Polyazomethines and their acid–base interactions with Nafion and Nafion–imidazole membranes for efficient fuel cells

Abstract

We propose, for the first time, aromatic polyazomethines applied in polymer electrolyte membrane fuel cells (PEMFC) as tetrafluoroethylene-perfluoro-3,6-dioxa-4-methyl-7-octene-sulfonic acid copolymer (PFSA) membrane modifiers. Two types of polyazomethines were selected to impregnate pristine PFSA membranes; one with bibenzimidazole groups (25Bo-BABPI) and the second with thiophene rings (2252Th-DMB). In addition, these composite membranes were doped using imidazole (Im) to increase the proton conductivity. The success of the impregnation process of all modified membranes has been confirmed based on FTIR spectroscopy. From the water sorption analysis performed, the obtained PFSA composite membranes were found to be thermally stable (TGA, DTG results) and revealed lower dependency on water, which served as an advantage. The typical membranes used were activated in boiling deionized water, then in 3.75% H₂O₂, and finally in boiling 0.2 M H₂SO₄. Moreover, both polyazomethines were oxidized with FeCl₃ to check their chemical oxidation abilities. Single PEMFCs assembled using polyazomethine composite membranes were characterized by improved performance, increase in the maximum power density by two folds, higher proton conductivity and larger electrochemical surface area of catalyst in comparison to reference samples. The maximum power density of 231 mW cm⁻² at the current density of 637 mA cm⁻² was detected for PFSA–25Bo-BABPI membrane. In accordance with the experiments carried out, polyazomethines have positive influence on PEMFCs' overall electrochemical properties. All membranes were tested by atomic force microscopy (AFM) and scanning electron microscopy (SEM) with EDX to analyze their morphology.

---

1. Introduction

The appropriate water management in fuel cells is a matter of great importance. Water is a by-product released constantly in electrochemical reactions occurring in these devices. Numerous fuel cells have the ability of self-humidification of their crucial components. The ion-conducting electrolyte in polymer electrolyte membrane fuel cells (PEMFC) contains water molecules incorporated in its structure. Therefore, the operating principle of fuel cells could be beneficial when the produced water is used to humidify their electrolyte. However, high dependency on water becomes a drawback in certain conditions. For instance, if the fuel cell stack is switched off for a long period, immediate cell operation on switching it on, under its nominal power, can be impossible due to membrane dehydration phenomenon. The other aspect is the continuous operation of fuel cells with high current density that leads to an increase in the temperature of the components of the fuel cell, which can affect its performance. In this case, at anode side, the humidification rate could be insufficient and that could lead to high voltage losses. Therefore, the appropriate balance of water in the fuel cells is necessary.^1–6

The modification of electrolyte structure was the subject of our previous study^7–9 with regard to various modifiers applied for tetrafluoroethylene-perfluoro-3,6-dioxa-4-methyl-7-octenesulfonic acid copolymer (Nafion)-based membranes. For example, in an earlier report,⁷ we have examined the impact of 2,2′-bis(4-aminophenyl)-5,5′-bibenzimidazole (BAPBI) and imidazole (Im) on the electrochemical properties of Nafion electrolyte with respect to the parameters such as ionic conductivity and power density. Based on the modification process, membrane water uptake scaled down and overall performance increased because of BAPBI/imidazole presence, which served as an advantage. In a relevant study,⁹ aromatic polyimides and imidazole were utilized as modifying compounds during impregnation process of Nafion-115 membrane. Similarly, the presence of external chemical compounds decreased the water uptake and caused significant improvement of fuel cell performance particularly at 60 °C in comparison to the reference sample.

Polyazomethine has been used in PEMFCs, as an electrolyte basic component, with sulfonated polyaryletherketone (SPAEK), acting as matrix material, and polysilsesquioxane according to a previous study.¹⁰ Such membranes are characterized by superior combination of properties as well as proton conductivity and methanol impermeability for use in direct methanol fuel cell. Neuse et al. synthesized polybenzimidazoles using a two-stage synthesis process that involves low-temperature solution polymerization of aromatic bis(o-diamines) with aromatic dialdehydes and subsequent conversion of the resulting azomethine-type prepolymers to aromatic polybenzimidazoles.¹¹

Other authors have published relevant reports with regard to bibenzimidazoles; however, they were used as electrolyte materials in high temperature proton exchange membrane fuel cells (HT PEMFC).¹² For instance, F. Mack et al.¹³ synthesized and characterized novel acid-based polybenzimidazole (PBI) blend membranes demonstrating high thermal and excellent chemical stabilities in terms of oxidative weight loss. Based on these blends, membrane electrode assemblies (MEAs) were prepared showing good performance. The research proved that acid–base PBI blends are suitable alternatives to commonly-used pristine PBI and AB-PBI (poly(2,5-benzimidazole)) for HT PEMFC because they have better thermal and chemical stability and higher ionic conductivity. J. Yang et al.¹⁴ prepared a novel composite membrane by introducing well-dispersed triazole functionalized graphene oxide into the polybenzimidazole. Based on this modification and after doping by phosphoric acid, an improvement in proton conductivity and tensile strength was observed compared with the pure PBI membrane, according to fuel cell performance evaluation.

In this study, we investigate, for the first time, the modification of PFSA membranes utilizing polyazomethines (PAZ) – the polymers synthesized as condensation product of diamine and dialdehyde. These materials were known from applications in organic photovoltaic cells due to their low-cost fabrication and ease of purification process because water is the only co-product of their chemical synthesis.^15–17

The main goal of this study was to show that it is possible to utilize polyazomethines in PEMFC in order to improve the electrochemical parameters of fuel cells. In the first part of our study, two types of polyazomethines were synthesized as fuel cell membrane modifiers. The first one is PAZ with bibenzimidazole groups (25Bo-BABPI) – the chemical groups that can be found in the polymer applied for high temperature PEMFC. This polyazomethine was obtained by two-step polycondensation reactions based on the previously synthesized BAPBI diamine.⁷ The second one is PAZ with thiophene rings (2252Th-DMB) comprising few good properties, like simplicity of dissolution or possibility of one-step-condensation synthesis process. In this case, PAZ has three thiophene rings per repeating unit.

