Universal Fluorine‐Free Proton Exchange Polymers for High‐Performance and Durable Fuel Cells Operable Under Severe Conditions

ABSTRACT

Fluorine‐free proton exchange membranes (PEMs) could offer a promising alternative to perfluorosulfonic acid (PFSA) membranes, addressing challenges such as negative environmental impact, high cost, and limited performance under harsh conditions. However, fluorine‐free PEMs inherently suffer from low proton conductivity due to weaker acidity and poorly developed hydrophilic/hydrophobic phase‐separation, along with limited mechanical and chemical stability. Here, we report a novel and universal fluorine‐free PEMs featuring aliphatic chains in the backbone, which for the first time to dissolve all the issues mentioned, such as 2.6 times higher proton conductivity than PFSAs. Further physical reinforcement (SP‐PAC12₆₇‐QP‐4.5‐PE‐7) effectively extended their lifetime to 335.1 times of the original SP‐PAC12₆₇‐QP‐3.5, and exceeded 1.3 times of the state‐of‐the‐art fluorinated Nafion XL reinforced membrane, in the severely accelerated durability test. Furthermore, the fluorine‐free PEMs exhibited outstanding performance than Nafion XL, especially at elevated temperatures, proving their practical applicability as a promising alternative to the fluorinated counterparts.

Due to excellent physical and chemical stability, the novel designed fluorine‐free polymers with aliphatic‐chain backbones (SP‐PAC12‐QP) are regarded as promising and universal alternatives to perfluorosulfonic acid membranes. Their performances are further improved by physical reinforcement, where the fuel cell performance under high temperature and combined chemical and mechanical durability (109,580 cycles) of the produced SP‐PAC12₆₇‐QP‐4.5‐PE‐7 outperforms commercially stabilized Nafion XL.

1. Introduction

Proton exchange membranes (PEMs) with ion‐selectivity and gas‐impermeability play a critical role in energy‐conversion and energy‐storage devices such as fuel cells [1, 2, 3, 4], redox flow batteries [5, 6], and electrolyzers [7, 8]. With the advancement of those electrochemical devices, demand for high‐performance PEMs has significantly increased [9]. Over the past several decades, PFSA membranes such as Nafion and their physically and chemically reinforced counterparts have been widely utilized as the commercially available benchmark PEMs [10, 11, 12, 13]. However, the complex synthesis procedures, high cost, and relatively high gas permeability remain issues intrinsic to the perfluorinated polymer structures. Furthermore, PFSA membranes are included among the category of per‐ and polyfluoroalkyl substances (PFAS) and thus would be subject to PFAS restrictions due to their possible impact on human health and the environment [14, 15, 16]. Therefore, alternative PEM materials free from fluorine are in increasingly high demand.

Among a number of types of fluorine‐free PEMs, aromatic polymers with acidic functional groups have been most explored but have yet been unsuccessful in replacing PFSA membranes [17, 18, 19, 20]. In particular, the intrinsically lower acidity of fluorine‐free aromatic sulfonic acid groups, as well as the less pronounced hydrophilic/hydrophobic differences in the polymer components compared to PFSA PEMs, are major obstacles to achieving high proton conductivity over a wide range of humidity and temperature [21, 22, 23]. To address these issues, Holdcroft et al. have designed polyphenylene‐based PEMs (SPPP‐H⁺ and SPPN‐H⁺) for constructing interconnected ionic domains for efficient proton transport. The resulting membranes exhibited high proton conductivity of 222 mS cm⁻¹ (SPPN‐H⁺) and 106 mS cm⁻¹ (SPPP‐H⁺) at 30°C and 95% RH, and 8.65 mS cm⁻¹ (SPPP‐H⁺) at 30°C and 40% RH [24, 25]. Winey et al. studied polyethylene backbones bearing pendant phenylsulfonic acid groups and found that the p5PhSA (with phenylsulfonic acid on every fifth carbon and an IEC of 4.2 mequiv g⁻¹) exhibited strong nanophase separation [26, 27]. Consequently, it achieved 280 mS cm⁻¹ proton conductivity at 40°C and 95% RH, four times higher than that of Nafion 117. Lin et al. reported that extending the alkyl side chain length in the hydrophobic segment of poly(alkyl‐fluorene isatin) significantly enhanced hydrophobic/hydrophilic phase separation, leading to a 3.9‐fold increase in proton conductivity upon replacing methyl groups with hexyl groups [28]. We previously discovered in our sulfonated polyphenyelne PEMs that sequencing the hydrophobic phenylene groups with controlled meta/para composition led to the development of ionic domains and channels, contributing to high proton conductivity (22 mS cm⁻¹ at 80°C and 40% RH) [29, 30]. Those researches have highlighted the high potential of sulfonated polyphenylenes as alternative, fluorine‐free PEMs but also their technical challenges. For example, some substituents triggered chemical degradation. After immersion in Fenton's solution (ex situ chemical stability), a polyphenylene‑based membrane with CH₃ substituents (SPP‐BP‐CH₃) became fragile, whereas one with CF₃ substituents (SPP‐BP‐CF₃) remained intact and exhibited no detectable degradation [22]. During a 400 h in situ chemical stability test, SPPB‑H⁺ experienced a substantial drop in open‐circuit voltage (OCV) from 0.965 V down to 0.71 V [25]. Furthermore, due to the intrinsic rigidity of the phenylene main chains, those fluorine‐free polyphenylene PEMs shared the limited mechanical strength with small strain, e.g., 3.3% for SPP‐BP‐CH₃, 17.5% for SPPB‑H⁺, 16.3% for HF3DP1‐PS [28], compared to 391.0% for Nafion NRE 211 [31].

