2848
J. Phys. Chem. B 2008, 112, 2848-2858
Relationships of Acid and Water Content to Proton Transport in Statistically Sulfonated Proton Exchange Membranes: Variation of Water Content Via Control of Relative Humidity Timothy J. Peckham,† Jennifer Schmeisser,†,§ and Steven Holdcroft*,†,‡ Department of Chemistry, Simon Fraser UniVersity, Burnaby, British Columbia V5A 1S6, Canada, and Institute for Fuel Cell InnoVation, National Research Council Canada, 3250 East Mall, VancouVer, British Columbia V6T 1W5, Canada ReceiVed: September 8, 2007; In Final Form: December 16, 2007
An in-depth analysis for proton exchange membranes to examine the effects of acid concentration and effectiVe proton mobility upon proton conductivity as well as their relationship to water content was carried out on two main-chain, statistically sulfonated polymers at 25 °C. These polymer systems consisted of poly(ethylenetetrafluoroethylene-graft-polystyrenesulfonic acid) (1) and sulfonated trifluorostyrene (BAM) membrane (2). Nafion (3) was used for comparison. Water content (as represented by Xv, the water volume fraction, where Xv ) volume of water in hydrated PEM ÷ volume of hydrated PEM), for each sample was varied by adjusting the relative humidity (RH) of the membrane environment from 50% to 98%. It was found that, at low RH (RH < 70%), the major factor determining proton conductivity is proton mobility. In order to remove the differences in acid strength for the membranes, proton mobility values at infinite dilution (Xv ) 1.0) and 25 °C were calculated and found to be 2.6 ( 0.2 × 10-3 (average of 1a-c), 1.6 ( 0.3 × 10-3 (average of 2a-e), and 2.32 ( 0.01 × 10-3 cm2 s-1 V-1 (3). These were then compared to the theoretical value for the mobility of a free proton at infinite dilution and to previously reported data. Possible differences in tortuosity and the juxtaposition of acid groups are proposed in order to account for the significant deviations of all samples from the theoretical value.
1. Introduction Research on proton exchange membranes (PEMs) has been an area of active interest for the past 10 years, mostly as a result of both public interest in the development of zero-emission vehicles and power sources as well as significant governmentfunded initiatives.1-3 As a primary component of PEM fuel cells (PEMFCs), the PEM is required to perform a number of functions that include gas separation, acting as an electrical insulator and providing an ionic path for proton transport from anode to cathode.4 To date, Nafion membranes have achieved the greatest success, providing the most attractive, commercially available combination of performance, durability, and reliability. However, with increasingly more stringent requirements for automotive and stationary PEMFC applications, new PEMs with improved properties are required.5,6 The large majority of attempts to develop new PEMs have used an iterative or “hit and miss” approach. Although a number of scientifically interesting materials have been prepared this way, having a fundamental understanding of structure-property relationships would enable a more effective development of new membranes. In particular, there has been a focus on understanding how polymer structure and morphology relate to the observed levels of proton conductivity.7-43 When researching new membranes, studies on proton transport efficiency generally involve an interpretation of proton * Corresponding author. E-mail:
[email protected]. † Simon Fraser University. ‡ National Research Council of Canada. § Current address: Department of Chemistry and Biochemistry, 273-1 Essex Hall, 401 Sunset Avenue, University of Windsor, Windsor, Ontario N9B 3P4, Canada.
conductivity data (as measured by AC impedance spectroscopy) as a function of ion exchange capacity (IEC) or water uptake. These results may be combined with information obtained from supplementary techniques (e.g., transmission electron microscopy (TEM) and X-ray diffraction (XRD)) and correlated for a series of different materials with differences between the connectivity and size of the water-saturated channels resulting from phase separation of hydrophobic and hydrophilic domains. The benefits of a systematic analysis of proton conductivity data for PEMs have been previously reported.5,26 In a recent paper,19 we described an approach to the analysis of PEM conductivity data that illustrated the strong links between proton conductivity and the combination of water and acid content through an examination of the conductivity data for a series of main-chain, statistically sulfonated PEMs. Proton conductivity was then further broken down into its constituent components of acid concentration and proton mobility using eq 1:
σH+ ) F[-SO3H]µ′H+
(1)
where σH+ is the proton conductivity, F is the Faraday constant, [-SO3H] is the bulk, analytical acid concentration, and µ′H+ is the effective proton mobility, which includes the mobility of “free” protons and the degree of dissociation of protons, the latter of which cannot be determined. Using this method, it was possible to determine that the observed differences in proton conductivity behavior for a series of different water-saturated, main-chain, statistically sulfonated PEMs as a function of water content were mainly due to differences in effectiVe proton mobility19 and acid concentration. This was most clearly seen for sulfonated trifluorostyrene
10.1021/jp077218t CCC: $40.75 © 2008 American Chemical Society Published on Web 02/21/2008
Variation of Water Content via RH Control
J. Phys. Chem. B, Vol. 112, No. 10, 2008 2849
Figure 1. Analysis of RH conductivity data.
