Introduction

Proton exchange membrane (PEM) is a significant material in PEM fuel cells (PEMFC) [1]. Perfluorosulfonated polymers such as Nafion are crucial materials for the development of PEMFC [2]. However, the commercially used Nafion membrane will lose their proton conductivity due to water evaporation at elevated temperature [3]. Considerable research has been focused on commercially available polybenzimidazole (PBI) [4]. The proton conductivity value for pure PBI membrane is low and was obtained nearly 10−9 mS/cm. However, its proton conductivity can be enhanced by adding different inorganic acids such as phosphoric acid (PA), sulfuric acid, nitric acid, or perchloric acid. Among these acids, PA is the most extensively used due to its high boiling point, high thermal stability, and high proton conductivity even in its anhydrous form [5, 6]. The enhanced proton conductivity of PA-PBI membranes in comparison with pure PBI membranes contribute to an effective acid doping of PBI membrane and provide new proton conduction pathways. PA as an acid with excellent thermal stability and high proton conductivity is always of interest to fuel cell researchers. When the pure PBI membrane is doped with PA, PA acts as a proton conducting transporter and no water is needed for proton conduction in the membranes. The proton conductivities of PA-doped PBI (PA-PBI) membranes have been extensively investigated [7–10]. However, the mechanical properties of the PA-PBI membranes may deteriorate due to the very high PA-doping level. This has been a serious problem for the improvement of the PA-PBI membranes. Oxide-type proton conductors are very important materials for a wide range of electrochemical applications such as fuel cells or hydrogen sensors because of their promising proton conductivity at high temperatures [11]. Nanocomposite membranes are new groups of membranes which include nanoparticles such as SiO2, TiO2, ZrO2, and other compounds [12, 13].

Fe-titanates are principally found as three minerals, namely: ilmenite (FeTiO3), pseudo-brookite (Fe2TiO5), and ulvospinel (Fe2TiO4). The –OH groups of Fe2TiO5 nanoparticles provide strong hydrogen bonding sites and increase the contents of the bound to free water into the membrane matrix. Also, nanoparticles possess a negative surface potential that are influenced by surface groups [14]. The anisotropic spin-glass behavior and photoelectrochemical and gas-sensing properties of Fe2TiO5 in the form of particles, thin films, or hollow spheres were studied by different groups [15–17]. The main cations in Fe2TiO5 nanoparticles are Ti4+ and Fe3+ [18]. According to the Lewis acid–base theory, these two cations are classified as hard acids. This means that the following polar groups will react with –OH groups of water. It seems that when Fe3+ cations are placed near Ti4+ cations in the Fe2TiO5 structure, it results in intensification of acidic effect and makes strong bonds with –OH groups of water. Fe2TiO5 single-phase nanoparticles have better hydrophilic properties in contrast with both TiO2 and Fe2O3 nanoparticles [19, 20].

Lately, our group presented new PEMs based on poly(vinyl alcohol) and nanoporous silica containing phenyl sulfonic acid [21] and poly(sulfonic acid)-grafted silica nanoparticles [22] and for PEMFCs. Recently, Nafion/Fe2TiO5 nanocomposite membranes were prepared by dispersion of Fe2TiO5 nanoparticles within the commercial Nafion membranes [23]. Incorporation of Fe2TiO5 nanoparticles in Nafion matrix improved the thermal stability of Nafion membranes, which is important for operation of PEMFCs at elevated temperatures.

The main aim of the present work is to study the nanocomposite membranes based on PBI and crystalline orthorhombic Fe2TiO5 nanoparticles [24]. The membranes were physico-chemically and electrochemically characterized for use as electrolytes in PEMFCs.

Experimental

Materials

Polybenzimidazole with chemical structure of poly(2,2′-m-(phenylene)-5,5′-bibenzimidazole) (PBI), with a glass transition temperature of 425–435 °C and molecular weight of 59,000–62,000 g mol−1 was obtained from FuMa-Tech GmbH corporation. N,N-dimethylacetamide (DMAc) was purchased from Merck. Phosphoric acid (PA) was supplied by Merck. Other reagents and solvents were obtained from Sigma-Aldrich.