The concurrent achievement of chemical and mechanical stability in PEMs, whether fluorinated or fluorine‐free, has been a central challenge for ensuring the durable operation of PEMFCs under practical, variable conditions [31, 32, 33, 34]. This challenge largely stems from the complex and interrelated chemical and mechanical degradations during long‐term operation. Mechanical degradation of PEMs is induced by swelling‐shrinking cycles and/or prolonged compression and eventually increases gas crossover and accelerates chemical decomposition, which in turn aggravates mechanical fatigue and ultimately triggers membrane fragmentation. The U.S. Department of Energy (DOE) introduced a standard protocol (combined OCV holding and periodic wet/dry cycling test) to simultaneously estimate the chemical and mechanical durability of membrane electrode assemblies (MEAs), defining a challenging target of 20,000 cycles with less than 20% loss in OCV [35]. To the best of our knowledge, none of the existing fluorine‐free PEMs have reached this target. Physical reinforcement with porous substrates and chemical stabilization with antioxidative additives have proven effective in enhancing the total durability of aromatic PEMs, based on the lessons of modified PFSA membranes, including Nafion XL and Gore‐Select [36, 37, 38]. For example, our polyphenylene‐based PEM (SPP‐TFP‐4.0) survived only 1,173 cycles under the aforementioned combined test, while SPP‐TFP‐4.0‐PVDF reinforced with porous non‐woven polyvinylidene fluoride (PVDF) fabric achieved 148,870 cycles and was more durable than the Nafion XL reinforced membrane with a lifetime of 88,008 cycles [31]. However, SPP‐TFP‐4.0‐PVDF contained fluorine groups both in the ionomer and porous substrate.

There were two main reasons why we adopted some fluorine‐containing groups in the components. First, fluorinated polymers were more capable in thin film‐forming and the resulting membranes were intrinsically more ductile than fluorine‐free polymers [30]. The reported SPP‐TFPs [39] containing trifluoromethyl groups exhibited 155% of maximum strain, compared with SPP‐QP (68%). Due to the markedly enhanced flexibility of the parent sulfonated polyphenylene with tetrafluoro‐terphenylene groups (BSP‐TP‐f, maximum strain: 90%) compared to SPP‐QP under identical conditions, the resulting BSP‐TP‐f‐4.1‐PE membrane reinforced with porous polyethylene [40] offered wet‐dry cycling durability (40,673 cycles) far superior to that of SPP‐QP‐PE‐9 (3,850 cycles), despite both membranes being fabricated on the similar porous PE substrate and cast from the same DMAc solvent. Second, fluorinated polymers were more soluble in common organic solvents, including lower alcohols (e.g., ethanol and isopropanol) than fluorine‐free polymers, making a significant impact on the quality of the reinforcement. Compared with SPP‐QP (soluble only in high boiling solvents such as DMSO and DMAc), the low alcohol‐soluble SPP‐TFPs were easier to reinforce with many porous materials. In fact, the fluorine‐free reinforced SPP‐QP‐PE‐9 membrane [41], prepared from SPP‐QP ionomer in DMAc solution with a porous polyethylene (PE) substrate, exhibited a limited wet‐dry cycling durability of only 3,850 cycles, while reinforced SPP‐TPF‐4.0‐PVDF survived 148,870 cycles. Jang et al. investigated the effect of casting solvent on reinforced membranes. RCM‐E was fabricated by casting a sulfonated poly(arylene ether sulfone) (SPAES‐40) ionomer DMAc solution onto a porous polytetrafluoretyhylene (PTFE) substrate, whereas RCM‐O was prepared similarly but with ethanol pre‐wetting of the PTFE to improve interfacial compatibility [42]. Consequently, RCM‐O exhibited improved mechanical properties (maximum stress: 46.19 MPa; maximum strain: 85.46%) and higher maximum power density (725 mW cm⁻²), compared to RCM‐E (37.65 MPa, 53.23%; 633 mW cm⁻²).

In order to address the issues of fluorine‐free aromatic polymer‐based PEMs, we designed sulfonated polyphenylene co‐ and terpolymers incorporating alkylene groups in the main chain to promote hydrophobic/hydrophilic phase separation, lower gas permeability, and enhance membrane flexibility, with negligible impact on chemical durability. By tuning the methylene‐to‐phenylene ratio and alkylene chain length, highly proton‐conductive terpolymers were obtained that were well soluble in low‐boiling‐point solvents such as ethanol and isopropanol, allowing their use as a universal ionomer with various porous substrates. Using this novel polymer design, we report for the first time a completely fluorine‐free reinforced PEM that surpasses the commercial chemically stabilized and physically reinforced PFSA membrane, Nafion XL, in an accelerated chemical and mechanical durability test, while maintaining superior fuel cell performance at elevated temperatures. The structure–property relationships of the co‐ and terpolymers and the effects of porous substrates were systematically investigated, providing insights into the design of durable, high‐performance, fluorine‐free PEMs as potential alternatives to PFSA systems.

2. Results and Discussion

2.1. Synthesis and Characterization

The fluorine‐free monomers (PACx, x = 6, 10, 12) with different alkyl spacer lengths were successfully synthesized by Friedel‐Crafts reaction and ketone reduction (Scheme S1; Figure S1). By Ni(0)‐mediated coupling reaction, the SP‐PACx‐3.5 copolymers, SP‐PAC12‐QP‐3.5 and 4.5 terpolymers were synthesized as shown in Figure 1a. ¹H NMR spectra confirmed their chemical structures (Figure S2). As listed in Table 1, all co‐ and terpolymers exhibited high molecular weight (Mw ≥ 131.7 kDa) and solubility in DMSO, while only SP‐PAC12₆₇‐QP‐4.5, with high‐IEC, dissolved well in lower alcohols (e.g., isopropanol). Since the polymerization mixture became highly viscous as proceeding the polymerization, the reaction was not so efficient at the last stage of the polymerization that the PDI of the resulting co‐ and terpolymers was relatively high (PDI = 3.92–10.89). Use of a mechanical stirrer would enable a more efficient polymerization reaction, in particular, in a larger scale. By casting polymer−DMSO solution onto a flat glass plate, bendable, transparent membranes (thickness: 25–35 µm) were obtained.

FIGURE 1.

Synthesis routes and membrane images. (a) The synthesis routes of SP‐PACx‐3.5 (x = 6, 10, 12) copolymers and SP‐PAC12_y‐QP terpolymers (y indicates the molar composition of PAC12 in the hydrophobic components). (b) Images of fabricated SP‐PAC12₆₇‐QP‐4.5‐PVDF, SP‐PAC12₆₇‐QP‐4.5‐PE‐6, SP‐PAC12₆₇‐QP‐4.5‐PE‐7, and SP‐PAC12₆₇‐QP‐4.5‐PE‐13 reinforced membranes.

TABLE 1.

The composition, molecular weight, solubility, and IECs of polymers and reinforced membranes.