(BAM) membranes (2) for which higher ion content led to lower proton conductivity values due to very high water uptakes and a resultant dilution of available protons, rather than the normally expected ever-increasing value of proton conductivity as a function of IEC. Furthermore, it was possible to calculate values for effective proton mobility at infinite dilution, and it was found that in at least two of the PEM series examined the values were significantly lower than the theoretical value for a free proton. These deviations were attributed to possible differences in tortuosity and the proximity of acid groups. For the above study, we used membranes that had been allowed to equilibrate in liquid water. This is a common and rapid method for determining the conductivity of PEMs, with water contents varying as a function of ionic content. It is known, however, that PEMs uptake more water when they are in contact with liquid water rather than vapor, even when the effective RH value of the membrane environment is 100%. This phenomenon is known as “Schroeder’s paradox”.44 For example, in the case of Nafion membranes, the λ value (mol H2O/mol -SO3H) can be as much as six H2O molecules greater in the case of the water-saturated membrane versus the situation for a membrane at RH ) 100%.45 While it is certainly quite possible that there may be regions of the PEM being exposed to liquid water during fuel cell operation, it is also likely that the PEM will be exposed to water in the vapor state. This would be especially true at the low RH levels ( 70%, conductivity values for 1 increase rapidly in comparison to 3, and, for RH > 90%, conductivity values of 2b-c are also higher than those for 3. In the case of 2d-e, roughly equivalent conductivity versus 3 is only observed at RH ) 98%. Conductivity for 2a, however, remains consistently low, even at RH ) 98%. More information can be obtained for these systems by examining their proton conductivity behavior as a function of water content (Figure 3b,c). At low water contents, 3 exhibits the highest “effective” use of water in comparison to the other PEMs. This can be seen most clearly in Figure 3c, wherein the conductivity for 3 is higher as a function of Xv over the range of 0.2-0.4. This is also in spite of the fact that 3 has the lowest Xv values as a function of RH (Figure 4a). As a function of λ, 3 only exhibits higher conductivity at λ < 10. However, the conductivity of 3 is approximately 3 times higher at λ ∼ 7, in comparison to the other PEMs. Again, this is in spite of the fact that 3 actually exhibits slightly lower λ values as a function of RH, as illustrated in Figure 4b, although the λ value for 3 is high as a function of Xv in comparison to the other PEMs examined, as shown in Figure 4c. This conductivity behavior of 3 has previously been reported by Kreuer in comparison to a main-chain, statistically sulfonated PEM, sulfonated poly(ether ether ketone ketone) (SPEEKK).5 At lower water contents, the lower dissociation constant of SPEEKK in comparison to 3 leads to less effective screening of the negative charge of the sulfonate group, thereby increasing localization of the proton. In fact, it has been calculated that complete separation of the proton from the tethered anion does not occur for λ < 6 (i.e., at λ < 10, all PEMs are close to this threshold).46 In addition, the smaller channels in SPEEKK and the larger number of “dead ends” lead to a higher percolation threshold in comparison to 3 with decreasing water content. Given that both 1 and 2 show only small amounts of microphase separation at best and with acid strengths similar to that of SPEEKK, the explanation proposed for SPEEKK by Kreuer may also apply to 1 and 2.