Synthesis of Fe2TiO5 nanoparticles

Fe2TiO5 nanoparticles were prepared according to the following procedure [24]: All of the analytical grade reagents that were used in preparation of Fe2TiO5 nanoparticles are iron acetyl acetonate, tetra-n-butyl titanate and stearic acid. Firstly, stearic acid was melted in at 73 °C. Secondly, iron acetyl acetonate was added and then dissolved to form a transparent brown solution. Thirdly, tetrabutyltitanate was added to the solution and stirred. In the fourth stage, it was cooled down at room temperature and dried in an oven to obtain dried gel. Finally, the obtained gel was calcined to obtain Fe2TiO5 nanoparticles in particle size range of 48 to 60 nm.

Ion exchange capacity

The ion exchange capacity (IEC) of PA-doped membranes was determined by the titration method. PA-doped membranes were soaked in 2 M sodium chloride solutions for 24 h at RT to ensure replacement of H+ groups with Na+ groups. Subsequently, the solution was titrated with 0.1 M sodium hydroxide.

Preparation of nanocomposite membranes

The PBI nanocomposite membranes were prepared by a solution-casting method using DMAc as a casting solvent. First, an appropriate amount of PBI powder was dissolved in DMAc under stirring at 120 °C. Then, different weight percentages of nanoparticles (2, 4, 8, 16 %) were dispersed in this solution using an ultrasonic bath. The obtained brown solution was cast onto a glass plate and solvent was removed by drying at 120 °C for 5 h. The membrane was removed from the glass plate by immersing in de-ionized water and the prepared membranes were immersed in PA. PBI-Fe2TiO5 nanocomposite membranes were named PFT. The thickness of PFT before doping PA was around 64–67 μm.

PFT x were named for x wt% of Fe2TiO5 nanoparticles in the nanocomposite membranes.

Water uptake measurements of nanoparticles

A certain amount of nanoparticles was put into a test tube in 100 % RH for 24 h which was followed by weighing. After this procedure, Fe2TiO5 nanoparticles were dried for 24 h at 80 °C, followed by another weighing. Water uptake was determined by the difference between the weights of dry and fully hydrated nanoparticles. The water uptake of the nanoparticles was calculated by the following Eq. (1).

$$ \mathrm{W}\mathrm{U}=\frac{W_{\mathrm{NP}/\mathrm{wet}}-{W}_{\mathrm{NP}/\mathrm{dry}}}{W_{\mathrm{NP}/\mathrm{dry}}}\times 100 $$
(1)

Where, W NP/wet and W NP/dry are weights of the wet and dry nanoparticles, respectively. The same procedure was repeated several times to ensure that the results are real.

Acid uptake of PFT x membranes and leaching test of PA-doped PFT x membranes

The PBI nanocomposite membranes were doped by immersion in aqueous PA (75 wt%) for 5 days. The membrane PA-doping contents were determined by weighing the membranes before and after doping of PA. In order to separate the water content from the PA doping before weighting, the membranes were dried by evaporating the water at 110 °C until the membranes’ weight did not change with time.

The PA-doped PFT x nanocomposite membranes were immersed in de-ionized hot water at 90 °C for 2 h. Then, the remaining acid was gained by a similar method as described above. Leaching test is a method for determining PA retained by the PA-doped PFT x nanocomposite membranes, which is considered one of the main degradation factors of this sort of membranes in the fuel cell after washing with hot water.

FT-IR ATR spectra

The FT-IR ATR spectra were recorded in the range 600–4000 cm−1 using a Bruker Equinox 55 (attenuated total reflectance) operated at a resolution of 4 cm−1.