Membrane	Fluorine content (wt%)	Methylene /phenylene a	m:n:o	IEC (mequiv g−1)			Mn (kDa)	Mw (kDa)	Solubility e
				Target b	NMR c	Titrated d			DMSO/ DMAc	Alcohols
SP‐PAC6‐3.5	0	1.56	1.84:1:0	3.50	3.37	2.77	21.2	228.5	√	×
SP‐PAC10‐3.5	0	2.34	2.26:1:0	3.50	3.42	2.97	18.4	200.4	√	×
SP‐PAC12‐3.5	0	2.68	2.48:1:0	3.50	3.40	2.80	20.1	137.1	√	×
SP‐QP‐3.5	0	0	2.94:0:1	3.50	3.48	3.12	33.5	131.7	√	×
SP‐PAC1250‐QP‐3.5	0	0.59	5.42:1:1	3.50	3.31	3.18	20.2	141.0	√	°
SP‐PAC1267‐QP‐3.5	0	1.42	7.90:2:1	3.50	3.19	3.11	40.6	159.1	√	°
SP‐PAC1275‐QP‐3.5	0	1.68	10.35:3:1	3.50	3.15	3.07	24.6	169.5	√	°
SP‐PAC1267‐QP‐4.5	0	0.98	15.50:2:1	4.50	4.05	3.92	30.0	141.5	√	√
SP‐PAC1267‐QP‐4.5‐PVDF	7.9	—	15.50:2:1	3.40 f		3.08	—		—
SP‐PAC1267‐QP‐4.5‐PE‐6	0	—		3.41 f		2.91
SP‐PAC1267‐QP‐4.5‐PE‐7	0	—		2.91 f		2.61
SP‐PAC1267‐QP‐4.5‐PE‐13	0	—		3.22 f		2.83

^a The molar ratio of methylene and phenylene.

^b Target IEC calculated from feed ratio.

^c NMR IEC calculated from ¹H NMR spectra.

^d Titrated IEC measured from titration.

^e soluble (√), partly soluble (°) and insoluble (×).

^f IEC calculated from SEM images.

Prior to preparing reinforced membranes, the contact angles of various solvents on applied porous substrates, including porous nonwoven PVDF nanofiber fabric and PE substrates, were evaluated at room temperature (Figures S4 and S5a). Among all solvents, isopropanol showed the smallest contact angle: PVDF (4.9°) < PE‐6 (6.4°) < PE‐13 (7.6°) < PE‐7 (9.8°). The DLS measurement of the SP‐PAC12₆₇‐QP‐4.5 ionomer was carried out in isopropanol (Figure S5b), which exhibited a single narrow peak with an average hydrodynamic diameter of 166.3 ± 3.4 nm, indicating uniform dispersion of the ionomer and the absence of significant particle aggregation in isopropanol. Therefore, SP‐PAC12₆₇‐QP‐4.5−isopropanol solution was filled into these porous substrates via the push coating method [31]. After drying, transparent, flexible reinforced membranes with a thickness of only 13–22 µm were obtained (Figure 1b).

2.2. Morphology

The structural characteristics of the porous PVDF and PE substrates and the fabricated reinforced membranes were examined using a scanning electron microscope (SEM) (Figure 2a–d). The electrospun PVDF fabric exhibited randomly oriented nanofibers and uniform pore size with high porosity (78%). The expanded PE‐7 substrate showed a uniform porous structure with lower porosity (44%), while PE‐6 and PE‐13 substrates displayed higher porosity but less uniform pore size and distribution. All reinforced membranes were composed of a triple‐layer structure, composite middle layer (porous substrate filled with SP‐PAC12₆₇‐QP‐4.5 ionomer) sandwiched by upper and bottom ionomer layers (parent SP‐PAC12₆₇‐QP‐4.5). The three layers were flawless without any cracks, pores, or defects, revealing that the porous substrates were fully impregnated with the ionomer. The F and S element mapping images also supported triple layer structure with an ionomer‐impregnated middle layer. The S atoms belonging to SP‐PAC12₆₇‐QP‐4.5 ionomer were distributed homogeneously in the composite middle layer at a lower concentration than that in the ionomer‐rich upper and bottom layers. Meanwhile, F atoms were detected exclusively in the composite middle layer of the SP‐PAC12₆₇‐QP‐4.5‐PVDF reinforced membrane, confirming the triple‐layer composition. In addition, the thickness of the reinforced membranes (12.7, 16.7, 14.0, and 20.0 µm for SP‐PAC12₆₇‐QP‐4.5‐PVDF, ‐PE‐6, PE‐7, and PE‐13, respectively) observed was consistent with the measurements by a micrometer.

FIGURE 2.

The SEM images. (a–d) The surface SEM images of porous substrates were observed at 5,000 magnifications, where the upper right corner was observed at 30,000 magnifications. (e–h) Cross‐section morphologies of reinforced membranes. (i–l) Sulfur element mapping images of reinforced membranes. (m–p) Fluorine element mapping images of reinforced membranes.

Transmission electron microscopy (TEM) images (Figures S6 and S7, Table S1) provided qualitative evidence of phase separation in the membranes, where hydrophilic (dark) and hydrophobic (bright) domain sizes were averaged from >100 spots. Among all SP‐PACx‐3.5 and SP‐QP‐3.5 membranes, SP‐PAC12‐3.5 showed the largest hydrophilic (0.99 ± 0.16 nm) and smallest hydrophobic (1.28 ± 0.15 nm) domains, indicating the entanglement of longer aliphatic chains promoting sulfonic acid group aggregation and the formation of distinct hydrophilic/hydrophobic phase separation. Moreover, obvious phase separation was also observed in terpolymer SP‐PAC12₆₇‐QP‐3.5. However, various reinforced membranes exhibited comparable domain sizes, suggesting that the phase separation was not affected by the porous substrates.

2.3. Mechanical Properties, Thermal and Ex Situ Oxidative Stabilities

As shown in Figure 3a and Table S2, replacing the rigid QP segment with a flexible aliphatic chain effectively improved the maximum strain from 8.3% (SP‐QP‐3.5) to 105.9% (SP‐PAC12‐3.5), albeit at the expense of a stress reduction from 34.5 to 9.0 MPa. To overcome the trade‐off relation between rigidity and flexibility and achieve balanced mechanical properties, a series of SP‐PAC12_y‐QP‐3.5 terpolymers was designed via precise control of the methylene/phenylene ratio. The obtained SP‐PAC12₆₇‐QP‐3.5 with a methylene/phenylene ratio of 1.42 exhibited the largest strain (104.7%) with a reasonable maximum stress (20.1 MPa) outperforming the two‐component copolymer membranes.