PEMs 1a-c exhibit consistently higher conductivity than 2a-e over the entire range of RH. In the case of 2a and 2b, this may be due to either a higher acid or water content of 1ac. However, in comparing 1a and 2c, wherein the IEC values are within ∼8% of each other and λ values are similar for a given RH value (Figure 4b), the conductivity of 1 is consistently higher (>30%) as a function of λ than that for 2c. A possible explanation for these observations is detailed in the next section. Within the set of BAM membranes (2), there are interesting deviations from expected conductivity behavior with increasing IEC. This was seen in our original study with water-saturated membranes where conductivity reached a maximum for this series at IEC ) 1.96 mequiv/g, and then decreased with a further increase in IEC. The behavior was explained by the increasing uptake of water with increasing IEC such that, at the higher IEC values, proton concentration was significantly lowered, leading to overall lower conductivity for the higher IEC samples.19 Similarly, the order of increasing conductivity level over the range of RH ) 95-98% is 2c > 2b > 2d ∼ 2e > 2a. At lower RH levels, 2b and 2c display similar conductivities as do 2a, 2d, and 2e at RH < 80%. Unavailable in the previous study, it is now interesting to see how conductivity values for each polymer in the BAM membrane series are affected by water content. Whereas 2c exhibited higher conductivity values as a function of RH than 2b, it appears that 2b actually achieves higher conductivity values as a function of λ. The reverse situation is true in conductivity as a function of Xv (Figure 3c), where 2c exhibits higher values than 2b. This latter result can be explained by examining the relationship between λ and Xv (Figure 4c). It can be seen from this plot that 2c achieves higher λ values as a function of Xv than 2b does, hence the results observed in Figure 3c. This does not, however, explain the results observed in Figure 3b, wherein it might be anticipated that the combination of higher IEC with increasing λ values would lead to higher levels of conductivity for 2d-e in comparison with the rest of series 2. This will be discussed in further detail in the next section within the context of proton mobility and acid concentration, as will the rather unusual conductivity behavior of both 2d and 2e.
Variation of Water Content via RH Control
J. Phys. Chem. B, Vol. 112, No. 10, 2008 2851
Figure 3. Proton conductivity of 1-3 as a function of RH (a), as a function of λ (b), and as a function of Xv (c).
Figure 4. Xv (a) and λ (b) of 1-3 as a function of RH; λ of 1-3 as a function of Xv (c).
2.2. Effective Proton Mobility and Acid Concentration as a Function of Acid and Water Content. The mobility of protons as a function of RH for polymers 1 - 3 can be seen in Figure 5a. At high RH (i.e., RH > 90%), there are clear differences in mobility between the three different polymer systems. As previously discussed in the literature,6,19,47,48 the proximity of acid groups has an effect upon proton mobility by determining the size of the energy barrier that has to be overcome in order for protons to move from the sphere of influence of one sulfonate group to the next; i.e., the greater the distance between acid groups, the higher the expected energy barrier. Therefore, the higher IEC values of 1, 2b, and 2c might therefore lead to the observed higher mobility values in comparison to 3 (IEC ) 0.91 mequiv/g). Also, with higher acid contents (IEC) versus 2b and 2c, 1a-c exhibit higher mobility values. Interestingly, however, 2a has a considerably lower mobility than 3, in spite
of its higher IEC (IEC ) 1.36 versus 0.91 mequiv/g, respectively). Perhaps most unusually, however, are the low mobility values for 2d-e. Even though they possess very high IEC values (2.20 and 2.46 mequiv/g, respectively) in comparison to the PEMs analyzed in this study, their mobility values only reach that of 3 at RH ) 98% and are significantly lower than those in the case of 1a-c and even samples of the same polymer system (i.e., 2b-c). As RH is reduced, both 1 and 2 exhibit sharper decreases in mobility than 3 such that, by RH ) 90%, all samples of 2 display lower mobility values than 3, and the differences between 1 and 3 are smaller. By RH ) 60%, all samples of 2 exhibit similar mobility values, irrespective of IEC, and 3 now exhibits the highest mobility value. As previously discussed, this may be due to the differences in acid dissociation constants between perfluorosulfonic acid-based systems (3, pKa ∼ -6) and benzenesulfonic acid-based systems (1 and 2, pKa ∼ -2):
2852 J. Phys. Chem. B, Vol. 112, No. 10, 2008
Figure 5. Effective proton mobility of 1-3 as a function of RH (a), as a function of λ (b), and as a function of Xv (c).
