with its many reactive sulfonic acid sites into the membranes contributes to an increase in the IEC. Cluster Correlation Peak: SAXS. Under wet conditions, the SAXS results indicate that there are significant changes in the properties of the membranes with variation in the sb-CD content, i.e., with variation in the water content. Figure 3a shows the relative scattering intensity as a function of scattering vector, S, for the hydrated Nafion membrane (NC0) and the hydrated Nafion/sb-CD composite membranes (NC1, NC3, and NC5). The scattering vector, S, is defined by S ) 2 sin θλ. The scattering intensity increased with increases in the water content. This is due to the enhancement of the difference in the electron density by decreasing the electron density of the ionic clusters containing more water content relative to the backbone.29 It can be also seen that the maximum SAXS peak shifts to lower scattering vector (i.e., scattering angle) as the sb-CD content (i.e., water content) increases. This maximum peak position is inversely proportional to the cluster correlation peak. Therefore, these decreases in scattering angle indicate the expansion of the hydrophilic ionic clusters due to the increased cluster correlation peak as shown in Figure 3(b). In addition, by revealing these changes in the cluster correlation peak, the SAXS data confirmed the presence of sb-CD particles in the ionic clusters. It has previously been reported that the increases in
1 D
(4)
where K is a constant depending solely on the physical properties of the liquid confined within the porous material. Measurement of the amount of liquid confined within the pores as a function of temperature enables the pore size, D, and its distribution to be estimated when K is known. In this study, water was used as the probe liquid, for which K is approximately 62 K nm.31 This equation indicates that the difference between the normal and depressed melting temperatures is inversely proportional to a linear dimension of the liquid confined within the pores. The spin-echo signal intensity, V, indicates the amount of liquid water confined within the pores at a particular temperature, T, and thus the volume of the pores with linear dimensions equal to the corresponding the pore diameter, D, can be calculated. The volume of the pores with pore diameters between D and D + ∆D is (dV/dD)∆D. If the pores are filled with liquid, the pore size distribution (dV/dD) can be determined from the derivative of the intensity-temperature (IT) curve:
dV dV dTm(D) dI ) ) dD dD dTm(D) dD
(5)
From eq 4, dTm(D)/dD ) K/D2, so eq 5 becomes
dV K dV ) dD D2 dTm(D)
(6)
Therefore, the measurement of dV/dTm(D) enables the pore size distribution of the samples to be determined, provided K is known. Figure 4 shows the spin-echo intensity and its derivative as functions of temperature for bulk water. The signal intensity increases abruptly at approximately 273 K, at which dI/dT is maximum. Thus, the transition temperature from the solid to the liquid state of bulk water is approximately 273 K, in very good agreement with the known value of approximately 273.15 K. This melting point was used as the reference temperature. In addition, Figure 4 shows that the first melting of bulk water starts at approximately 271 K, corresponding to a temperature gradient across the sample of approximately 2 K. This is due to sample height of 1.5 cm, which is rather large. The Hansen group32 also reported that the temperature across the sample was changed by approximately 2 K. In order to confirm the change of relaxation properties for bulk water, T2’s were
Swollen Nafion/Sulfated β-Cyclodextrin Membranes
Figure 4. NMR data for bulk water: the spin-echo signal intensity (9) and the derivative (O) of the intensity as functions of temperature.
Figure 5. Spin-echo signals of water confined within the ionic cluster pores of a pristine Nafion membrane, NC0, at 210, 230, 240, and 273 K.