FIGURE 3.

Mechanical properties. (a) Stress/strain curves of SP‐PACx‐3.5 (x = 6, 10 and 12), SP‐QP‐3.5 and SP‐PAC12_y‐QP‐3.5 (y = 50, 67 and 75) at 80°C and 60% RH. Stress/strain curves of SP‐PAC12₆₇‐QP‐4.5‐filled reinforced membranes at 80°C under (b) 60% RH and (c) 0% RH. (d) The rupture energy of SP‐PAC12₆₇‐QP‐3.5, SP‐PAC12₆₇‐QP‐4.5‐filled reinforced membranes, Nafion NRE 211, and Nafion XL.

The effect of porous substrates on the mechanical properties was evaluated at 80°C under both high (60%) and low (0%) RH (Figure 3b,c; Table S3). The stress/strain curves of the isotropic PVDF fabric and anisotropic PE substrates in the machine direction (MD) and transverse direction (TD) are provided as a reference (Figure S8). Owing to the flexible porous substrates and ionomer, all reinforced membranes exhibited increased maximum strain (≥ 116.4%). It was apparent that the employed porous substrates strongly governed the mechanical strength of the reinforced membranes. For instance, SP‐PAC12₆₇‐QP‐4.5‐PE‐7 fabricated from a robust PE‐7 substrate, exhibited the highest stress (42.3 MPa in MD and 23.7 MPa in TD), while SP‐PAC12₆₇‐QP‐4.5‐PE‐13 exhibited the lowest stress (8.8 MPa in TD) because of the intrinsic weakness of the PE‐13 substrate (7.7 MPa in TD). At 0% RH, the reduced plasticizing effect resulted in lower strain for the membranes, but SP‐PAC12₆₇‐QP‐PE‐7 still possessed high strain, up to 96.9% and 102.7% in MD and TD, respectively. Furthermore, among all membranes, SP‐PAC12₆₇‐QP‐PE‐7 exhibited the highest rupture energies (calculated from the stress/strain curves) under both conditions, achieving a 4.8–5.3‐fold and 5.9–7.3‐fold improvement over the original SP‐PAC12₆₇‐QP‐3.5 at 0% and 60% RH, respectively. In particular, under 60% RH, the rupture energy of SP‐PAC12₆₇‐QP‐4.5‐PE‐7 reached 100.1 and 123.4 MJ m⁻³ in TD and MD, respectively, approximately twice those of Nafion XL. The lowest change in rupture energy (16.3%) of isotropic SP‐PAC12₆₇‐QP‐4.5‐PVDF upon humidity increase from 0% to 60% RH reflected the minor humidity impact on mechanical stability [31], although the rupture energies, 12.9 (0% RH) and 15.0 MJ m⁻³ (60% RH) were relatively low.

The humidity and temperature dependence of viscoelastic properties was evaluated at constant temperature (80°C) and different humidity (60% and 10% RH) (Figure S9). The storage modulus (E′) of SP‐PACx‐3.5 membranes decreased with increasing the aliphatic chain. However, the decrease was suppressed by introducing aromatic QP monomer into SP‐PAC12‐3.5, where the variation of E′ was similar to that of SP‐PAC6‐3.5. It was probably associated with a comparable methylene/phenylene ratio, 1.42 for SP‐PAC12₆₇‐QP‐3.5 and 1.56 for SP‐PAC6‐3.5. In the temperature dependence at 60% RH (Figure S9b), SP‐PAC10‐3.5 and SP‐PAC12‐3.5 exhibited broad peaks in loss modulus (E′′), which was attributable to the glass transition temperature caused by the micro‐Brownian motion of the long aliphatic chain groups. With the decreasing methylene/phenylene ratio, SP‐PAC6‐3.5 and SP‐PAC12₆₇‐QP‐3.5 did not exhibit obvious glass transition behaviors. The viscoelastic properties of all membranes were more stable and exhibited no glass transition in the temperature dependence at 10% RH (Figure S9c), implying that the thermal stability of the membranes, including the aliphatic chain in the backbone was highly dependent on absorbed water. The reinforced membranes were also evaluated under the same conditions (Figure S10). At 80°C, as the increasing the humidity, a similar change in E′ and E′′ to SP‐PAC12₆₇‐QP‐3.5 was observed for reinforced membranes. At 60% RH, all reinforced membranes showed similar broad peaks in E′′ between 40 and 80°C, whereas no such peaks were detected in SP‐PAC12₆₇‐QP‐3.5 membrane. The E′′ value of reinforced membranes remained stable at low (e.g., 0.0626 Gpa for SP‐PAC12₆₇‐QP‐4.5‐PE‐7‐TD, 27°C) and high temperatures (e.g., 0.0547 GPa for SP‐PAC12₆₇‐QP‐4.5‐PE‐7‐TD, 110°C). Therefore, the glass transition in the reinforced membranes was considered to originate from the porous substrates and did not alter in the reinforced membranes.

The effect of aliphatic chains in the backbone on thermal stability of the proton exchange membranes was investigated via thermogravimetric analyses (TGA) under N₂ atmosphere within the temperature range of room temperature to 800°C to evaluate their applicability to medium‐ and high‐temperature fuel cells (Figure S11). The weight loss (ca. 12–15 wt.%) from room temperature to ca. 150°C was due to the evaporation of water and residual solvent. The second weight loss between ca. 250 and 400°C was attributable to the decomposition of the sulfonic acid groups. The following third stage, occurring from ca. 450°C was observed in the polymers containing aliphatic chains in the backbone (SP‐PACx‐3.5 and SP‐PAC12₆₇‐QP‐3.5), but was absent in SP‐QP‐3.5, suggesting that the weight loss was the result of the decomposition of the aliphatic chains. The last stage above ca. 510°C was attributable to the decomposition of phenyl groups in the backbone. TGA results demonstrated that the membranes containing alkyl spacers possessed sufficiently high thermal stability to match the operational requirement in medium and high‐temperature PEMFCs (≤ 180°C).