Proton conductivity measurements

The impedance spectra were obtained by employing an AC impedance spectroscopy with PGSTAT303N potentiostat/galvanostat (Ecochemie). The sample membrane was immersed in aqueous PA (75 wt%) for 5 days at RT and then sealed between two platinum plates electrodes. The measurements were carried out on the potentiostatic mode. The frequencies were swept from 100 to 1 MHz recording 100 points with signal amplitude of 50 mV. The proton conductivty (σ) was calculated from the measured current resistance (R) using the equation (σ = L/RA), where, σ, L, R, A, respectively, refer to proton conductivity (S/cm), thickness (cm), resistance from the impedance data (Ω), and cross-sectional area (cm2) of the membranes.

SEM, EDX, and mechanical properties measurements

The morphology of membranes was studied using a scanning electron microscope (SEM) (JSM-5600, Jeol Co.), coupled with energy dispersive X-ray (EDX) spectroscopy. The samples were freeze-fractured in liquid N2 and then their surface was coated with gold plate before SEM observations were performed. The mechanical properties of the prepared membranes were measured by using Zwick/Roell Z030 tensile test machine. All the membranes were cut to the standard shape and all tests were performed at a crosshead speed of 10 mm/min and room temperature.

Thermal properties

The thermal degradation behavior of the PA-doped nanocomposite membranes were measured using a thermogravimetric analysis (TGA) TA Instruments 2050 system, operated from 25 to 600 °C at the heating rate of 20 °C/min in nitrogen atmosphere.

XRD analysis

X-ray diffraction analysis (XRD) of PBI-based membranes was carried out on a rotating anode Philips PW-1700 diffractometer (l = 1.5418 Å, Cu Ka).

Fuel cell tests

The catalyst was Pt–C (E-TEK, 20 wt% Pt) and the Pt loadings of anode and cathode were 0.5 mg/cm2. Pt–C/PBI/LiCl/DMAc (3.6/1/0.2/38 by wt%) catalyst solution was prepared by ultrasonic disturbing for 1 h. The Pt–C inks were loaded on to the carbon paper (Toray TGP-H-090) by a painting method and dried at 190 °C in a conventional oven to calculate catalyst loading. The catalyst coated carbon papers were then doped with PA by dipping in 10 % PA solution. The acid-doped membrane was sandwiched between two pieces of gas diffusion electrodes on each side and hot-pressed under a pressure of about 50 kg/cm2 at 130 °C for 5 min.

Results and discussion

FT-IR ATR spectra

The results of the FT-IR spectra for PBI and PFT4 nanocomposite and FT-IR spectra of neat-Fe2TiO5 are shown in Fig. 1. The PFT4 nanocomposite membranes displayed numerous absorption peaks.

Fig. 1
figure 1

FT-IR spectra of PBI and PFT4 nanocomposite membranes and Fe2TiO5

Full size image

The characteristic absorption of PA molecules in PA-PBI membranes was investigated with FT-IR analysis. The bands at 500–1300 cm−1 refer to the vibration of HPO4 2− and PA groups.

Three characteristic bands around 1090, 1008, and 970 cm−1 were respectively attributed to HPO4 2−, P-OH, and H2PO4 − stretching vibrations [9–11].

The bands at 1445 cm−1 suggest the deformation of benzimidazole “Breathing” mode of imidazole rings [25]. The peak at 1600 cm−1 was assigned to the C═C and C═N stretching groups. The absorption peaks at 2250–2500 cm−1 and 2500–300 cm−1 were owing to the stretching vibration of O–H and N+–H in the presence of PA, respectively [26].

The peak at 2900 cm−1 corresponds to the stretching vibration of aromatic C–H groups. The bands at 3195 and 3390 cm−1 were attributed to the hydrogen bonded N–H groups and non-hydrogen bonded N–H stretching groups, respectively. The peak around 3615 cm−1 suggests the O–H stretching due to absorbed water [26].