measured at various temperatures. As a result, it was found that T2’s of solid ice were constant and very short below melting temperature. This indicates that the relaxation properties of water do not change below melting temperature, implying that there is no significant temperature-dependent loss. Figure 5 shows the 1H spin-echo signals of water confined within the ionic cluster pores of a pristine Nafion membrane, NC0, at various temperatures. It can be clearly seen in these signals that the spin-echo intensities increase gradually with increasing temperature. At low temperatures, the spin-echo intensities from solid water were found to be 0%. For samples at temperatures close to 273 K, corresponding to the melting point of molten water, the intensities of the liquid phase were determined to be approximately 100%. In addition, Figure 5 shows two echoes at T ) 273 K, a broad echo and a narrow echo which is superposed on a broad echo. These broad and narrow echoes may be caused due to bulk water and poreconfined water, respectively. On the other hand, Pineri et al.33 reported that the desorption of a certain amount of water out of the ionic phase was followed by freezing of the water, thereby resulting in the decrease of NMR signal. However, in this study, the evaporation and desorption of water does not occur during freezing because the samples containing hydrophilic water in the NMR tube are filled with hydrophobic perfluorooctane. Since these perfluorooctane molecules contain practically no protons, only water molecules are seen in this experiment. Figure 6 shows the IT curves of water confined within the ionic cluster pores of the Nafion/sb-CD composite membranes. In the case of the solid phase, the relative signal intensities were denoted as 0, corresponding to totally frozen water. The relative signal intensities from the liquid phase of totally molten water were denotes as 1. These curves show that the signal intensities increase gradually with increasing temperature (below 273 K) due to the gradual melting of the frozen water confined within
J. Phys. Chem. B, Vol. 111, No. 32, 2007 9441
Figure 6. Relative signal intensities vs temperature (IT curves) of water confined within the ionic cluster pores of the Nafion/sb-CD composite membranes.
the ionic cluster pores in the membranes. The abrupt increase in intensity at the melting point of bulk water, 273 K, is due to the melting of the bulk supernatant water. If the amount of liquid confined within the pores of the membranes and K are known, their cluster size distribution can be estimated using eq 6. It has generally been reported that K is in the range 41-73 K nm, when water is used as the probe liquid. We assumed that K ) 62 K nm to obtain the cluster size distributions. These data were used in eq 6 to obtain the relative pore volume as a function of pore diameter, as shown in Figures 7a-d. On the other hand, colligative effect will also affect the IT curves due to the addition of different sb-CD content. This effect is a phenomenon by which the freezing/ melting point of a solution is lowered when more solute is dissolved in the solution. In order to validate the colligative effect of water/sb-CD solutions confined within ionic clusters of composite membranes, the melting point of watersb-CD solutions was measured by using the NMR spin-echo method. The watersb-CD solutions confined within ionic clusters of NC1, NC3, and NC5 composite membranes had various sb-CD concentrations of 4, 11, and 17 wt%, respectively. These concentrations were determined by assuming that water (the amount of water measured by the gravimetric method) and sb-CD particles were confined within only ionic clusters of composite membranes. As a result, the melting point depressions of the water/sb-CD solutions with 4, 11, and 17 wt% concentrations compared to pure water were found to be 2, 6, and 8 K, respectively. Therefore, the original IT curves of the composite membranes were corrected by adding their melting point depressions to original temperature because the original IT curves in Figure 6 shifted to lower temperature due to colligative effect. The dashed lines in Figure 7 represent the pore size distribution curves corrected by these melting point depressions, which are caused by colligative effects. From this result, it was confirmed that the ionic cluster size of Nafion/nanoparticle composite membranes measured by NMR cryoporometry seemed to be lowered by colligative effect. In Figure 7, it is clear that the cluster size distribution of the composite membranes broadens gradually with increases in their sb-CD content; i.e., there is an increase in the ionic cluster pore size. This is probably due to the expansion of the hydrophilic ionic clusters due to the increased water content, which is caused by the addition of sb-CD with its many sulfonic acid groups into the hydrophilic ionic clusters. This increase in cluster size is in agreement with the trend in the results obtained from the SAXS measurements, but the cluster sizes obtained with NMR cryoporometry are smaller as listed in Table 2. The Hansen group23 has reported that the melting point depression of water is less sensitive to pore radius than those of benzene and cyclohexane because
9442 J. Phys. Chem. B, Vol. 111, No. 32, 2007
Jeon and Kwak
Figure 7. Pore size distribution curves of the Nafion/sb-CD composite membranes: (a) NC0, (b) NC1, (c) NC3, (d) NC5. The vertical scale is arbitrary. The dashed lines of NC1, NC3, and NC5 are corrected by the melting point depressions, which are caused by colligative effect.