The SP‐PAC12₆₇‐QP‐3.5 was subjected to Fenton's reagent at 60°C for 1 h to scrutinize its ex situ oxidative stability. The post‐test SP‐PAC12₆₇‐QP‐3.5 retained its flexibility and solubility in high‐boiling solvent (e.g., DMSO) with only 4.7% weight loss. For further analysis of the degradation behavior, the molecular weight and chemical structure of the post‐test membrane were characterized (Figure S12). Compared with the original SP‐PAC12₆₇‐QP‐3.5, the post‐test polymer showed a slightly decreased number‐average molecular weight (M _n) from 40.6 to 32.2 kDa and an increased weight‐average molecular weight (M _w) from 159.1 to 237.5 kDa, suggesting that minor degradation occurred. In addition, the nearly identical NMR spectra before and after the Fenton's test, together with 93.7% remaining IEC calculated from the NMR spectrum, indicate that the sulfonic acid groups, rather than the hydrophobic segments, were partly lost during the Fenton's test. Overall, the SP‐PAC12₆₇‐QP‐3.5 demonstrated excellent chemical stability against radical attack generated from the Fenton's reagent.

2.4. Simulation, Water Uptake and Proton Conductivity

To understand the effect of molecular structures on the packing efficiency of polymer chains, molecular dynamics simulations were performed on model polymers with rigid phenylene or flexible aliphatic linkages (Figure 4a–c). The fully phenylene‐bridged SP‐QP exhibited the largest microporous cavities due to the rigidity and steric hindrance of the consecutive phenylene groups, which restricted chain rearrangement and impeded their dense packing. In contrast, incorporation of a flexible aliphatic chain into the main chain (SP‐PAC12‐QP and SP‐PAC12) significantly enhanced conformational flexibility of the polymer chain, enabling tighter packing and smaller interstitial voids. Pore size distribution analysis using Zeo++ with a 0.2 nm probe diameter (Figure 4d) revealed that SP‐PAC12 possessed the smallest and narrowest pore size distribution, followed by SP‐PAC12‐QP and SP‐QP. This enhanced packing efficiency of the aliphatic‐containing polymers (e.g., SP‐PACx‐3.5) was corroborated by their significantly lower water uptake (<72.0% across all humidity conditions), compared to SP‐QP‐3.5 (113.9% at 95% RH) (Figure 4e). Interestingly, SP‐PAC12₆₇‐QP‐3.5, with a balanced composition of flexible and rigid segments, retained low water uptake (under ≤ 80% RH), similar to SP‐PAC12‐3.5, yet preserved the dimensional robustness of the rigid SP‐QP framework (Figure S13). Additionally, the narrow pore size distribution contributed to the formation of efficient ion transport pathways, which accounted for the high proton mobility (µ_H+) calculated from the Nernst‐Einstein equation (Figure S14a). Consequently, SP‐PACx‐3.5 outperformed SP‐QP‐3.5 in proton conductivity under > 20% RH (e.g., 50.4 mS cm⁻¹ vs. 33.6 mS cm⁻¹ at 60% RH, Figure 4f). Impressively, the structurally optimized SP‐PAC12₆₇‐QP‐3.5 achieved the highest conductivity (479.8 mS cm⁻¹ at 95% RH) across the full humidity range, highlighting the synergistic effect of the combined flexible‐rigid segmental design. Moreover, all newly designed fluorine‐free membranes exceeded the commercial Nafion NRE 211 (187.9 mS cm⁻¹ under 95% RH) membrane under test conditions.

FIGURE 4.

Properties of polymers and membranes. (a–c) Computational modeling results of the pore surfaces for SP‐PAC12, SP‐PAC12‐QP and SP‐QP (red, the equilibrated polymer; blue, microporous cavities) at 353 K, applying a 0.2 nm probe diameter. (d) The pore size distribution obtained by molecular dynamic simulation. (e) Water uptake and (f) proton conductivity as a function of relative humidity for SP‐PACx‐3.5, SP‐PAC12_y‐QP‐3.5, SP‐QP‐3.5, Nafion NRE 211, and Nafion XL membranes. (g) Water uptake and h) proton conductivity as a function of relative humidity for SP‐PAC12₆₇‐QP‐4.5‐filled reinforced membranes. (i–l) Conductive durability of SP‐PAC12₆₇‐QP‐4.5‐filled reinforced membranes at 80°C and 90% RH for 100 h.

The lower water uptake and in‐plane swelling ratio in SP‐PAC12₆₇‐QP‐4.5‐filled reinforced membranes than those in SP‐PAC12₆₇‐QP‐3.5 (Figure 4g; Figure S13) reflected the high capability of porous substrates to suppress water absorption, particularly, at > 80% RH. However, the similar humidity dependence of water uptake at 80°C in four reinforced membranes suggested that the used ionomer dominated the water absorption properties. In Figure 4h, SP‐PAC12₆₇‐QP‐4.5‐PVDF and SP‐PAC12₆₇‐QP‐4.5‐PE‐6 using highly porous (65%–78% porosity) and thin (6–7 µm) substrates exhibited significantly higher proton conductivity (724.6 and 550.7 mS cm^{− 1}, respectively at 95% RH) than those of SP‐PAC12₆₇‐QP‐4.5‐PE‐13 (363.6 mS cm^{− 1}) and SP‐PAC12₆₇‐QP‐4.5‐PE‐7 (291.6 mS cm^{− 1}) using thick (13 µm thick of PE‐13) and low porosity (44% for PE‐7) substrates. The proton conductivity of the reinforced membranes at elevated temperatures (100 and 120°C, Figure S15) retained the same order as that at 80°C. Among them, SP‐PAC12₆₇‐QP‐4.5‐PE‐7 exhibited the lowest proton conductivity, but all values exceeded those of Nafion XL under all test conditions. For a more quantitative comparison, the molar proton conductivities (σ/IEC_V), normalized by the volumetric IEC, are plotted as a function of λ at 80°C (Figure S14b) to eliminate the influence of the effect of IECs among the copolymers, terpolymers, and reinforced membranes. The similarity in molar proton conductivity at comparable λ indicates that those membranes shared a similar proton transport mechanism to Nafion NRE 211 and Nafion XL.

Considering the high‐IEC (4.5 mequiv g⁻¹) of SP‐PAC12₆₇‐QP‐4.5 ionomer, the stability of the proton conductivity for all SP‐PAC12₆₇‐QP‐4.5‐filled reinforced membranes was monitored in a high‐humidity environment (90% RH) at 80°C for 100 h (Figure 4i–l). Despite minor fluctuations caused by humidity instability, the proton conductivity remained highly stable, confirming the chemical and physical durability of the high‐IEC ionomer under prolonged high‐humidity conditions.

