Lin Guoa,
Akihiro Masudab and
Kenji Miyatake*cde
aIntegrated Graduate School of Medicine, Engineering and Agricultural Science, University of Yamanashi, Kofu, Yamanashi 400-8510, Japan
bToray Research Center, Inc., Otsu 520-8567, Japan
cClean Energy Research Center, University of Yamanashi, Kofu, Yamanashi 400-8510, Japan. E-mail: miyatake@yamanashi.ac.jp
dFuel Cell Nanomaterials Center, University of Yamanashi, Kofu, Yamanashi 400-8510, Japan
eDepartment of Applied Chemistry, and Research Institute for Science and Engineering, Waseda University, Tokyo 169-8555, Japan
First published on 11th April 2023
The mechanical and chemical durability is one of the most crucial properties for proton exchange membranes in practical fuel cell applications. In the present paper, we report the physical reinforcement of chemically stable, highly proton conductive tandemly sulfonated, partially fluorinated polyphenylenes using porous polyethylene (PE). With the PE pores completely and homogeneously filled by ionomers through a push coating approach, the resulting reinforced membranes were more proton conductive (183.1–389.2 mS cm−1) than the commercial perfluorinated ionomer (Nafion: 120.6–187.2 mS cm−1) membrane at high humidity (80–95% RH). Benefiting from the tough PE supporting layer, the reinforced membranes outperformed the parent ionomer membranes in stretchability with maximum strain up to 453%. The combination of intrinsic chemical stability of partially fluorinated polyphenylene ionomers and physical reinforcement with PE substrates contributed for the reinforced membranes to achieving superior durability to survive more than 20000 cycles in severe accelerated durability test combining OCV hold and wet/dry frequent cycling.
Reinforcing with more mechanically robust substrates is an effective and promising approach to improve the mechanical robustness of PEMs without scarifying inherent advantages of parent ionomers. In fact, PFSA membranes (e.g., GORE-SELECT and Nafion XL membranes) being used in commercial PEMFCs are reinforced with expanded polytetrafluoroethylene (ePTFE).15,16 We have also reported reinforced SPP-QP membranes with porous polyethylene (PE) substrates.17 The highly conductive SPP-QP ionomer was effectively impregnated into the nanopores of PE via push-coating method. The reinforced SPP-QP-PE membranes exhibited improved mechanical properties (maximum stress = 28–47 MPa, maximum strain = 106–134%, compared to 34 MPa maximum stress and 68% maximum strain of the parent SPP-QP membrane at 80 °C and 60% RH). In operando fuel cells, the reinforced membrane was durable for 3850 cycles in wet/dry cycle test (in nitrogen). Pintauro et al. fabricated a composite membrane (cPPSA-ePTFE) from sulfonated polyphenylene copolymer (cPPSA) and ePTFE.18 The cPPSA-ePTFE membrane exhibited improved mechanical properties (maximum stress = 25 MPa (machine direction) and 21 MPa (transverse direction), maximum strain = 10% (machine direction) and 18% (transverse direction), at 25 °C and 50% RH) compared to the parent cPPSA membrane. The fuel cell performance of cPPSA-ePTFE membrane (maximum power density = 690 mW cm−2 (100% RH) and 600 mW cm−2 (50% RH) at 80 °C with H2/O2) was significantly greater than that of the commercial Nafion XL membrane (maximum power density = 632 mW cm−2 (100% RH) and 485 mW cm−2 (50% RH), at 80 °C with H2/O2).
More recently, we have developed sulfonated polyphenylene membranes containing tandemly linked sulfophenylene groups (BSP-TP-f), which showed superior proton conductivity (10.7 mS cm−1 at 80 °C and 20% RH) as well as comparable fuel cell performance (power density (at 0.6 V) = 593.7 mW cm−2, 80 °C and 30% RH, H2/O2) with Nafion membrane (power density (at 0.6 V) = 601.6 mW cm−2, 80 °C and 30% RH, H2/O2).19 However, the large water absorbability (81.4% at 80 °C and 95% RH) and insufficient mechanical strength (maximum stress = 39.6 MPa and maximum strain = 90% at 80 °C and 60% RH) impeded its longevity (1640 cycles) in wet/dry cycle test at OCV. Herein, we report reinforced membranes composed of BSP-TP-f and porous expanded PE substrate via push coating method. The properties and fuel cell performance/durability of the reinforced membranes have been assessed.