TABLE 2: SAXS and NMR Cryoporometry Data for the Swollen Composite Membranes cluster size distribution (nm) from NMR measurements
membrane
cluster correlation peak (nm) from SAXS measurements
original data
data corrected by colligative effecta
NC0 NC1 NC3 NC5
4.0 4.4 4.6 4.7
1.2-2.4 1.2-2.5 1.2-2.7 1.3-3.1
1.2-2.4 1.2-2.7 1.4-3.4 1.6-3.9
a Corrected data were calculated assuming that water and sb-CD particles were confined within only ionic clusters of composite membranes.
water has a smaller K value and the surface layer of water acts physically as part of the pore wall, resulting in a decrease in the effective pore radius. Despite this weakness, water has been used in NMR cryoporometry as a probe molecule because water is easily imbibed into hydrophilic pores and is a better probe molecule for characterizing small pores.23 Overall, these results show that 1H NMR cryoporometry can be used to determine the ionic cluster size distributions of Nafion-based membranes and that this method provides a means with which to determine pore sizes on the nanometer scale in an aqueous environment, which cannot be achieved with other methods. Conclusions Nafion/sulfated β-cyclodextrin (Nafion/sb-CD) composite membranes with sb-CD content of 1, 3, and 5 wt% were prepared with the solution casting method. The H+-form sbCD was obtained by recrystallization after adjusting the pH of the Na+-form sb-CD solution and used to prepare the membranes. Thermal analysis was performed to investigate the water uptake of the composite membranes and the uptake values were
found to increase with increases in their sb-CD content, which is in good agreement with the gravimetric results. This result shows that the amount of water uptake in the membranes can be controlled by varying their sb-CD content. The acid-base titration results show that the IECs of the membranes increase from 0.89 to 0.96 mequiv/g with increases in their sb-CD content, indicating that the introduction of sb-CD with its many reactive sites (sulfonic acid groups) into the membranes results in an increase in IEC. SAXS and 1H NMR cryoporometry results were obtained to determine the cluster correlation peak and ionic cluster size distributions of the membranes, respectively. The SAXS results show that the cluster correlation peak increases with sb-CD content. The NMR cryoporometry method utilizes measurements as a function of temperature, T, with a spinecho pulse sequence (90°-τ-180°-τ-echo) of the signal intensities, I, which are directly proportional to the non-frozen volume fraction of the liquid. From the intensity-temperature (IT) curves, it was determined that the cluster sizes of the membranes increase with increases in their sb-CD content due to their increased water content; i.e., there is an increase in the ionic cluster pore size. Thus the presence of sb-CD with its many sulfonic acid groups in the Nafion membranes leads to increases in their water uptake, IEC, and ionic cluster size. In conclusion, 1H NMR cryoporometry was found to be a very effective method for characterizing the ionic cluster size distributions of Nafion-based membranes, which are strongly correlated with the performance of membranes for polymer electrolyte fuel cells (PEMFCs). Acknowledgment. The authors are grateful to the Korea Science and Engineering Foundation (KOSEF) for supporting this study through the Hyperstructured Organic Materials Research Center (HOMRC). The corresponding author, S.-Y. Kwak, gratefully acknowledges Professor Colin A. Fyfe at the