2.5. Fuel Cell Performance and In Situ Chemical Stability

The original SP‐PAC12₆₇‐QP‐3.5 membrane and all SP‐PAC12₆₇‐QP‐4.5‐filled reinforced membranes were fabricated into catalyst‐coated membranes (CCMs) (Figure S16) and subsequently assembled into MEAs for evaluation of their fuel cell performance and in situ stability, supplying H₂ to the anode and air to the cathode without back‐pressure. In the signal cells, all components were the same except for the membrane. The initial electrochemically active surface area (ECSA) of the cathodes was estimated by cyclic voltammetry (CV) at 40 °C and 100% RH (Figure S17 and Table S4). With the same Pt loading amount (0.5 mg cm⁻² for both electrodes), all cells exhibited similar ECSA (75.8−81.8 m² g_pt ⁻¹), comparable to that of Nafion NRE 211 (83.0 m² g_pt ⁻¹) and Nafion XL (81.9 m² g_pt ⁻¹), suggesting that the cathode catalyst layers were not affected by the membranes. After CV, the linear sweep voltammetry (LSV) was conducted to estimate the hydrogen permeability of the membranes at 80 °C and 100% RH (Figure 5a; Table S4). Compared with Nafion NRE 211 (1.27 mA cm⁻², 25 µm thick) and Nafion XL (1.07 mA cm⁻², 30 µm thick), the SP‐PAC12₆₇‐QP‐3.5 membrane (27 µm thick) exhibited much smaller hydrogen cross‐over current density (0.25 mA cm⁻²) due to the densely packed, fluorine‐free polymer structure. The thinner SP‐PAC12₆₇‐QP‐4.5‐filled reinforced membranes (13‐20 µm thick) also exhibited significantly lower hydrogen crossover current densities (0.26–0.37 mA cm⁻²), indicating that the substrate pores were effectively filled with the SP‐PAC12₆₇‐QP‐4.5 ionomer and that the intrinsically gas impermeable properties of the ionomer functioned in the substrate pores. Considering the different thicknesses of the applied membranes [43], the fabricated SP‐PAC12₆₇‐QP‐4.5‐filled reinforced membranes showed lower hydrogen permeability than the pristine SP‐PAC12₆₇‐QP‐3.5 (Table S4). Low hydrogen permeability was also supported by the high OCVs (> 0.96 V) during the subsequent fuel cell performance measurements (Figures S18 and S19).

FIGURE 5.

Linear sweep voltammograms (LSVs) and power densities. (a) LSV at 80°C and 100% RH, feeding H₂ (0.1 slpm) and N₂ (0.1 slpm) at anode and cathode, respectively. Power densities at (b) 100°C, 53% RH, (c) 100°C, 30% RH and (d) 120°C, 30% RH, feeding H₂/air at anode/cathode. The measured membranes were: SP‐PAC12₆₇‐QP‐3.5, SP‐PAC12₆₇‐QP‐4.5‐filled reinforced membranes, Nafion NRE 211, and Nafion XL. Test conditions: 0.5 mg cm⁻² Pt loading on both electrodes, without back‐pressure.

At 80°C and 100% RH, benefiting from high proton conductivity, the fabricated SP‐PAC12₆₇‐QP‐4.5‐filled reinforced membranes exhibited lower ohmic resistances (0.031‐0.075 Ω cm²) than those of Nafion NRE 211 (0.076 Ω cm²) and Nafion XL (0.110 Ω cm²). Among the fabricated membranes, the SP‐PAC12₆₇‐QP‐4.5‐PE‐7 exhibited the highest maximum power density (560.5 mW cm⁻²), due to its higher mass activity (89.5 A g_pt ⁻¹). The mass activity of the Pt catalyst in the cathode was calculated from the IR‐free IV curves (Figure S20), assuming negligible anodic overpotential at low current densities. The different mass activities (Nafion NRE211> Nafion XL> SP‐PAC12₆₇‐QP‐4.5‐PE‐7> SP‐PAC12₆₇‐QP‐4.5‐PE‐6> SP‐PAC12₆₇‐QP‐3.5> SP‐PAC12₆₇‐QP‐4.5‐PVDF> SP‐PAC12₆₇‐QP‐4.5‐PE‐13) in those cells reflected the interfacial compatibility of the membranes with the catalyst layers, demonstrating that the membranes can affect the apparent catalytic activity. It also explained the high performance of Nafion NRE 211 and Nafion XL in comparison with the others. However, with decreasing humidity and increasing temperature (up to 120°C), their mass activities decreased by more than 58%, presumably due to the intrinsic low glass transition temperature of Nafion (Tg, ≈100 °C). Operating near or above Tg (e.g., 120 °C and 30% RH) induced micro‐Brownian motion in the polymer main chains, which eventually collapsed the microphase‐separated morphology with ionic pathways and impaired proton transport [11, 44]. The decrease in water content at elevated temperatures further exacerbated the performance degradation. In contrast, the reinforced membranes filled by thermally stable and highly conductive SP‐PAC12₆₇‐QP‐4.5 ionomer, SP‐PAC12₆₇‐QP‐4.5‐PVDF (403.1 mW cm⁻²), ‐PE‐7 (349.3 mW cm⁻²), and ‐PE‐6 (322.0 mW cm⁻²) performed much better than Nafion NRE 211 (301.2 mW cm⁻²) and Nafion XL (308.3 mW cm⁻²) at 100°C and 53% RH (Figure 5b). At even lower humidity (30% RH), SP‐PAC12₆₇‐QP‐4.5‐PVDF exhibited even more pronounced merit in output maximum power density (Figure 5c,d). For instance, at 120 °C and 30% RH, the maximum power density of SP‐PAC12₆₇‐QP‐4.5‐PVDF membranes was 264.6 mW cm⁻², substantially superior to the 162.3 and 166.9 mW cm⁻² values for Nafion XL and Nafion NRE 211, respectively. In addition, the SP‐PAC12₆₇‐QP‐4.5‐PE‐6, and ‐PE‐7 also consistently delivered higher power densities than that of Nafion XL at elevated temperatures (≥100 °C), significantly highlighting the outstanding suitability of the fluorine‐free SP‐PAC12₆₇‐QP‐4.5‐filled reinforced membrane at high temperature.