The prepared reinforced membranes BSP-TP-f-4.1/PE and BSP-TP-f-4.5/PE exhibited robust mechanical properties (maximum strain = 284% or 453%, 80 °C and 60% RH), much higher than the above-mentioned physically reinforced SPP-QP-PE7 (maximum strain = 134%, 80 °C and 60% RH) and cPPSA-ePTFE (maximum strain = 10% (machine direction) and 18% (transverse direction), 25 °C and 50% RH).17,18 The proton conductivity of BSP-TP-f-4.1/PE and BSP-TP-f-4.5/PE (6.9–389.2 mS cm−1, 80 °C, 20–95% RH) was comparable or even higher compared to that of the benchmark Nafion membrane (7.9–187.2 mS cm−1, 80 °C, 20–95% RH) under the same conditions. In the accelerated durability test, the reinforced membranes presented longevity with 40673 cycles (BSP-TP-f-4.1/PE) and 22235 cycles (BSP-TP-f-4.5/PE) with wet/dry frequently cycling under OCV conditions, exceeding the US-DOE target (20000 cycles).19
Porous polyethylene (PE) substrate (thickness = 7 μm, porosity = 44%, pore size = 62 nm) was kindly supplied by Toray Industries Inc. Nafion membrane (Nafion NR211, 25 μm thick) was purchased from the Chemours Company.
Fig. 1 (a) Chemical structure of BSP-TP-f, and photos of (b) BSP-TP-f-4.1 membrane, (c) porous PE substrate, and (d) BSP-TP-f-4.1/PE reinforced membrane. |
The resulting BSP-TP-f ionomers were of high-molecular-weight (Mw = 242.4–533.0 kDa, Mn = 86.6–148.6 kDa) (Table 1) with good solubility in polar organic solvents such as N,N-dimethylacetamide and dimethyl sulfoxide but insoluble in water. From the 1H NMR spectra, the IEC was estimated to be 4.09 and 4.39 mequiv. g−1, respectively, nearly comparable to the target values. The ionomers provided brown, transparent, bendable membranes by solution casting (Fig. 1b), whose titrated IECs were 3.45 and 3.75 mequiv. g−1, indicating ca.16% of the sulfonic acid groups did not participate in the ion exchange reactions for both membranes probably because they were located in the hydrophobic domains. BSP-TP-f-4.1 and BSP-TP-f-4.5 ionomers were composited with the porous PE substrate (Fig. 1c) to obtain reinforced membranes (BSP-TP-f-4.1/PE and BSP-TP-f-4.5/PE) by a push coating method. Compared to the semi-transparent PE substrate, the reinforced membranes were more transparent without detectable defects such as pinholes and wrinkles (Fig. 1d). The titrated IECs of the BSP-TP-f-4.1/PE and BSP-TP-f-4.5/PE membranes were 1.97 and 2.31 mequiv. g−1 (Table 1), respectively, which were slightly higher than the calculated IEC values, 1.86 mequiv. g−1 for BSP-TP-f-4.1/PE and 2.19 mequiv. g−1 for BSP-TP-f-4.5/PE, taking the porosity and density of the PE substrate and assuming the full impregnation of the ionomers in the pores. The results suggest the existence of pure ionomer layer(s) with the composite layer, which is discussed below with the cross-sectional SEM images.
Sample | IEC (mmol g−1) | Molecular weightd (kDa) | ||||
---|---|---|---|---|---|---|
Targeta | NMRb | Titratedc | Mw | Mn | Mw/Mn | |
a Calculated from the feed monomer ratio.b Calculated from the integrals of the relevant peaks in the 1H NMR spectra.c Determined by acid-base titration.d Determined by GPC. | ||||||
BSP-TP-f-4.1 | 4.1 | 4.09 | 3.45 | 533.0 | 148.6 | 3.6 |
BSP-TP-f-4.5 | 4.5 | 4.39 | 3.75 | 242.4 | 86.6 | 2.8 |
BSP-TP-f-4.1/PE | — | — | 1.97 | — | — | — |
BSP-TP-f-4.5/PE | — | — | 2.31 | — | — | — |
The cross-sectional SEM images of the reinforced membranes are shown in Fig. 2a and c, in which the sandwich-like structures (triple layers) were observed in both reinforced membranes. The top and bottom layers were composed of the parent, pure BSP-TP-f ionomer while the middle layer contained BSP-TP-f ionomer and PE substrate. The composite layers were homogeneous with pores fully impregnated with the ionomer throughout the view. The parent ionomer layers in the top and bottom were well-adhered to the middle composite layers, suggesting good interfacial compatibility between the BSP-TP-f ionomer and PE substrate. The thickness of each layer of the BSP-TP-f-4.1/PE membrane was ca. 4 μm (top), 7 μm (middle), and 3 μm (bottom), respectively, making up the total 14 μm-thickness. The thickness of the BSP-TP-f-4.5/PE reinforced membrane was similar (13.5 μm) with ca. 3 μm (top), 7 μm (middle), and 3.5 μm (bottom) thickness of each layer. The triple layer structure was responsible for the slightly higher titrated IEC of the reinforced membranes than the calculated values as mentioned above. The cross-sectional sulfur distribution in the EDS analyses revealed that the sulfur intensity was much smaller in the middle layer than in the top and bottom layers as expected (Fig. 2b and d). With normalizing the maximum sulfur density of the top or bottom layer as 100%, the average sulfur density of the middle layer was ca. 39% in BSP-TP-f-4.1/PE and ca. 41% in BSP-TP-f-4.5/PE, respectively, corresponding to the porosity of the PE substrate to further support the complete impregnation of the ionomers.