The primary goal for application of polyazomethines to PEMFCs was to improve the fuel cell electrochemical performance as well as to decrease the activation and resistance losses. The second goal was to mitigate membrane dependency on water; however, this in fact was achieved as an additional benefit.

It was shown in previous reports^7–9,13 that acid–base blend membranes consisting of a polymer containing sulfonic groups and a basic compound with N-heterocycle groups are one of the most promising materials in the last decade for the development of PEMFCs. For this reason, herein, we also investigate the electrochemical properties of composite fuel cell membranes with regard to acid–base interactions occurring between polymer matrix (PFSA) and incorporated modifiers (PAZ, Im). Moreover, both PAZ and Im were complexed with FeCl₃ and investigated via UV-vis spectroscopy to check their oxidation properties. Finally, all the synthesized membranes were tested using atomic force microscopy (AFM) and scanning electron microscopy (SEM) with EDX to analyze their morphology.

As a final point, we compared the properties of obtained PEMFCs based on PAZ with PEMFCs previously investigated by us that were modified using BAPBI. According to our research, we could conclude that the polymers proposed in this study, which contain imine bonds, are the most promising materials for utilization in polymer fuel cells among all previously investigated chemical compounds.^7–9

2. Experimental

2.1. Materials and synthesis procedure

2,2′:5′,2′′-Terthiophiene-5,5′′-dicarboxaldehyde, 2,5-bis(octyloxy)terephthalaldehyde, 3,3′-dimethoxybenzidine, N,N-dimethylacetamide (DMA), 3,3′-diaminobenzidine, 4-aminobenzoic acid, imidazole, poly(phosphoric acid) (PPA), CHCl₃ and FeCl₃ were purchased from Sigma-Aldrich and used as received. Methanol and acetone were purchased from POCH and used as received. Tetrafluoroethylene-perfluoro-3,6-dioxa-4-methyl-7-octene-sulfonic acid copolymer (PFSA, Nafion) film was purchased from Quintech Brennstoffzellen Technologie.

2,2′-Bis(4-aminophenyl)-5,5′-bibenzimidazole (BAPBI) was obtained according to the synthesis described in a previous literature.⁷ The 2252Th-DMB polymer was obtained and characterized in our previous study used as a donor in polymer solar cells.¹⁵

2.2. Fabrication of composite PFSA membranes and single fuel cells

Nafion-115 (DuPont) was selected to prepare composite PFSA membranes. All PFSA membranes were fabricated via immersion of purified Nafion membrane into the typical polymer solutions of DMA used as solvent according to the following procedure. Equal quantities of 2252Th-DMB and 25Bo-BABPI polyazomethines (25 mg) were separately dissolved in 25 g of DMA. Two more solutions were prepared by addition of 25 mg of 2252Th-DMB or 25Bo-BABPI polyazomethine to 100 mg of imidazole (Im) and 25 g DMA. Furthermore, 5 cm² PFSA films were immersed four times in polyazomethine–DMA or polyazomethine–Im–DMA solutions for 15 minutes with drying at 105 °C after each step. In summary, four groups of composite membranes were obtained and marked as PFSA–2252Th-DMB, PFSA–2252Th-DMB–Im, PFSA–25Bo-BABPI and PFSA–25Bo-BABPI–Im. In addition, two PFSA membranes were used to prepare reference fuel cells – one with pristine PFSA membrane and one containing PFSA doped with imidazole (PFSA–Im). Before the membrane electrode assembly (MEA) preparation, all composite membranes were activated by cleaning in boiling deionized water for one hour, purifying in 3.75% H₂O₂ for one hour and protonating in boiling 0.2 M H₂SO₄ also for one hour.

For MEA preparation, including both the anode and cathode, a commercially available gas diffusion electrodes (SLGDE FuelCellsEtc) were selected to provide the same conditions for fuel cell study. The electrodes are characterized by the standard Pt loading of 0.5 mg cm⁻² (60 wt% on carbon), thickness 410 μm, the basic weight 180 g m⁻² and through-plane resistance of 13 mΩ cm². They were coated with Nafion for improved water management and good adhesion to the composite membrane structure.

Gas diffusion electrodes and activated membrane were hot pressed at 100 °C and 30 bar for 5 minutes to form MEAs of the typical single fuel cells studied.

Architecture of constructed PEMFCs along with images of the membranes after three-step activation process is shown in Fig. 1.

Fig. 1 Architecture of constructed PEMFC with polyazomethines (a) along with photos of the created membranes (b) PFSA (1), PFSA–2252Th-DMB (2), PFSA–25Bo-BABPI (3), PFSA–2252Th-DMB–Im (4), PFSA–25Bo-BABPI–Im (5).