at lower RH values, protons are less dissociated in the case of 1 and 2 and, therefore, exhibit lower mobility values. Alternatively, or in conjunction, it may also be due to a more fully defined percolation network in the case of 3 at these low RH values. Examining proton mobility for these systems as a function of water content reveals additional information. Figure 5b shows proton mobility as a function of λ. In general, there is decreasing mobility with decreasing λ values. This is expected since, with less water, there is less effective screening of the anionic charge of the sulfonate groups, and thus proton transport becomes increasingly more difficult. In addition, the percolation pathway may be compromised, leading to narrow “necks” and “dead ends”. For λ > 15, mobility is highest for 1, although it should also be noted that λ in the case of 2a and 3 does not rise higher than ∼17, whereas λ reaches ∼23 for 1, 2b, and 2c and ∼4060 in the case of 2d-e. In the region of λ ∼ 10-15, 2b-c
Peckham et al. exhibit similar mobility values, whereas those for 2a are lower. This may be due to the higher acid content of 2b-c versus 2a. With acid groups potentially in closer proximity in the former case versus the latter, it might be expected that mediation of a proton between immobile sulfonate groups has a lower energy barrier to overcome5,19,47,48 in the case of 2b-c, assuming, of course, that the tortuosity of the proton conduction pathways are similar, which might not be the case. However, this does not explain the significantly lower mobility values of 2d-e in comparison to all the other polymer samples, including 2a-c. Another area of interest is to compare the mobility of 1 and 2 as a function of λ. Whereas the acid content (IEC) of 1a is similar to that of 2c (within 8%) and slightly higher than that of 2b (within 14%), the mobility of 1a is significantly higher for a comparable value of λ, where λ > 10 in the case of both polymer systems. For example, approximating linear behavior for mobility as a function of λ, it can be seen that the mobility of 1a as a function of λ is approximately 30% greater than that of either 2b or 2c. This is in contrast to comparisons between 2b and 2c where acid content is similar (within 5%) as are mobility values (within 5%) at λ ) 20. This also applies to comparisons between projected values for 2a with 2b and 2c. With mobility values possibly higher than would be expected based solely on acid content differences, this may suggest that there is another factor contributing to increased mobility in the case of 1a-c versus 2a-c, such as a comparatively reduced degree of tortuosity. A similar explanation could be used for the higher than expected mobility values (based on acid content) for 3 in comparison to either 1 or 2. In the case of 2d-e, there may be some degree of tortuosity not observed in 2a-c, which results in lower than expected mobility values. Further information on the effect of water content on mobility can be seen in Figure 5c, where mobility is plotted as a function of Xv. In this case, some of the ordering has changed. For example, 2a exhibits lower mobility values as a function of λ (Figure 5b) than 2b, but higher values as a function of Xv (Figure 5c). These can be explained by the fact that 2a achieves higher λ values as a function of Xv (Figure 4c); i.e., in order for 2a to have an Xv value similar to that of 2b, its corresponding λ value must be higher because its IEC is lower. The information in Figure 5c can, however, be used to predict mobility values for polymers 1-3 at infinite dilution. These results will be discussed in section 2.4. Analytical acid concentration is another important factor for proton conductivity in PEMs. Figure 6a shows the relationship between acid concentration and RH for polymers 1-3. As expected, the decrease in water content with decreasing RH leads to higher acid concentrations for all the studied polymers. Decreasing water content leads to less dilution of the acid groups and hence [-SO3H]. In our previous study,19 we stated that the proximity of acid groups likely has an effect upon proton mobility; i.e., the closer the acid groups are located to one another, the easier it will be for protons to be mediated from one sulfonate group to the next, and the higher the resultant proton mobility. Thus, a higher acid concentration may lead to a higher mobility. However, this is based on the assumption that λ values and hence proton dissociation remain relatively high (i.e., λ > 15). In the case of this study, acid concentration increases with decreasing water content. Reducing the amount of water leads to less effective screening and dissociation of the proton from the sulfonate group and a concomitant decrease in proton mobility. Therefore, increasing acid concentration as a function of RH is not necessarily an inherent advantage and is more likely
Variation of Water Content via RH Control
Figure 6. Acid concentration of 1-3 as a function of RH (a), as a function of λ (b), and as a function of Xv (c).
to have a detrimental effect upon proton conductivity. Nevertheless, as will be seen, differences in acid concentration may be used to explain some of the observed differences in proton conductivity. In Figure 6a, the general trend, as might be expected, is that higher acid content (i.e., IEC) leads to higher acid concentration over the observed RH range. An exception is observed at RH > 95%, where the polymer with the highest acid content, 2e (IEC ) 2.46 mequiv/g), experiences a sharp drop in acid concentration. This was previously observed for the watersaturated membranes wherein the high IEC (>2 mequiv/g) samples of 2 absorb large amounts of water, leading to proton dilution and a detrimental effect upon proton conductivity. At lower RH levels, all polymers lose water, and thus acid concentration effectively increases.