Swollen Nafion/Sulfated β-Cyclodextrin Membranes Department of Chemistry, University of British Columbia (UBC), for his valuable suggestions and comments regarding the NMR results. References and Notes (1) Ren, X.; Zelency, P.; Thomas, S.; Davey, J.; Gottesfeld, S. J. Power Sources 2000, 86, 111-116. (2) Smitha, B.; Sridhar, S.; Khan, A. A. Macromolecules 2004, 37, 2233-2239. (3) Chou, J.; McFarland, E. W.; Meitu, H. J. Phys. Chem. B 2005, 109, 3252-3256. (4) Gierke, T. D.; Hsu, W. Y. Macromolecules 1982, 15, 101-105. (5) Jiang, R.; Chu, D. Electrochem. Solid State Lett. 2002, 5, A156159. (6) Kim, J. Y.; Lee, H. K.; Kim, S. C. J. Membr. Sci. 1999, 163, 159166. (7) Calvo, J. I.; Herna´ndez, A.; Pradanos, P.; Martinez, L.; Bowen, W. R. J. Colloid Interface Sci. 1995, 176, 467-478. (8) Schneider, P.; Uchytil, P. J. Membr. Sci. 1994, 95, 29-38. (9) Pra´danos, P.; Rodriguez, M. L.; Calvo, J. I.; Herna´ndez, A.; Tejerina, F.; de Saja, J. A. J. Membr. Sci. 1996, 117, 291-302. (10) Ishikiriyama, K.; Todoki, M.; Motomura, K. J. Colloid Interface Sci. 1995, 171, 92-102. (11) Gane, P. A. C.; Ridgway, C. J.; Lehtinen, E.; Valiullin, R.; Furo´, I.; Schoelkopf, J.; Paulapuro, H.; Daicic, J. Ind. Eng. Chem. Res. 2004, 43, 7920-7927. (12) Fujimura, M.; Hashimoto, T.; Kawai, H. Macromolecules 1981, 14, 1309-1315. (13) Gebel, G.; Lambard, J. Macromolecules 1997, 30, 7914-7920. (14) Elliot, J. A.; Hanna, S.; Elliott, A. M. S.; Cooley, G. E. Macromolecules 2000, 33, 4161-4171. (15) Haubold, H.-G.; Vad, Th.; Jungbluth, H.; Hiller, P. Electrochim. Acta 2001, 46, 1559-1563.
J. Phys. Chem. B, Vol. 111, No. 32, 2007 9443 (16) Elliot, J. A.; Hanna, S.; Eliott, A. M. S.; Cooley, G. E. Polymer 2000, 42, 2251-2253. (17) Strange, J. H.; Rahman, M.; Smith, E. G. Phys. ReV. Lett. 1993, 71, 3589-3591. (18) Schmidt, R.; Hansen, E. W.; Sto¨cker, M.; Akporiaye, D.; Ellestad, O. H. J. Am. Chem. Soc. 1995, 117, 4049-4056. (19) Gibbs, J. W. The Collected Works of J. Williard Gibbs; Academic Press: New York, 1928. (20) Thompson, W. Philos. Mag. 1871, 42, 448-452. (21) Jackson, C. L.; McKenna, G. B. J. Phys. Chem. 1990, 93, 90029011. (22) Aksnes, D. W.; Kimtys, L. Solid State Nucl. Mag. 2004, 25, 146152. (23) Hansen, E. W.; Schmidt, R.; Sto¨cker, M. J. Phys. Chem. 1996, 100, 11396-11401. (24) Hansen, E. W.; Fonnum, G.; Weng, E. J. Phys. Chem. B 2005, 109, 24295-24303. (25) Del Valle, E. M. M. Process Biochem. 2004, 39, 1033-1046. (26) Grossi, N.; Espuche, E.; Escoubes, M. Sep. Purif. Technol. 2001, 22-23, 255-267. (27) Moore, R. B.; Martin, C. R. Macromolecules 1988, 21, 13341339. (28) Gebel, G.; Aldebert, P.; Pineri, M. Macromolecules 1987, 20, 1425-1428. (29) Marx, C. L.; Caufield, D. F.; Cooper, S. L. Macromolecules 1973, 6, 344-353. (30) Dore, J.; Webber, B.; Strange, J.; Farman, H.; Descamps, M.; Carpentier, L. Physica A 2004, 333, 10-16. (31) Valckenborg, R.; Pel, L.; Kopinga, K. Magn. Reson. Imag. 2001, 19, 489-491. (32) Hansen, E. W.; Sto¨cker, M.; Schmidt, R. J. Phys. Chem. 1996, 100, 2195-2200. (33) Pineri, M.; Volino, F. J. J. Polym. Sci., Part B: Polym. Phys. 1985, 23, 2009-2020.