2.3. Fuel cell characterizations

PFSA composite membranes were characterized directly via thermogravimetric method (TGA, Mettler-Toledo AG apparatus at a heating rate of 10 °C min⁻¹ under nitrogen), infrared spectroscopy (FTIR, Nicolet 5700 (ThermoElectron) in the range of 4000–400 cm⁻¹ at a resolution of 2 cm⁻¹ and for accumulated 32 scans) and water sorption analysis (Micromeritics ASAP2020, Accelerated Surface Area and Porosimetry System). Indirect analysis was carried out by conducting electrochemical experiments. Three types of indirect measurement techniques were used: polarization curve (I–V, H&H DC load), electrochemical impedance spectroscopy (EIS, Solartron SI1260), cyclic voltammetry (CV, Solartron SI1287) and linear sweep voltammetry (LSV). Fuel cell test fixture (Pragma Industries), equipped with 1 cm² graphite single-serpentinemonopolar plates, was used. I–V characterization was carried out delivering hydrogen/air with flow rate of 60/80 mL min⁻¹, at ambient pressure, in voltage mode for 60 s, from open circuit voltage to 0.25 V at 60 °C and Rh 90%. Both hydrogen/air lines were humidified. The I–V polarization curves were performed as long as steady state conditions were achieved. The same gas configuration was chosen in case of EIS technique, during which fuel cells were scanned in the frequency range from 40 kHz to 0.1 Hz with a signal amplitude of 250 mV at OCV condition to obtain Nyquist and Bode plots. Finally, using CV and LSV method, the fuel cells were analyzed by feeding the cathode with argon (50 mL min⁻¹, ambient pressure), which acted as counter and reference electrodes, and anode with hydrogen (the same gas configuration), which acted as the working electrode. This measurement was acquired at steady state by applying voltage at a scan rate of 10 mV s⁻¹.

3. Results and discussion

Both polyazomethines were prepared using a simple one- or two-step high-temperature condensation procedure using a catalyst in the solution. The polymerization conditions were not optimized. In both cases, polycondensation reaction of polyazomethines was carried out in DMA solution at 160 °C for 22 hours in the presence of p-toluenesulfonic acid. The synthesis of both polyazomethines (i.e., 2252Th-DMB and 25Bo-BABPI) is outlined in Fig. SI1 in ESI.† Details about synthesis and characterisation of 25Bo-BABPI are also described in the ESI.†

2252Th-DMB polyazomethine was soluble in chloroform and DMA, while 25Bo-BABPI was soluble only in DMA. The structures of the polyazomethines were confirmed by FTIR and ¹H NMR spectroscopy; the results were in a good agreement with the proposed chemical structure. The presence of imine groups was confirmed by FTIR spectroscopy; the characteristic band of the –HC=N– stretching deformations at 1608 cm⁻¹ for 2252Th-DMB and at 1615 cm⁻¹ for 25Bo-BABPI was observed. Moreover, the characteristic band of the carbonyl group was observed at approximately 1660 cm⁻¹ and 1672 cm⁻¹ for 2252Th-DMB and 25Bo-BABPI, respectively. The FTIR spectra of both polymers are shown in ESI (Fig. SI2a).† In the proton NMR spectrum of 2252Th-DMB, the azomethine proton signal at approximately 8.70 ppm was observed, as expected.¹⁵

The results from the TGA analysis in a nitrogen atmosphere suggested that both polyazomethines (2252Th-DMB and 25Bo-BABPI) possessed good thermal stability with a 5% weight loss at 377 °C and 373 °C, respectively. The high thermal stability of the polymers prevents deformation of their morphology and is important for real applications.

3.1. PEMFC composite membrane evaluation

In the first step of this study, FTIR spectroscopy of binary and ternary membranes was performed to determine the interactions between PFSA, polyazomethines and Im (see Fig. SI2b†). In the FT-IR spectra of modified membranes, the changes were not observed in the range of 500–1500 cm⁻¹. The characteristic band of the –HC=N– stretching deformations at 1608 cm⁻¹ for 2252Th-DMB and at 1615 cm⁻¹ for 25Bo-BABPI was shifted to higher wavenumber (1647 and 1636 cm⁻¹, respectively). This is an effect of protonation of hydrogen acceptor sites (–N=C–) in 2252Th-DMB and 25Bo-BABPI by hydrogen atom of SO₃H group in PFSA membranes, as is schematically presented in Fig. 2.

Fig. 2 Possible acid–base interactions between PFSA–PAZ–Im.

Moreover, along with protonation, the change of the shape of this peak was observed. For PFSA–Im membrane the peak at 1686 cm⁻¹ was observed, indicating a shift to a higher wavenumber in comparison to PFSA membrane. For ternary membrane, this peak was shifted to 1728 cm⁻¹. The –HN=C– stretch of PAZ was shifted to higher wavenumber in the ternary membrane (see Fig. SI2b†).

The spectra of pristine PFSA membrane exhibit characteristic absorption peaks at 1200 cm⁻¹ and 1140 cm⁻¹ that correspond to the –CF₂– stretching vibrations (asymmetric and symmetric, respectively) of the fluorinated main chain of Nafion. The band at 1056 cm⁻¹ is related to the symmetric stretching vibration of the S–O group, while the peak at 976 cm⁻¹ is attributed to the symmetric stretching vibration of the C–O–C and CF₃ groups. A broad peak at about 3441 cm⁻¹ is attributed to the O–H stretch.^8,18–20 The peak at 1730 cm⁻¹ that is probably attributed to unsaturated bonds of –CF=CF– was detected by us and by Alentiev et al.²¹ Moreover, Gruger et al.²² described the peak at 1730 cm⁻¹ as an effect of PFSA membrane hydration and as the result of the presence of OH bending vibrations in this region. In summary, for PFSA, two bending vibrations are observed at ca. 1630 cm⁻¹ (shoulder) and ca. 1730 cm⁻¹ (see Fig. SI2b†), which could be described as the presence of two types of protonic species coming from S–OH groups in PFSA and OH groups from water, respectively. Interactions between OH and –SO₃H in PFSA are also possible; however, in the presence of Im or PAZ they could not be distinguished.