Fig. 2 (a) and (c) Cross-sectional SEM images and (b and d) relative sulfur atom (Kα1) intensity in the EDS of the reinforced BSP-TP-f membranes. |
The reinforcement effect with the porous PE substrate for the BSP-TP-f ionomer membranes was also evaluated by stress/strain curves (Fig. 4 and Table S3†) and dynamic mechanical analyses (DMA) (Fig. S2†). The parent BSP-TP-f-4.1 and BSP-TP-f-4.5 membranes exhibited high Young's modulus (0.61 and 0.48 GPa) and yield stress (25.9 and 21.9 MPa) and maximum stress (39.6 and 28.5 MPa) but relatively low maximum strain (90% and 63%), typical for polyphenylene ionomer membranes. Compared with the parent membranes, the reinforced membranes BSP-TP-f-4.1/PE and BSP-TP-f-4.5/PE exhibited similarly high Young's modulus (0.36 and 0.27 GPa) and somewhat lower maximum stress but much larger maximum strain (453% and 284%), proving the significant reinforcement effect. In fact, the maximum strain of the reinforced membranes was comparable or even higher than that of Nafion membrane (335%).
Fig. 4 Stress–strain curves of parent and reinforced BSP-TP-f and Nafion membranes at 80 °C and 60% RH. |
The viscoelastic properties were measured at 80 °C as a function of the humidity (Fig. S2†). The storage moduli (E′) of parent and reinforced BSP-TP-f membranes was 0.7 × 108–2.9 × 109 Pa at dry condition and higher than that of Nafion (9.2 × 107–1.5 × 108 Pa). All membranes showed loss in E′ as increasing the humidity due to the softening effect of the absorbed water.21 There were no obvious peaks related with the glass transition in the loss moduli (E′) and tanδ (=E′′/E′) curves.
Fuel cells were evaluated with the parent BSP-TP-f-4.1 membrane, reinforced BSP-TP-f-4.1/PE and BSP-TP-f-4.5/PE, and Nafion membranes at 80 °C and 100% RH, 53% RH and 30% RH (Fig. 6). When feeding O2 as the oxidant, the open circuit voltages (OCVs) of the BSP-TP-f-4.1 (0.997 V), BSP-TP-f-4.1/PE (1.006 V) and BSP-TP-f-4.5/PE (0.999 V) cells at 100% RH were all higher than 0.99 V, supporting the aforementioned lower hydrogen permeability although they were only slightly lower than that of Nafion (1.024 V) at the same conditions. At 53% RH and 30% RH, the OCVs decreased to 0.993 V (53% RH) and 0.968 V (30% RH) for BSP-TP-f-4.1, 1.002 V (53% RH) and 0.962 V (30% RH) for BSP-TP-f-4.1/PE, 0.997 V (53% RH) and 0.991 V (30% RH) for BSP-TP-f-4.5/PE, and 1.021 V (53% RH) and 1.060 V (30% RH) for Nafion, respectively, because of the higher partial pressure of oxygen at lower humidity. The power density at 0.6 V were in the order of Nafion (0.944 W cm−2) > BSP-TP-f-4.1 (0.903 W cm−2) > BSP-TP-f-4.5/PE (0.837 W cm−2) > BSP-TP-f-4.1/PE (0.718 W cm−2) with feeding O2 as the oxidant at 100% RH (Table 2). The performance was in the same order at 53% RH and 30% RH. It is noted that the fuel cell performance of the reinforced BSP-TP-f-4.5/PE was nearly comparable to the parent BSP-TP-f-4.1 and Nafion at 100% RH and 53% RH. The minimum ohmic resistance at 100% RH was in the order of Nafion (0.073 Ω cm2) < BSP-TP-f-4.1 (0.065 Ω cm2) < BSP-TP-f-4.5/PE (0.086 Ω cm2) < BSP-TP-f-4.1/PE (0.101 Ω cm2), which were somewhat higher than those calculated from the membrane thickness and in-plane proton conductivity at 80 °C and 95% RH (0.013 Ω cm2 for Nafion, 0.004 Ω cm2 for BSP-TP-f-4.1, 0.003 Ω cm2 for BSP-TP-f-4.5/PE and 0.005 Ω cm2 for BSP-TP-f-4.1/PE) probably due to the contact resistance with the catalyst layers. The ohmic resistance became higher as decreasing the humidity as expected due to the lowered proton conductivity.