After measuring the polarization curves, the membrane was subsequently subjected to the in situ chemical stability test at a constant current density (0.2 A cm⁻²) at 120°C and 30% RH for 300 h (Figure S21). It is known that •OH and •OOH radicals, mainly generated from the reaction between hydrogen and crossed‐over oxygen, can oxidatively attack membranes, causing chemical degradation and a decrease in cell voltage. In addition to the polymer backbones, due to the limited dissociation of sulfonic acid groups (–SO₃H) at elevated temperature and low humidity, the undissociated –SO₃H groups are susceptible to hydrogen abstraction by the radicals [45]. The chemical stability of the cells was significantly improved by applying suitable reinforced membranes compared to the original SP‐PAC12₆₇‐QP‐3.5 membrane (220.0 µV h⁻¹), where the lowest decay rate of the cell voltage was achieved with SP‐PAC12₆₇‐QP‐4.5‐PVDF cell (136.6 µV h⁻¹), followed by SP‐PAC12₆₇‐QP‐4.5‐PE‐7 (193.3 µV h⁻¹) and SP‐PAC12₆₇‐QP‐4.5‐PE‐6 (196.7 µV h⁻¹). To understand the degradation states, the changes of chemical structure and molecular weight of post‐test SP‐PAC12₆₇‐QP‐3.5 CCM were characterized by NMR spectroscopy and GPC (Figure S22a,b). The polymer in the post‐test CCM was less soluble in DMSO‐d₆ (for NMR spectroscopy characterization) and DMF containing 0.01 g L⁻¹ LiBr (as the eluent for GPC test) than the pristine polymer, probably due to the occurrence of minor crosslinking during the degradation [46, 47]. Compared with the pristine sample, no new peaks appeared, and the peaks corresponding to the hydrophobic components remained at similar intensities. However, the proton peak (Ha) adjacent to the ─SO₃H groups became slightly smaller. The IEC_NMR calculated from the NMR spectrum of the post‐test membrane was 13% smaller. In addition, after in situ chemical stability test, the M _n decreased, but the M _w increased. Based on those results and the previous reports [46, 47, 48, 49], the possible degradation mechanisms are proposed (Figure S22c). Due to the attack of generated radicals, the C−S bond was broken, resulting in the formation of phenol; 2) the C−C linkage connecting the sulfonated phenyl ring and phenyl ring was decomposed to hydroxylated, sulfonated phenyl ring chain scission and phenyl radical, leading to crosslinking by recombination reactions. Although the fluorine‐free SP‐PAC12₆₇‐QP‐3.5 membranes exhibited a higher ohmic resistance decay rate (601.2 µΩ cm² h^{− 1}) compared to commercial fluorinated membranes such as Nafion NRE 211 (69.1 µΩ cm² h^{− 1}) and Nafion XL (158.8 µΩ cm² h^{− 1}), the stability was significantly improved through mechanical reinforcement. The decays were 427.8 µΩ cm² h⁻¹ for SP‐PAC12₆₇‐QP‐4.5‐PE‐6 (from 0.393 to 0.521 Ω cm²), 454.2 µΩ cm² h⁻¹ for SP‐PAC12₆₇‐QP‐4.5‐PE‐7 (from 0.540 to 0.0.676 Ω cm²), 464.5 µΩ cm² h⁻¹ for SP‐PAC12₆₇‐QP‐4.5‐PVDF (from 0.202 to 0.341 Ω cm²), and 590.9 µΩ cm² h⁻¹ for SP‐PAC12₆₇‐QP‐4.5‐PE‐6 (from 0.557 to 0.734 Ω cm²), reflecting enhanced dimensional and oxidative stability.

2.6. Combined Chemical and Mechanical Durability Test at 90°C

To evaluate the resistance of the fluorine‐free SP‐PAC12₆₇‐QP‐4.5‐filled reinforced membranes (13–19 µm) to concurrent mechanical and chemical stresses encountered during the practical operation of fuel cell vehicles, we followed the aforementioned DOE protocol [35] of the combined chemical (OCV holding) and mechanical (periodic wet (15s) /dry (2s) cycles) durability test at 90°C, in which > 2.5‐fold ohmic resistance difference was maintained between wet and dry states (Figure 6a–c; Figures S23 and S24, and Table S5). All cells exhibited a dry/wet ohmic resistance ratio of 2.6–6.6 times and similar initial OCVs of 0.90‐0.92 V (wet) and 0.81‐0.89 V (dry). After the durability test, the OCV in the wet/dry states was decreased by 20.7%/40.4%, 19.6%/28.4%, 15.6%/29.6%, 6.6%/58.1%, 15.2%/45.9%, 16.5%/29.1% and 5.5%/64.0% for SP‐PAC12₆₇‐QP‐3.5, SP‐PAC12₆₇‐QP‐4.5‐PE‐6, SP‐PAC12₆₇‐QP‐4.5‐PE‐13, SP‐PAC12₆₇‐QP‐4.5‐PVDF, SP‐PAC12₆₇‐QP‐4.5‐PE‐7, Nafion NRE 211 and Nafion XL, respectively. Notably, under the dry conditions, the OCV losses of all fabricated membranes were lower than Nafion XL. Compared with the original SP‐PAC12₆₇‐QP‐3.5 membrane (327 cycles), physically reinforced membranes exhibited longer lifetimes by factors of 5.0–335.1. In particular, SP‐PAC12₆₇‐QP‐4.5‐PVDF (54,574 cycles) and SP‐PAC12₆₇‐QP‐4.5‐PE‐7 (109,580 cycles) exceeded the DOE target by factors of 2.7 and 5.5, respectively. In particular, the fluorine‐free SP‐PAC12₆₇‐QP‐4.5‐PE‐7 demonstrated the longest lifetime, exceeding that of Nafion XL by a factor of 1.3.

FIGURE 6.