The membrane is a crucial component in terms of durability of PEM fuel cells. Pristine Nafion, the material commonly used for fuel cell fabrication, exhibits satisfactory performance in terms of chemical and thermal stability. The aim of this research was to examine the thermal stability of Nafion and to analyze if the impregnation process could deteriorate the thermal properties of Nafion. TGA analysis was performed in nitrogen atmosphere, in the range from 25 to 800 °C, maintaining a constant heating rate of 10 °C min⁻¹. For this study, all membranes (including reference ones) were prepared following the same method as mentioned above. However, we observed some differences in initial hydration of typical specimens that may have an impact at temperatures below 100 °C. The results of thermogravimetric analysis are shown in ESI.†

The high decomposition temperature suggests that polyazomethines-based membranes should be thermally stable under fuel cell working conditions during long-term operation.

Typical water sorption isotherms recorded for PFSA–2252Th-DMB, PFSA–25Bo-BABPI and reference membranes are shown in Fig. 3. ESI† contains details about water sorption analysis. In case of pristine PFSA membrane, the isotherm shape reveals two steps of water adsorption. For the applied pressure of up to 0.8p/p⁰, the water sorption rate was slow, but above the pressure of 0.8p/p⁰, sudden increase of water adsorption occurred. It could be considered that the first step corresponds to solvation of ions by water molecules in the membrane. In the second step, water fills the pores and swells the polymer. This phenomenon is significantly reduced in case of PFSA composite membranes. The impregnation process introduced polyazomethine/imidazole moieties into the Nafion free absorption centers. This resulted in the limited H₂O content in membrane volume, as shown in Fig. 3. The water uptake calculated at 0.9p/p⁰ for PFSA membrane was 19.5%, whereas that for PFSA–Im, PFSA–2252Th-DMB and PFSA–25Bo-BABPI was 7.6, 10.2 and 8.0%, respectively. The results show approximately double reduction of membrane capacity to absorb water, which means that composite membrane dependency on water was remarkably mitigated.

Fig. 3 Water sorption isotherms obtained at 25 °C for the typical membranes used.

A higher water uptake leads to a higher ionic conductivity, but it also lowers the fixed charge concentration and typically reduces permselectivity of the membrane. Therefore, a balance should be established between water uptake, ionic conductivity and membrane permselectivity for efficient fuel cell operation.²³

In addition, the morphology of the modified membranes was investigated by AFM and SEM (see ESI†). In summary of the AFM study, we could conclude that the investigated membranes exhibited noticeable differences in the surface morphology. The phase image showed a homogenous structure for all investigated materials in terms of mechanical properties (viscous–elastic). The highest surface area ratio, related to the smallest grains, had an impact on the high efficiency of the PFSA–25Bo-BABPI membrane. PFSA–25Bo-BABPI–Im membrane exhibited the highest values of R_a and R_ms (see Table SI2†). The lowest values of R_a and R_ms were observed for PFSA–Im membrane. The SEM images show that the PFSA membrane is uniform, compact, and devoid of visible pores. In case of modified PFSA membranes, few changes in the morphology of the materials were visible, whereas in the case PFSA–25Bo-BABPI–Im and PFSA–2252Th-DMB membranes, numerous morphological changes were observed. This behavior confirmed that typical compounds (modifiers) used were introduced in the PFSA membranes.

3.2. Electrochemical performance of composite membrane-based fuel cells

To investigate the impact of polyazomethines on the electrochemical performance of fuel cells, I–V polarization curves, electrochemical impedance spectroscopy, cyclic and linear sweep voltammetry tests were carried out. Typical results of the electrochemical characterization of the fuel cells studied are summarized in Table 1.

Table 1 Electrochemical parameters of polyazomethine-based fuel cells and reference samples

Code	PFSA	PFSA–Im	PFSA–2252Th-DMB	PFSA–2252Th-DMB–Im	PFSA–25Bo-BABPI	PFSA–25Bo-BABPI–Im
a Not determined.
J max [mA cm^−2]	299	463	485	437	637	468
J at 0.6 V [mA cm^−2]	84	157	269	98	327	122
P max [mW cm^−2]	95	140	192	123	231	136
Efficiency at Pmax [%]	27	27	34	24	32	26
OCV [V]	0.91	0.89	0.97	0.91	0.97	0.91
σ [mS cm]	45.9	54.1	94.3	78.7	75.2	68.5
ECSA [m^2 g^−1]	24.6	8.28	23.4	7.2	35.0	3.8
Water uptake [%]	19.5	7.6	10.2	—^a	8.0	—^a

---

The polarization curve method provides information regarding electrochemical characterization of fuel cells, including polarization losses and electrochemical performance. In Fig. 4 and 5, the I–V curves for all investigated polymer fuel cells are presented. The I–V curves in Fig. 4 were obtained for fuel cells working at nominal conditions (60 °C, Rh 90%), whereas those in Fig. 5 were obtained under conditions of decreased humidity and temperature (25 °C, Rh 50%) to demonstrate the mitigation of water dependency of the modified membranes.

Fig. 4 Polarization curves of typical polyazomethine-based and reference single PEMFCs, obtained at 60 °C, Rh 90%, H₂/air 60/80 mL min⁻¹.

Fig. 5 Polarization curves of typical polyazomethine-based and reference single PEMFCs, obtained at mitigated humidity of Rh 50% and temperature 25 °C.