Acid uptake and proton conductivity of PA-doped membranes

The acid uptake and proton conductivity of PA- PBI membranes were obtained at 60 % and 1.4 mS/cm, respectively. Figure 2 shows the acid uptake of PFT x nanocomposite membranes. All of the PFT x nanocomposite membranes confirmed a high acid uptake compared with pure PBI membranes. After acid doping with PA, the PA-PBI membranes exhibit enhancement of proton conductivity compared with the pure PBI membrane [7, 27–32]. The enhanced acid uptake of PFT x nanocomposite membranes compared with PA-PBI membranes was attributed to the interaction of PA and Fe2TiO5 nanoparticles in PFT x nanocomposite membranes. The added nanoparticles in the nanocomposite membranes enhanced the ability to trap PA, which improved the proton conductivity of the nanocomposite membranes (Fig. 3). Hence, PFT nanocomposite membranes displayed a high acid uptake and proton conductivity compared with PA-PBI membranes. In the PA-PBI nanocomposite membranes, protons can hop between the nitrogen of benzimidazole and PA by developing benzimidazolium cation and dihydrogen phosphate anion, assisting proton conduction by a Grotthuss-type mechanism [30]. In this process, the proton hops between molecules (acid–acid, acid–water or acid–benzimidazole ring) in Grotthuss mechanism. The presence of HPO4 2− and H2PO4 − anions indicates that the proton transfer could occur according to the Grotthuss mechanism. In this mechanism, proton transfer between PA, HPO4 2−, H2PO4 −, and PBI and strong acids could form membrane complexes due to the acid–base interaction or hydrogen bonding interactions between imidazole group of PBI membrane and acid molecules [30].

Fig. 2
figure 2

Acid uptake plot of PFT x nanocomposite membranes

Full size image
Fig. 3
figure 3

Proton conductivity plots of PFT x nanocomposite membranes at RT

Full size image

At high content of Fe2TiO5 nanoparticles (>4 wt%), there was a decrease in the acid uptake of PFT x nanocomposite membranes. These results were attributed to the self-aggregate of Fe2TiO5 nanoparticles in PFT x nanocomposite membranes. Figure 4 shows the SEM images of the cross-section of PBI, PFT4, and PFT16 nanocomposite membranes. Figure 4a shows the SEM image of PBI and Fig. 4b shows that the PFT4 nanocomposite membranes are homogenous. Significant agglomerations of Fe2TiO5 nanoparticles were clearly observable in the PFT16 nanocomposite membranes (Fig. 4c). This image shows that the aggregation of Fe2TiO5 nanoparticles happened at high content of Fe2TiO5 nanoparticles. The self-aggregate of nanoparticles reduces the active surface area of the nanoparticles and consequently the membrane acid uptake and proton conductivity were decreased.

Fig. 4
figure 4

SEM micrograph of the cross-section of (a) PFT4, (b) PFT16 nanocomposite membranes

Full size image

Figure 5 demonstrated Nyquist and Bode modulus plots of PFT nanocomposite membranes (2–16 wt%) at fully hydrated state. Figure 5a shows that PFT4 nanocomposite membranes displayed the lowest resistance (highest proton conductivity) compared with other PFT nanocomposite membranes. Bode modulus plots (Fig. 5b) approved a result obtained from Nyquist plots, which exhibited lower resistance for PFT4 nanocomposite membrane corresponding to the Nyquist plots.