Combined chemical (OCV holding) and mechanical (periodic wet/dry cycling) durability test. OCV vs. the cycle number and test time of (a) SP‐PAC12₆₇‐QP‐4.5‐PVDF and (b) SP‐PAC12₆₇‐QP‐4.5‐PE‐7. (c) The comparison of cycle number among all membranes; Spider graphs of (d) SP‐PAC12₆₇‐QP‐4.5‐PVDF and e) SP‐PAC12₆₇‐QP‐4.5‐PE‐7 summarize possible factors impacting the combined chemical and mechanical durability, including rupture energy in MD and TD at 60% RH, change rate of rupture energy in MD and TD, the ratio of MD/TD (calculated from the rupture energy in MD and TD), through‐plane swelling and in situ chemical stability. Test condition: anode/cathode: 0.2/0.1 mg cm⁻² Pt loading, wet/dry time:15/2 s, at 90°C without backpressure, supplying H₂ (0.06 slpm)/air (0.06 slpm) at anode/cathode.

After the test, the recovered CCMs retained their integrity without any visible defects except for the original SP‐PAC12₆₇‐QP‐3.5 and Nafion NRE 211 [31] (Figure S25). An obvious crack and pinholes in the post‐test SP‐PAC12₆₇‐QP‐3.5 and Nafion NRE 211 indicated that the termination of the test was associated with this mechanical failure. For further analysis, the SEM images in surface and cross‐sections of recovered CCMs were obtained (Figure S26). In contrast to only minor cracks in SP‐PAC12₆₇‐QP‐4.5‐PE‐6 and ‐PE‐13 CCMs, several cracks were observed in the post‐test SP‐PAC12₆₇‐QP‐4.5‐PVDF, ‐PE‐7, and Nafion XL [31] CCMs, reflecting the impact of the prolonged accelerated stress (< 15 h for the former vs. >250 h for the latter) on catalyst layer degradation. The cross‐sectional SEM images of SP‐PAC12₆₇‐QP‐4.5‐PE‐6, ‐PE‐13, and PVDF revealed cracks in the ionomer layer on the cathode side [34], which likely resulted from the ionomer tearing during the combined durability test. These results indicate that the failure can be most likely associated with insufficient mechanical stability (low and nonuniform rupture energies under wet and dry conditions) of these membranes. This conclusion was supported by the minimal thickness decrease (only 0.3–0.7 µm) of these membranes after the testing, suggesting that chemical degradation of membranes was not the major cause of failure. In contrast, the cross‐sectional image of SP‐PAC12₆₇‐QP‐4.5‐PE‐7 after 109,580 cycles exhibited both cracks and ionomer loss in the ionomer layer on the cathode side. These structural defects impaired the interfacial compatibility between the membrane and the cathode catalyst layer, thereby resulting in increased ohmic resistance of this cell, particularly under dry conditions. Additionally, the membrane thickness significantly decreased from an initial 15 µm to a final 10 µm. These observations suggest that the concurrence of chemical (SP‐PAC12₆₇‐QP‐4.5 ionomer and Nafion binder) and mechanical degradation in the SP‐PAC12₆₇‐QP‐4.5‐PE‐7 membrane accelerated the mechanical failure during the durability test. As discussed in our previous report [31], the thinning of the Nafion XL membrane (from 30 to 27 µm) and many large cracks in the catalyst layer suggested that chemical decomposition in the Nafion binder and membrane caused the termination.

According to the above results and discussion, we proposed several possible causative factors on combined chemical and mechanical durability, considering rupture energy values and changes occurring under various conditions, including humidity, uniformity of materials, dimensional stability, and chemical stability (Figure 6d,e; Figure S27). It was discovered that no single factor but the total of all factors (or colored area in the spider plot) contributed to the lifetime, with a larger area corresponding to a longer lifetime. The most durable SP‐PAC12₆₇‐QP‐4.5‐PE‐7 membrane exhibited the largest colored area (1.14 times larger than that of Nafion XL), primarily benefiting from its well‐balanced properties, including the largest rupture energy values and the best dimensional stability. Nafion XL followed closely, mainly attributed to the lowest change rate of rupture energy across a wide humidity range. Owing to the isotropicity of the substrate, the best chemical stability and the slightly higher change rate of rupture energy, SP‐PAC12₆₇‐QP‐4.5‐PVDF achieved a comparable area with Nafion XL (99.5% that of Nafion XL) and the third best combined durability. These findings demonstrate that extending the lifetime under the combined mechanical and chemical degradation requires reinforced membranes to possess superior mechanical (including dimensional) and chemical stability, achieved by designing highly proton‐conductive, densely packed, and lower‐alcohol‐soluble ionomers and by fabricating thin, uniform (isotropic) porous materials.

3. Conclusion

In summary, we designed and synthesized a series of fluorine‐free copolymers (SP‐PACx‐3.5) featuring alkyl chain‐containing backbones and further developed a mechanically robust terpolymer (SP‐PAC12₆₇‐QP‐3.5), balancing the effects of flexible and rigid segments. The terpolymers contained densely packed structures and a narrow pore size distribution, which underpin the high proton conductivity and low gas permeability. The high solubility of high‐IEC SP‐PAC12₆₇‐QP‐4.5 polymer in lower alcohols enabled its effective impregnation into various porous substrates (nonwoven PVDF fabric and expanded PE‐6, ‐7, and ‐13) to fabricate physically reinforced membranes for enhancing the durability of fluorine‐free PEMs under practical, variable conditions. The resulting SP‐PAC12₆₇‐QP‐4.5‐PVDF (54,574 cycles) and ‐PE‐7 (109,580 cycles) reinforced membranes both met the DOE target in the harsh combined periodic wet/dry cycling and OCV‐hold durability test. In particular, the fluorine‐free SP‐PAC12₆₇‐QP‐4.5‐PE‐7 demonstrated unprecedented combined durability among both unreinforced and reinforced fluorine‐free membranes, and even outperformed the chemically and mechanically stabilized perfluorinated Nafion XL reinforced membrane (109580 cycles vs. 88,008 cycles). In addition, compared with Nafion XL, the fluorine‐free reinforced membranes exhibited 2.1–5.2 times higher proton conductivity, as well as superior H₂/air fuel cell performance under low humidity and elevated temperatures without backpressure. These findings demonstrate the promise of fluorine‐free SP‐PAC12₆₇‐QP polymers as viable and sustainable candidates for high‐performance and durable PEMs in next‐generation, environmentally benign fuel cells. Their good solubility in lower alcohols, low gas permeability, high ionic conductivity, and robust mechanical strength establish a versatile foundation to support future integration with various porous substrates, further accelerating the deployment of fluorine‐free technologies.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: adma72392‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.