In case of both reference fuel cells (PFSA and PFSA–Im), the maximum current densities obtained were 299 and 463 mA cm⁻², and maximum power densities were 95 and 140 mW cm⁻², respectively (Fig. 4). As was already published in our previous report,⁷ the doping of PFSA by imidazole could significantly increase the ionic conductivity of electrolytes; therefore, the overall performance could be improved. However, the so-called poisoning effect of imidazole on platinum could limit the practical usefulness of such a doping compound. In this context, both polyazomethines, impregnated to PFSA electrolyte, provide positive feedback in terms of fuel cell performance. Polyazomethine-modified fuel cells obtained high current and power densities in comparison to reference samples. For PFSA–2252Th-DMB and PFSA–25Bo-BABPI, the maximum values of current and power density were 485 mA cm⁻² and 192 mW cm⁻² and 637 mA cm⁻² and 231 mW cm⁻², respectively, according to the data obtained from the polarization curves.

The mitigation of membrane dependency on water was proven performing polarization tests at lowered humidity (Rh 50%) and temperature (25 °C). In these conditions both polyazomethines-based fuel cells obtained good results in comparison to reference samples (Fig. 5). The maximum current and power density for PFSA–2252Th-DMB and PFSA–25Bo-BABPI were 541 mA cm⁻² and 146 mW cm⁻² and 499 mA cm⁻² and 169 mW cm⁻², respectively. In contrast to these samples, initial dehydration of reference samples resulted in low performance according to polarization curves (204 mA cm⁻², 77 mW cm⁻² and 337 mA cm⁻², 89 mW cm⁻² for PFSA and PFSA–Im single cells).

Based on the polarization curves (Fig. 4), the activation and concentration polarization as well as the ohmic voltage losses were calculated (see ESI†). The obtained results of evaluation and the necessary coefficients to calculate voltage losses are summarized in Table 2.

Table 2 Voltage losses and given coefficients characterized polyazomethine-based and reference fuel cells

Parameters	PFSA	PFSA–Im	PFSA–2252Th-DMB	PFSA–2252Th-DMB–Im	PFSA–25Bo-BABPI	PFSA–25Bo-BABPI–Im
Tafel slope [mV per decade]	129	146	101	109	137	117
Cathode transfer coefficient	0.51	0.45	0.66	0.61	0.48	0.57
Exchange current density [mA cm^−2]	0.0044	0.0362	0.0026	0.0008	0.0989	0.0025
Limiting current density [mA cm^−2]	350.0	606.3	510.8	584.4	685.6	623
Activation polarization losses, 50 mA [mV]	525	461	432	519	370	502
Ohmic losses, 0.2 A [mV]	57	39	30	35	38	40
Concentration polarization losses, Jmax [mV]	82	67	108	52	117	55

---

The obtained data revealed the reduction of activation polarization and ohmic losses observed for polyazomethine-based fuel cells, leading to their improved performance in comparison to PFSA and PFSA–Im. For instance, in accordance with polarization curve data, at the current density of 50 mA cm⁻², the activation polarization losses in case of PFSA and PFSA–Im are 525 and 461 mV, whereas for PFSA–2252Th-DMB and PFSA–25Bo-BABPI the values are 432 and 370 mV, respectively. Ohmic losses occurring at 200 mA cm⁻², caused by electrolyte resistance measured in high frequency region (HFR) of Nyquist plots, created the voltage drop of 57 mV and 39 mV in case of PFSA and PFSA–Im and 30 mV and 38 mV for PFSA–2252Th-DMB and PFSA–25Bo-BABPI, respectively. Furthermore, taking into consideration the higher heating value of hydrogen (HHV), the voltage efficiency at the maximum power density for both polyazomethine-modified fuel cells was elevated i.e. 34% for PFSA–2252Th-DMB and 32% for PFSA–25Bo-BABPI in comparison to reference fuel cells (both 27%). Such improvement confirms the mitigation of activation voltage and ohmic voltage losses.

Due to higher performance of PFSA fuel cells doped with imidazole in comparison to pristine one, relevant doping was applied during impregnation process of polyazomethine membranes. The I–V polarization curves for PFSA–2252Th-DMB–Im and PFSA–25Bo-BABPI–Im revealed an unexpected effect of fuel cell performance reduction. The maximum power density for both Im-doped polyazomethine fuel cells was lower than that for polyazomethine fuel cell without imidazole and even lower than that for PFSA–Im, i.e. 123 mW cm⁻² in case of PFSA–2252Th-DMB–Im and 136 mW cm⁻² in case of PFSA–25Bo-BABPI–Im. The reason could be that imidazole created strong chemical bonds with a part of the platinum catalyst surface, resulting in high activation polarization losses (according to the polarization curve data at 50 mA cm⁻², for PFSA–2252Th-DMB–Im and PFSA–25Bo-BABPI–Im at 519 mV and 502 mV, respectively). Ohmic losses were approximately the same as in case of imidazole-free polyazomethine-based fuel cells, i.e. at 200 mA cm⁻² at 35 mV and 40 mV.

According to the results summarized in Table 2, the best fuel cells evaluated in this study are characterized by approximately two-fold concentration polarization voltage losses in comparison to other fuel cells (108 and 117 mV for PFSA–2252Th-DMB and PFSA–25Bo-BABPI). The effect caused by the limited diffusion of reactant gases showed that the setup of experiments was not properly adjusted, and that the polyazomethine-based fuel cells would obtain even higher results of performance evaluation in better testing conditions.

It was found that polyazomethines improved the electrochemical properties of PEMFC in the most effective way than other modifiers investigated in our previous study. For example, in comparison to our data published in an earlier report⁷ with respect to the electrochemical performance, 25Bo-BABPI polyazomethine achieved significantly better results than BAPBI modifier of fuel cell membrane. For PFSA–25Bo-BABPI, the values of maximum current density were ca. 40% higher and maximum power density ca. 60% higher than in case of PFSA–BAPBI.