Fig. 5
figure 5

Nyquist (a) and Bode modulus (b) plots for PFT x nanocomposite membranes

Full size image

Water uptake measurement results of nanoparticles demonstrated that the Fe2TiO5 nanoparticles displayed a higher water uptake (10 %) compared with TiO2 nanoparticles (3 %) for equal weights of the TiO2 and Fe2TiO5 nanoparticles in 25 °C. This result shows that Fe2TiO5 nanoparticles have better hydrophilicity properties compared with TiO2 nanoparticles. The proton conductivity of PFT4 nanocomposite membranes has an intense increase compared with Nafion nanocomposite membrane [33–35] and other PBI-based nanocomposite membrane [36–41]. Table 1 shows a comparison between the proton conductivity of PFT4 nanocomposite membranes and the other works. As it can be seen, the PFT4 nanocomposite membranes show proton conductivity comparable to that of commercial Nafion membrane and significant increase in proton conductivity (78 mS/cm at 180 °C) in comparison with others. It seems that in Fe2TiO5 nanoparticles, once Fe3+ cations are located near Ti4+ cations, the acidic properties of these ions are increased. Therefore, PA ensures strong interaction with Fe2TiO5 nanoparticles in PA-doped nanocomposite membranes and as a result the proton conductivity of the PFT4 nanocomposite membranes is developed because of the increasing capability of PA trap. To evaluate the proton conductivity of the PFT x membranes in high temperature, the PFT4 and the PFT8 membranes which had much better proton conductivity than the PBI and other PFT x membranes at RT were used for evaluation of proton conductivity in different temperatures up to 180 °C. Figure 6 shows the conductivity of PBI, PFT4 and PFT8 composite membranes at dry environment up to 180 °C. As it is seen in Fig. 6, these two membranes (PFT4 and PFT8) still exhibited higher proton conductivities than the pristine PA-PBI membrane. This finding could be due to the higher doping levels achieved in these membranes.

Table 1 Specification of several of PA-doped PBI composite membranes
Full size table
Fig. 6
figure 6

Proton conductivity of PBI, PFT4, and PFT8 composite membranes at dry environment up to 180 °C

Full size image

Leaching test of PA-doped PFT x membranes

The acid leaching tests for PFT nanocomposite membranes were carried out in order to determine the PA preservation ability of the membranes. The acid leaching test is considered as one of the main degradation aspects of PBI-based membranes in the PEMFCs [42, 43]. Figure 7 displayed the results achieved from the leaching tests. The PFT4 nanocomposite membranes preserve higher amounts of the PA than the other PFT x nanocomposite membranes. This result shows that the presence of Fe2TiO5 in the PFT4 nanocomposite membranes improved the capability properties of the membrane to retain acid after being washed with hot water. The formed agglomerates for PFT16 nanocomposite membranes increase the PA leaching from the membrane than the other PFT x nanocomposite membranes.

Fig. 7
figure 7

Remaining acid plot of PFT x nanocomposite membranes

Full size image

Ion exchange capacity

The ion exchange capacity (IEC) of PFT x nanocomposite membranes is displayed in Fig. 8. IEC values of PFT nanocomposite membranes increase with the increasing of the nanoparticles content due to higher PA trap of nanoparticles, which increased the dissociable H+ ions. PFT4 nanocomposite membranes displayed high IEC (meq g−1) compared with other nanocomposite membranes. But in the high content of nanoparticles (>4 wt%) IEC values decreased due to the aggregation of nanoparticles.

Fig. 8
figure 8

Ion exchange capacity (IEC) of PFT x nanocomposite membranes

Full size image

EDX analysis and mechanical properties

EDX analysis established the presence of Fe2TiO5 in the PFT4 nanocomposite membrane. EDX distribution of nanoparticles in the cross-section of PFT4 nanocomposite membranes were demonstrated in Fig. 9a. A homogenous distribution of Fe and Ti nanoparticles in the cross-section of PFT4 nanocomposite membranes were detected.

Fig. 9
figure 9

SEM-EDX images of the cross-section of PFT4 nanocomposite membranes (a) and mechanical stability plot of PA-PBI and PFT4 membranes (b)

Full size image

From Fig. 9b, it was found that the PFT4 nanocomposite membranes, due to strong interactions of Fe2TiO5 nanoparticles with PBI-based membrane, displayed a higher mechanical stability than PBI-based membrane. Uniform dispersion of nanoparticles in the PFT4 nanocomposite membranes, which increases the PBI-nanoparticles interactions, plays also a key role in improvement of its mechanical stability.

Thermogravimetric analyses

The results of the thermal stability for the PFT x nanocomposite membranes are demonstrated in Fig. 10. From Fig. 10, it can be seen that the PFT4 nanocomposites membranes displayed higher thermal stability than other nanocomposite membranes.