Fig. 6 IR-included polarization curves, power densities and ohmic resistances for BSP-TP-f-4.1, BSP-TP-f-4.1/PE, BSP-TP-f-4.5/PE and Nafion fuel cells at 80 °C, (a) 100% RH, (b) 53% RH, (c) 30% RH. |
Membrane | Minimum ohmic resistance (Ω cm2) | Power density at 0.6 V (W cm−2) | ||||
---|---|---|---|---|---|---|
100% RH | 53% RH | 30% RH | 100% RH | 53% RH | 30% RH | |
Nafion | 0.073 | 0.108 | 0.114 | 0.944 | 0.778 | 0.602 |
BSP-TP-f-4.1 | 0.065 | 0.095 | 0.113 | 0.903 | 0.778 | 0.568 |
BSP-TP-f-4.5/PE | 0.086 | 0.113 | 0.169 | 0.837 | 0.656 | 0.444 |
BSP-TP-f-4.1/PE | 0.101 | 0.144 | 0.206 | 0.718 | 0.495 | 0.376 |
In the case of supplying air as the oxidant, the power density at 0.6 V of the pristine BSP-TP-f-4.1 cell decreased from 0.555 W cm−2 at 100% RH to 0.382 W cm−2 at 53% RH and 0.183 W cm−2 at 30% RH, which were comparable to those of the Nafion cell (100% RH: 0.560 W cm−2; 53% RH: 0.368 W cm−2; 30% RH: 0.223 W cm−2) (Table 3). The power density at 0.6 V of the reinforced BSP-TP-f-4.5/PE cell was 0.487 W cm−2 (100% RH), 0.300 W cm−2 (53% RH) and 0.123 W cm−2 (30% RH), only slightly lower than those of the pristine BSP-TP-f-4.1 cell. The fuel cell performance of the reinforced BSP-TP-f-4.1/PE cell was more dependent on the humidity than the other cells probably because the reinforced membrane tended to lose water with air supplied at high flow rate (>80 mL min−1 when the current density was larger than 0.5 A cm−2). The idea was supported by the larger ohmic resistance of the reinforced membrane cells with air than that with oxygen, in particular, at low humidity (30% RH).
Membrane | Minimum ohmic resistance (Ω cm2) | Power density at 0.6 V (W cm−2) | ||||
---|---|---|---|---|---|---|
100% RH | 53% RH | 30% RH | 100% RH | 53% RH | 30% RH | |
Nafion | 0.069 | 0.119 | 0.214 | 0.560 | 0.368 | 0.223 |
BSP-TP-f-4.1 | 0.059 | 0.115 | 0.165 | 0.555 | 0.382 | 0.183 |
BSP-TP-f-4.5/PE | 0.082 | 0.149 | 0.409 | 0.487 | 0.300 | 0.123 |
BSP-TP-f-4.1/PE | 0.099 | 0.198 | 0.687 | 0.497 | 0.227 | 0.073 |
The IR-free polarization curves are plotted in Fig. S4.† The performance of the reinforced BSP-TP-f-4.1/PE and BSP-TP-f-4.5/PE cells were comparable or only slightly lower than those of the parent BSP-TP-f-4.1 and Nafion cells, indicating good compatibility with the Nafion-binded catalyst layers. For more quantitative discussion, the mass activity of the Pt catalyst in the cathode at 0.85 V was calculated assuming negligibly small anodic overpotential and is summarized in Table S4.† The reinforced BSP-TP-f-4.1/PE (136.6 A gPt−1) and BSP-TP-f-4.5/PE (138.2 A gPt−1) membrane cells exhibited higher mass activity than that of the parent BSP-TP-f-4.1 (112.5 A gPt−1) membrane cell at 100% RH with O2 as the oxidant. The difference became even larger at lower humidity. With air as the oxidant, the mass activity was lower than that with O2, in particular, at low humidity. The reinforced BSP-TP-f-4.1/PE and BSP-TP-f-4.5/PE membrane cells still showed higher mass activity than that of the parent BSP-TP-f-4.1 membrane cell at any humidity investigated. The results suggest that the compatibility with the catalyst layers as well as small thickness of the reinforced BSP-TP-f-4.1/PE (14 μm thick) and BSP-TP-f-4.5/PE (13.5 μm thick) membranes contributed to higher utilization of the catalysts.