To confirm the favorable influence of 2252Th-DMB and 25Bo-BABPI polyazomethines on fuel cell membrane and the catalyst layer, EIS and CV measurements were performed. According to EIS spectra depicted in Fig. 6, PFSA–Im, PFSA–2252Th-DMB and PFSA–25Bo-BABPI demonstrated reduced charge and mass transfer resistance than PFSA, PFSA–2252Th-DMB–Im and PFSA–25Bo-BABPI–Im (app. width of semicircle at OCV 3.33 vs. 3.94 Ω cm²). Such transport phenomenon improvement may be ascribed to the fact that both polyazomethines, which were added during membrane impregnation process, could interact with the catalyst layer as a consequence of MEA formation. Similar interactions were also confirmed using the CV method.

Fig. 6 Electrochemical impedance spectra at OCV (quasi-OCV i.e. without external electrical load; at internal resistance of FRA analyzer) of single PEMFCs.

Based on EIS results in HFR region and taking into account the thickness of the typical membranes used, proton conductivity of electrolytes (σ) was calculated. In case of PFSA fuel cell, the value of σ was 45.9 mS cm⁻¹, whereas the addition of imidazole increased the conductivity to the value of 54.1 mS cm⁻¹. Significant improvement could be observed for both PFSA–2252Th-DMB and PFSA–25Bo-BABPI fuel cells, for which the proton conductivity was found at 94.3 and 75.2 mS cm⁻¹, respectively. These results were partially confirmed for imidazole-doped polyazomethine fuel cells (σ equals to 78.7 and 68.5 mS cm⁻¹). Relevant results are shown in Bode plots (the impedance magnitude vs. logarithm of the frequency) depicted in Fig. 7 along with an equivalent circuit used for the evaluation of measured impedance spectra, where polyazomethine-modified PFSA membrane resistance could be easily distinguished from cathode mass-transport and charge transfer resistance.

Fig. 7 Impedance spectra of single PEMFCs (calculated data: solid line: experimental data: symbols) along with equivalent circuit used for the evaluation of measured impedance spectra (R_A∥C_A – element correspond to the rate determining processes at the anode, R_C∥C_C – element correspond to the rate determining processes at the cathode, C_A, C_C – constant phase element, R_N∥C_N – the finite diffusion element, R_M – membrane resistance and series resistance, L_s – feed line inductance).

In Table 3 the equivalent circuit parameters for all single PEMFCs obtained from the fitting impedance curves are presented. The frequency dependence of the C_A,C impedance takes the form of the following equation: , where CT and |Cp| ≤ 1 are parameters. When 0 < p < 1, the parameters correspond to the distribution of charge transport processes rather than to a single one. Its value could be used to distinguish between hopping and diffusion transport mechanisms. In our study, depending on the fuel cell, various activation polarization losses could result from different porosity and thickness of the electrolytes and various fractions of Pt catalyst in the electrode layers that were partially brought into ionic contact with the membranes. Our preliminary investigation of membranes showed that the thickness increases in the following order: PFSA–Im (105 microns) < PFSA (130 microns) < PFSA–25Bo-BABPI–Im (136 microns) < PFSA–2252Th-DMB–Im (139 microns) < PFSA–25Bo-BABPI and PFSA–2252Th-DMB (144 microns). Due to different magnitude of polymer swelling, depending on the type of modifier used, it was challenging to maintain the membranes with identical thickness. The various thicknesses of composite membranes suggest that polyazomethines probably fill the pores and cover the surface of the PFSA membrane. On the other hand, PFSA–Im membrane exhibited lower value of thickness than Nafion-115, which shows that Im influenced on the chemical structure of PFSA and altered the mechanical properties of the membrane. For a given membrane, higher thickness and lower ionic conductivity could be achieved due to increase of through-plane resistance; therefore, lower fuel cell performance is expected. However, in our study, this correlation has been reversed on comparing the reference and modified single cells: the thickest membranes are those modified by polyazomethines for which MEAs achieved the best results.

Table 3 The equivalent circuit parameters for all single PEMFCs obtained from the fitting (R_A∥C_A – element correspond to the rate determining processes at the anode, R_C∥C_C – element correspond to the rate determining processes at the cathode, C_A, C_C – constant phase element, R_N∥C_N – the finite diffusion element, R_M – membrane resistance and series resistance, L_s – feed line inductance)

Code	PFSA–2252Th-DMB	PFSA–2252Th-DMB–Im	PFSA–25Bo-BABPI	PFSA–25Bo-BABPI–Im	PFSA	PFSA–Im
χ ^2 [×10^−6]	507	486	568	358	224	339
R M [Ω]	0.150	0.168	0.182	0.175	0.264	0.186
L S [nH]	642	626	630	623	612	633
R A [Ω]	0.436	0.822	0.915	0.623	1.374	1.150
CT A [s^p Ω^−1 × 10^−3]	76.07	51.16	51.09	62.78	39.16	12.35
Cp A [—]	1.000	1.000	0.853	0.932	0.790	0.966
R C [Ω]	0.496	0.881	0.248	0.473	0.248	0.494
CT C [s^p Ω^−1 × 10^−3]	58.34	38.90	255.9	121.3	271.5	037.66
Cp C [—]	0.747	0.770	0.551	0.623	0.503	0.732
R N [Ω]	2.456	2.371	2.157	2.757	2.376	1.713
C N [mF]	41.58	49.62	45.27	35.52	37.76	28.09

---

In comparison to PFSA–BAPBI,⁷ according to EIS analysis, polyazomethine-modified PEMFC are characterized by increased electrolyte conductivity and improved mass and charge transport mechanism. This a great feature of PAZ for improving the PFSA-based PEM fuel cell electrochemical properties.