Fig. 10
figure 10

Thermograms of the of PFT x nanocomposite membranes

Full size image

Incorporation of 4 wt% of Fe2TiO5 nanoparticles in PFT4 nanocomposites membranes, which is an effective way for improving the thermal stability of nanocomposites membranes, leads to an increase in the decomposition temperature of PFT4 nanocomposite membranes compared with other nanocomposite membranes.

Figure 10 shows that all the samples show two well-defined weight decays. The first goes from room temperature to 125 °C that is due to the desorption of absorbed water from polymer. And the second one, appearing at around 160 °C, is due to the thermal changes in PA, forming pyrophosphoric and triphosphoric acid, as shown by the following equations:

$$ 2{\mathrm{H}}_3{\mathrm{P}\mathrm{O}}_4\to {\mathrm{H}}_4{\mathrm{P}}_2{\mathrm{O}}_7+{\mathrm{H}}_2\mathrm{O} $$
(2)
$$ 3{\mathrm{H}}_4{\mathrm{P}}_2{\mathrm{O}}_7\to 2{\mathrm{H}}_5{\mathrm{P}}_3{\mathrm{O}}_{10}+{\mathrm{H}}_2\mathrm{O} $$
(3)

The PFT4 nanocomposite membranes show outstanding thermal stability. It also has displayed high proton conductivity and worked at high temperatures.

X-ray diffraction analyses PFT x membranes

The X-ray patterns of the PFT nanocomposite membranes are shown in Fig. 11. The PFT nanocomposite membranes displayed a broad peak at around 2θ = 25°. In Fig. 11, the Fe2TiO5 X-ray pattern is shown. All the PFT nanocomposite membranes displayed the main peak that is the characteristic of Fe2TiO5 nanoparticles. The more intense the main peak, the higher the content of Fe2TiO5 will be appear. This fact confirms the presence of the Fe2TiO5 nanoparticles in the PBI membranes and that the structure of it did not vary when incorporated into the membrane.

Fig. 11
figure 11

X-ray diffraction patterns of the undoped standard PBI, PFT x nanocomposite membranes, and Fe2TiO5 nanoparticles

Full size image

Fuel cell performance tests

To evaluate the fuel cell performance of the PFT x membranes, the PFT4 membrane was used to prepare MEAs and the fuel cell performance test was carried out. The membrane thickness was around 65 μm. The PEMFC unit cell performance of these MEAs were tested at 100, 150, and 180 °C under ambient pressure with non-humidified H2/O2 flows. The flow rates for both hydrogen and oxygen gases were kept as 300 ml/min and 500 ml/min, respectively. Table 2 summarizes the PEMFC open circuit voltages (OCVs), power density in 0.5 V and the current density in 0.5 V. The PFT4 composite membrane showed 380 mW/ cm2 power density and 760 mA/cm2 current density in 0.5 V at 180 °C at dry condition. Polarization curves were obtained using a fuel cell evaluation system (FCT—150 s). Figure 12 shows polarization curves for a fuel cell based on the PFT4 nanocomposite membranes obtained at different temperatures (RT—180 °C) and PBI membranes obtained at 180 °C.

Table 2 Fuel cell performance for unit cell based on PFT4 nanocomposite membranes
Full size table
Fig. 12
figure 12

Polarization curves for a fuel cell based on the PFT4 nanocomposite membranes obtained at different temperatures (25–180 °C) and PBI membranes obtained at 180 °C

Full size image

Conclusion

In the present study, new and advanced PBI-Fe2TiO5 nanocomposite membranes were prepared by solution-casting method. The results showed that the acid uptake and proton conductivity of the PFT nanocomposite membranes were higher than that of PBI-based membrane because of the unique properties of Fe2TiO5 nanoparticles. The PA-doped PFT4 nanocomposite membrane showed a higher proton conductivity specially at high temperatures up to 180 °C (78 mS/cm) compared with the PA-doped PBI and showed 380 mW/cm2 power density and 760 mA/cm2 current density in 0.5 V and 180 °C at dry condition. These results indicated that the PBI nanocomposite membranes based on Fe2TiO5 could be utilized as good candidates as proton exchange membranes for high-temperature PEMFCs.