Fig. 7 shows durability of the cells using parent BSP-TP-f-4.1, reinforced BSP-TP-f-4.1/PE and BSP-TP-f-4.5/PE, and Nafion membranes in the accelerated durability test (wet/dry cycle test under OCV conditions), where the ohmic resistances at dry were ensured to be larger than 2.5 times than those at wet according to the US-DOE protocol.20 The initial OCV, accelerated durability, and average OCV decay are summarized in Table 4. The initial OCV at wet was 0.895 V for parent BSP-TP-f-4.1 cell, 0.912 V for reinforced BSP-TP-f-4.1/PE cell, 0.903 V for reinforced BSP-TP-f-4.5/PE cell and 0.905 V for Nafion cell, respectively, all slightly lower than those in the polarization curves (>0.97 V) due to the lower Pt loading (0.1 ± 0.02 mg cm−2 in the cathode and 0.2 ± 0.02 mg cm−2 in the anode) in the accelerated durability test. The OCV gradually decreased with the testing time until a sudden drop which was aroused by mechanical failure of the membranes. The accelerated durability (number of cycles at the time of the sudden drop of OCV) was in the order of BSP-TP-f-4.1/PE (40673 cycles) > BSP-TP-f-4.5/PE (22235 cycles) > Nafion (8788 cycles) > BSP-TP-f (1640 cycles). The average decay of the OCV was 5.90 mV h−1 for BSP-TP-f-4.1 cell, 0.47 mV h−1 for BSP-TP-f-4.1/PE cell, 1.16 mV h−1 for BSP-TP-f-4.5/PE cell and 2.40 mV h−1 for Nafion cell, respectively. The durability of BSP-TP-f-4.1/PE and BSP-TP-f-4.5/PE membranes both exceeded the requirement in the US-DOE protocol (20000 cycles). In particular, the durability of BSP-TP-f-4.1/PE membrane was more than 2 times larger than the US-DOE target. The superior durability of the reinforced BSP-TP-f-4.1/PE and BSP-TP-f-4.5/PE membranes than the parent BSP-TP-f-4.1 membrane was based on their improved mechanical properties acquired by the reinforcement.
Fig. 7 The accelerated durability test at 90 °C with H2 and air of (a) BSP-TP-f-4.1, (b) BSP-TP-f-4.1/PE, (c) BSP-TP-f-4.5/PE, and (d) Nafion cells. |
Membrane | Initial OCVa (V) | Accelerated durability (cycles) | OCV decayb (mV h−1) |
---|---|---|---|
a Initial OCV at wet condition (100% RH).b Calculated from (drop off OCV − initial OCV)/test time, where the drop off OCV and test time were determined at the cycle where the potential suddenly dropped. | |||
BSP-TP-f-4.1 | 0.895 | 1640 | −5.90 |
BSP-TP-f-4.1/PE | 0.912 | 40673 | −0.47 |
BSP-TP-f-4.5/PE | 0.903 | 22235 | −1.16 |
Nafion | 0.905 | 8788 | −2.40 |
After the test, the reinforced BSP-TP-f-4.1/PE and BSP-TP-f-4.5/PE membranes were recovered from the cells and the ionomers were extracted with DMSO and analyzed by NMR spectra and GPC (Fig. S5 and Table S5†). While the 19F NMR spectrum did not change, the 1H NMR spectra of the recovered ionomers showed minor changes. The peak integral indicated that the post-test BSP-TP-f-4.1 and BSP-TP-f-4.5 ionomers lost ca. 32% and ca. 27% of the sulfonic acid groups, respectively. The residual molecular weight was ca. 67% and 74% (based on Mw) for BSP-TP-f-4.1 and BSP-TP-f-4.5, respectively. The larger losses in IEC and molecular weight of the BSP-TP-f-4.1 were probably attributed to its longer durability cycles.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d3ra01041d |
This journal is © The Royal Society of Chemistry 2023 |