Cyclic voltammetry experiments confirmed the poisoning effect of imidazole, deteriorating the active surface area of the catalyst. Based on CV voltammograms (ESI, Fig. SI9†), using the area of hydrogen desorption peaks, electrochemical surface area (ECSA) was calculated for each fuel cell. ECSA is a physical parameter that describes the real area of Pt catalyst that takes part in an electrochemical reaction. In case of PFSA fuel cell, ECSA was 24.6 m² g⁻¹. The negative influence of imidazole caused large reduction of ECSA for PFSA–Im, PFSA–2252Th-DMB–Im and PFSA–25Bo-BABPI–Im, i.e. 8.3, 7.2 and 3.8 m² g⁻¹, respectively. There was no influence of 2252Th-DMB polyazomethine on ECSA (23.4 m² g⁻¹); however in case of PFSA–25Bo-BABPI increased value of ECSA was obtained (35.0 m² g⁻¹). Thus, it could be assumed that polyazomethine, containing bibenzimidazole groups, has a positive impact on the catalyst layer of the fuel cell. Such an improvement in ECSA was not observed in case of PFSA–BAPBI,⁷ for which Pt surface active area was approximately identical to that for PFSA reference fuel cell.

Imidazole interaction with platinum catalyst was already studied and discussed in the literature.²⁴ In this study, the chemical compounds, used as modifiers, reduced the electrochemical surface area (ECSA) of the catalyst via occupying the hydrogen adsorption centers on the surface of platinum that could be confirmed according to the cyclic voltammetry investigation results. This seems to be the only possible explanation of significant ECSA mitigation obtained in this study. The results of CV experiments are summarized in Table 1 of the manuscript, according to which the sudden drop of ECSA is observed for imidazole-based fuel cells against other imidazole-free fuel cells.

Finally, linear sweep voltammetry (LSV) showed that only in the case of PFSA–Im fuel cell, the hydrogen crossover through the membrane was observed (9.98 × 10⁻⁹ mol cm⁻² s⁻¹). This fact results from the decrease of membrane thickness because of imidazole impact. This finding shows that in addition to the polarization tests, EIS and CV measurements infer that Im has a negative influence on the parameters of polymer fuel cells in terms of deteriorating of the membrane physical properties. LSV curve of PFSA–Im is presented in ESI (Fig. SI10).†

Analyzing the chemical structure of membranes, we could conclude that the investigated PEMFCs exhibited vehicle-type and Grotthuss-type mechanism, depending on the composition of the membrane (two- or three-component system). In the case of PFSA–polyazomethine membranes, Grotthuss-type mechanism would be predominant, while in the case of PFSA–polyazomethine–Im membranes vehicle-type mechanism would prevail. However, based on the experiments that were carried out in this study, it is impossible to distinguish the type of the proton-conducting mechanism.

The interactions between hydrogen atoms of –SO₃H group in PFSA and nitrogen atoms in imine groups (see Fig. 2) take active part in proton transport, replacing water from its structure. Low influence of water contribution into proton transport was also confirmed by low water adsorption value (below 10%). The six nitrogen atoms of PFSA–25Bo-BAPBI per mer unit act as proton donors (benzimidazole) and acceptors (imine and benzimidazole), while in the case of PFSA–2252Th-DMB Grotthuss-type mechanism based on two nitrogen atoms (imine) and three sulfur atoms (thiophene) per mer unit act only as acceptors (see Fig. 2). These interpretations are in agreement with the study published by Fu et al.^25,26

Conclusion

Polyazomethine-based composite membranes PEMFCs were successfully assembled. The utilization of chemical compounds such as polyazomethines with bibenzimidazole groups, thiophene rings and imidazole resulted in the decrease of membrane water uptake in consequence of free sorption centers occupied by the typical modifiers used in this study. This led to the mitigation of electrolyte conductivity dependency on water inside the PFSA membrane structure. Therefore, such composite membranes have an ability to operate with more stability under fuel cell operating conditions that may vary substantially.

Both polyazomethines had a strong influence on the electrochemical properties of the investigated fuel cells. Because of modification process, polyazomethine fuel cells exhibited an improved maximum power density (192 and 231 mW cm⁻²) in comparison to reference samples (95 and 140 mW cm⁻²). In addition, the energy conversion efficiency at this point improved from 27% to 33%. These achievements were obtained due to the interactions that occurred between the polyazomethines, the PFSA electrolyte, and the catalyst layer. The membrane conductivity was higher (94.3 and 75.2 vs. 45.9 and 54.1 mS cm⁻¹) and electrochemical surface area larger (35.0 vs. 24.6 m² g⁻¹) in comparison to those of the reference fuel cells.

Finally, we could conclude that the polyazomethines with relevant chemical structure seem to be very promising for usage in PEMFCs, while taking into consideration their chemical oxidation stability and good thermal properties.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

A. I. expresses her gratitude to The National Centre for Research and Development for a supplementary financial support under the 2012–2015 project No. PBS1/A5/27/2012. Authors thank Mr L. Gorecki for TGA experiments, and Mr K. Parafiniuk for synthesis and IR, and NMR analysis of the BAPBI and PAZs.