J. Phys. Chem. B, Vol. 112, No. 10, 2008 2853 It is interesting to note, however, that there are two polymers (2b and 3, where IEC ) 1.86 and 0.91 mequiv/g, respectively) that appear to be out of sequence as based on acid content. This can also be seen in Figure 6b,c. In the case of 3 versus 2a, the probable explanation is that it absorbs less water (λ) than 2a. A similar explanation can be used for the comparison between 2b and 2c. This would thus infer a greater degree of dilution of the acid groups and hence a lower acid concentration. Given the different structures of 2a and 3, this behavior may be expected. However, as 2b and 2c are members of the same series and thereby would be expected to possess similar structures, it is not clear what is/are the underlying cause(s) for the differences between 2b and 2c. By examining values for proton mobility and acid concentration in conjunction, it is possible to explain some of the observed trends in conductivity as a function of water content. For example, 2b exhibits slightly higher conductivity than 2c as a function of λ. This is primarily due to the higher acid concentration of 2b since 2b and 2c exhibit very similar mobility values. This situation also applies in the case of 1 versus 3 in the region of λ ∼ 15-20 where the higher conductivity of 1 is due to its greater acid concentration (mobility values for both systems in this region are similar). In the case of the higher conductivity of 1a compared to 2b (for which IEC values are similar, 2.12 and 1.96 mequiv/g, respectively); however, it is the higher mobility of 1a that leads to the observed conductivity values since the acid concentration values of 2b are actually higher than 1a as a function of λ. Similarly, the almost equal conductivity values of 2a in comparison to 2d-e at a given λ value (Figure 3b), despite the large differences in IEC, are primarily due to the higher mobility values of 2a, as acid concentration is considerably greater for 2d-e versus 2a. Finally, it is interesting to note that, in spite of generally having the lowest acid concentration of all the polymers with the exception of 2a (acid concentration values of 2a and 3 are comparable as a function of λ), the high proton conductivity of 3 at λ < 10 is due to its higher mobility. Again, this would be expected based on the lower pKa of perfluorosulfonic acids in comparison to that of arylsulfonic acids as well as the presumably more connected percolation pathways in 3, hence the expected greater degree of dissociation (and hence mobility) of protons in 3 at low λ values in comparison to protons in 1 and 2. 2.3. Proton Mobility at Infinite Dilution. In our previous study,19 we removed the effect of the different acid strengths of the analyzed polymer systems from µ′H+ by projecting a calculated µ′H+ at Xv ) 1.0 (i.e., at infinite dilution). This allowed us to determine whether the mobility of protons in membranes with tethered sulfonate groups is affected by contributing factors other than acid strength. In this paper, data for polymers 1-3 were similarly analyzed using linear regression. The calculated values for µ′H+ at Xv ) 1.0 can be seen in Table 1. The results for 1 and 2 can be compared with those obtained from our previous water-saturated membrane conductivity data analysis.19 In that study, 1 (IEC ) 2.12-3.27 mequiv/g) and 2 (IEC ) 1.36-2.46 mequiv/g) gave infinite dilution mobility values of 2.9 ( 0.4 × 10-3 and 2.1 ( 0.2 cm2 s-1 V-1, respectively. These are in reasonable agreement with calculated values from our current study for the comparable membranes (2.6 ( 0.2 and 1.6 ( 0.3 cm2 s-1 V-1, respectively). The value for 3 differs from a previously reported value5 of ∼3.6 × 10-3 cm2 s-1 V-1 despite our previous results with SPEEK using the water-saturated membrane method providing very similar infinite dilution mobility values to those reported.5
2854 J. Phys. Chem. B, Vol. 112, No. 10, 2008 TABLE 1: Calculated Proton Mobility Values at Infinite Dilution (Xv ) 1.0) polymer
µ′H+ at Xv ) 1.0 (10-3 cm2 s-1 V-1)
1a 1b 1c avg. 1a-c 2a 2b 2c 2d 2e avg. 2a-e 3
2.96 ( 0.05 2.6 ( 0.2 2.13 ( 0.08 2.6 ( 0.2 1.43 ( 0.06 1.8 ( 0.1 1.8 ( 0.1 1.1 ( 0.1 1.2 ( 0.2 1.6 ( 0.3 2.32 ( 0.01