References

1. W. Dai, H. Wang and X.-Z. Yuan, et al., A review on water balance in the membrane electrode assembly of proton exchange membrane fuel cells, Int. J. Hydrogen Energy, 2009, 34, 9461–9478.
2. U. Reimer, J. Ehlert, H. Janßen and W. Lehnert, Water distribution in high temperature polymer electrolyte fuel cells, Int. J. Hydrogen Energy, 2016, 41, 1837–1845.
3. J. Mishler, Y. Wang, P. P. Mukherjee, R. Mukundan and R. L. Borup, Subfreezing operation of polymer electrolyte fuel cells: ice formation and cell performance loss, Electrochim. Acta, 2012, 65, 127–133.
4. A. D. Santamaria, J. Bachman and J. W. Park, Cold-start of parallel and interdigitated flow-field polymer electrolyte membrane fuel cell, Electrochim. Acta, 2013, 107, 327–338.
5. H. Nara, T. Momma and T. Osaka, Impedance analysis of the effect of flooding in the cathode catalyst layer of the polymer electrolyte fuel cell, Electrochim. Acta, 2013, 113, 720–729.
6. H. Li, Y. Tang and Z. Wang, et al., A review of water flooding issues in the proton exchange membrane fuel cell, J. Power Sources, 2008, 178, 103–117.
7. A. Iwan, M. Malinowski, A. Sikora, I. Tazbir and G. Pasciak, Studies of imidazole and bibenzimidazole influences on electrochemical properties of polymer fuel cells, Electrochim. Acta, 2015, 164, 143–153.
8. M. Malinowski, A. Iwan, K. Parafiniuk, L. Gorecki and G. Pasciak, Electrochemical properties of PEM Fuel Cells based on Nafion–polybenzimidazole–imidazole hybrid membranes, Int. J. Hydrogen Energy, 2015, 40, 833–840.
9. A. Iwan, M. Malinowski and A. Sikora, et al., Nafion-115/aromatic poly(etherimide) with isopropylidene groups/imidazole membranes for polymer fuel cells, J. Appl. Polym. Sci., 2015, 132, 42436.
10. M. Zhu, Y. Song and W. Hu, et al., SPAEK-based binary blends and ternary composites as proton exchange membranes for DMFCs, J. Membr. Sci., 2012, 415, 520–526.
11. E. W. Neuse and M. S. Loonat, Two-stage polybenzimidazole synthesis via poly(azomethine) intermediates, Macromolecules, 1983, 16, 128–136.
12. T. Ossiander, M. Perchthaler and C. Heinzl, et al., Influence of membrane type and molecular weight distribution on the degradation of PBI-based HTPEM fuel cells, J. Membr. Sci., 2016, 509, 27–35.
13. F. Mack, K. Aniol, C. Ellwein, J. Kerres and R. Zeis, Novel phosphoric acid-doped PBI-blends as membranes for high-temperature PEM fuel cells, J. Mater. Chem. A, 2015, 3, 10864–10874.
14. J. Yang, C. Liu, L. Gao, J. Wang, Y. Xu and R. He, Novel composite membranes of triazole modified graphene oxide and polybenzimidazole for high temperature polymer electrolyte membrane fuel cell applications, RSC Adv., 2015, 5, 101049–101054.
15. A. Iwan, B. Boharewicz and I. Tazbir, et al., New environmentally friendly polyazomethines with thiophene rings for polymer solar cells, Sol. Energy, 2015, 117, 246–259.
16. A. Iwan, M. Palewicz and I. Tazbir, et al., Influence of ZnO: Al, MoO₃ and PEDOT: PSS on efficiency in standard and inverted polymer solar cells based on polyazomethine and poly(3-hexylthiophene), Electrochim. Acta, 2016, 191, 784–794.
17. A. Iwan, B. Boharewicz, I. Tazbir and M. Filapek, Enhanced power conversion efficiency in bulk heterojunction solar cell based on new polyazomethine with vinylene moieties and [6,6]-phenyl C₆₁ butyric acid methyl ester by adding 10-camphorsulfonic acid, Electrochim. Acta, 2015, 159, 81–92.
18. D. Aili, M. K. Hansen, C. Pan, Q. Li, E. Christensen, J. O. Jensen and N. J. Bjerrum, Phosphoric acid doped membranes based on Nafion®, PBI and their blends – Membrane preparation, characterization and steam electrolysis testing, Int. J. Hydrogen Energy, 2011, 36, 6985–6993.
19. J. Yang, Q. Che, L. Zhou, R. He and R. F. Savinell, Studies of a high temperature proton exchange membrane based on incorporating an ionic liquid cation 1-butyl-3-methylimidazolium into a Nafion matrix, Electrochim. Acta, 2011, 56, 5940–5946.
20. Z. Liang, W. Chen, J. Liu, S. Wang, Z. Zhou, W. Li, G. Sun and Q. Xin, FT-IR study of the microstructure of Nafion® membrane, J. Membr. Sci., 2004, 233, 39–44.
21. A. Alentiev, J. Kostina and G. Bondarenko, Chemical aging of Nafion: FTIR study, Desalination, 2006, 200, 32–33.
22. A. Gruger, A. Regis, T. Schmatko and P. Colomban, Nanostructure of Nafion® membranes at different states of hydration: An IR and Raman study, Vib. Spectrosc., 2001, 26, 215–225.
23. G. M. Geise, H. B. Park, A. C. Sagle, B. D. Freeman and J. E. McGrath, Water permeability and water/salt selectivity tradeoff in polymers for desalination, J. Membr. Sci., 2011, 369, 130–138.
24. C. Yang, P. Costamagna, S. Srinivasan, J. Benziger and A. B. Bocarsly, Approaches and technical challenges to high temperature operation of proton exchange membrane fuel cells, J. Power Sources, 2001, 103, 1–9.
25. Y. Fu, A. Manthiram and M. D. Guiver, Acid–base blend membranes based on 2-amino-benzimidazole and sulfonated poly(ether ether ketone) for direct methanol fuel cells, Electrochem. Commun., 2007, 9, 905–910.
26. Y. Fu, W. Li and A. Manthiram, Sulfonated polysulfone with 1,3-1H-dibenzimidazole-benzene additive as a membrane for direct methanol fuel cells, J. Membr. Sci., 2008, 310, 262–267.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7se00296c

---