In contrast to our previous work, however, it is possible to examine the predicted infinite dilution mobility for each sample within a given polymer system rather than an average result for a series with different IEC values. This can be seen in the case of 2 for which there is a broad range of predicted infinite mobility values. As was described earlier, it is thought that closer juxtaposition of acid groups would lead to higher mobility. Thus, a higher acid content with presumed smaller distances between acid groups would result in higher mobility, assuming, that is, there are no other contributing factors (e.g., tortuosity effects). Therefore, the higher acid content of 2c versus 2a may possibly explain at least part of the predicted higher infinite dilution mobility of 2c. Similarly, some of the difference between 2b and 2c might be possibly explained by an acid proximity effect upon mobility. However, both 2a and 2b are predicted to have the same mobility at infinite dilution in spite of the greater acid content of 2b. This suggests that there may be offsetting differences in tortuosity between the different samples of 2, especially in the case of 2d-e, which have significantly lower predicted µ′H+ values at Xv ) 1.0 than even 2a (IEC ) 1.36 mequiv/g versus 2.20 and 2.46 mequiv/g for 2d and 2e, respectively). The synthetic method used to produce these materials (copolymerization of different trifluorostyrene-based vinyl monomers followed by selective sulfonation of specific moieties) may be in some way responsible for these differences in tortuosity. Copolymerization of monomers with potentially different reactivities may lead to blocks within the polymer backbone that are further magnified by the sulfonation. Possibly, the connections between hydrophilic domains are not wellformed until 2c. However, at higher IEC values, there may be another change in morphology that leads to an increase in effective tortuosity of the proton conduction pathways and hence a decrease in mobility. Previous studies on these materials have not revealed any strong evidence for significant microphase separation although greater ionic aggregation was observed in a small-angle X-ray scattering study on 2 for intermediate IEC samples versus higher IEC samples.7 Further study on this polymer system may provide information for the reasons for the observed mobility differences. For all the studied polymer systems, the predicted infinite dilution mobility values are significantly lower than the theoretical value for a free proton at infinite dilution of 3.6 × 10-3 cm2 s-1 V-1.49 In the case of 1, it was previously theorized that this may be due in part to both the bound SO3- groups restricting proton mobility as well as the potentially cross-linked nature of 1, leading to increased tortuosity and thus lower mobility.19 In the case of 3, proximity of acid groups may possibly be an issue at very high water contents, given its low IEC value and the fact that the sulfonate groups are tethered. There may also be some tortuosity in the conduction pathway that leads to the predicted lower value.
Peckham et al. Unexpectedly low values for 2 had been seen in our previous study. These results are interesting and are reproducible, but the reasons remain unclear. However, it is noteworthy that 2c with the highest predicted infinite dilution mobility also exhibits the highest conductivity within polymer system 2, both as a function of RH (this study) as well as on the basis of its IEC value.19 Similarly, 2d-e exhibit the lowest mobility and, were it not for their comparatively high acid concentration, would exhibit lower conductivity values than 2a with its significantly lower acid content. As previously discussed, this suggests that there are considerable morphological differences among the samples of this polymer system, thereby leading to differences in tortuosity. Prior morphological studies on these membranes do not appear to show striking morphological segregations.7 However, further studies on the structure-property relationships of 2, as well as on block and graft copolymers in which morphological differences have been demonstrated,13,50 may provide additional evidence for these differences and possible knowledge that could be applied to improve proton conduction pathways not only for 2 but other polymer systems as well. 3. Conclusion The relationships of proton conductivity to acid and water content for a series of statistically sulfonated PEMs (1-3) have been examined, wherein acid content was varied by using samples with different IEC values and whereby water content was adjusted by controlling the RH value (50-98%) of the membrane environment. It has been found that decreasing water content leads to decreasing levels of proton mobility and hence conductivity, presumably due to less screening of the tethered sulfonate groups and hence increasing proton localization. At low RH levels (
