An efficient toluene barrier membrane for high-performance direct toluene hydrogenation via an electrochemical process

Abstract

Direct electro-hydrogenation of toluene to methyl cyclohexane (MCH) is gaining attention as a green process to store hydrogen in a liquid organic hydrogen carrier (LOHC), ultimately for long-distance hydrogen storage and transportation. A critical challenge in this electrochemical process is preventing toluene crossover to maintain high conversion efficiency and cell performance. This study introduces a novel approach utilizing sulfonated poly(arylene ether sulfone) (SPAES), a hydrocarbon-based proton exchange membrane (PEM) with narrow hydrophilic domains, to significantly reduce toluene diffusivity. Our findings reveal that toluene diffusivity in the SPAES PEM is 19.6-fold lower than in commercially available Nafion, resulting in a 60% reduction in toluene permeability. The enhanced barrier properties of the SPAES PEM substantially improved the Faradaic efficiency of toluene hydrogenation to 72.8%, whereas Nafion achieved 68.4% at a high current density of 600 mA cm⁻². Long-term operation (48 hours) at 150 mA cm⁻² demonstrated the superior performance of the SPAES PEM, with a degradation rate of 728 μV h⁻¹ compared to Nafion's 1270 μV h⁻¹. This research elucidates the effects of toluene crossover in direct electro-hydrogenation electrolyzers and demonstrates the advantages of hydrocarbon-based PEMs for LOHC electro-conversion applications.

---

Introduction

As global interest in carbon neutrality increases, renewable energies have become essential considerations in planning both current and future energy systems.^1,2 However, their uneven distribution between regions and intermittent supply limit large-scale utilization.^3–5 In this context, hydrogen, an electrochemically convertible chemical, is considered as an energy carrier to address the limitations of renewable energy sources and as a fuel for electricity generation.^6,7

Hydrogen itself, however, has limitations in transportation and storage due to its low mass energy density (33.33 kW h kg⁻¹), volume energy density (0.003 kW h L⁻¹), and volume density (0.089 kg m⁻³).^8,9 To achieve more efficient hydrogen storage, various physical and chemical approaches have been adopted.^10,11 Among them, liquid organic hydrogen carriers (LOHCs) are of particular interest as compounds that store hydrogen through the chemical process of hydrogenation.^12–16

LOHCs represent an attractive solution for large-scale hydrogen storage, offering advantages over other methods such as hydrogen compression, liquefaction, and chemical adsorption and absorption.^15–19 These traditional methods often require significant additional expenses for specialized materials and infrastructure.²⁰ Moreover, they frequently demand challenging physical conditions, such as extremely high pressures or very low temperatures.²⁰ In contrast, LOHCs can operate under more moderate conditions, potentially simplifying both the storage and handling processes of hydrogen. Notably, LOHCs can store hydrogen in small organic molecules, providing a high hydrogen storage capacity per unit volume and weight, while minimizing the risk of unexpected hydrogen leakage and environmental impact during the hydrogenation and dehydrogenation cycles.^15,16

Initial research on potential candidates for hydrogen storage and transportation via LOHCs has focused on pairs such as decalin/naphthalene,^21–23 benzene/cyclohexane,^24,25 and toluene/methyl cyclohexane (MCH).^26,27 While the decalin/naphthalene and benzene/cyclohexane LOHC systems may theoretically approach the hydrogen storage target proposed by the U.S. Department of Energy (DOE) (7.5 wt%), there are limitations in handling of these pairs.²⁸ Specifically, decalin and benzene may exceed acceptable levels of environmental toxicity during transportation and usage.²⁹ In contrast, the toluene/MCH system is relatively low in toxicity and maintains a liquid state at ambient temperature and pressure due to its low vapor pressure.^30,31 Additionally, it shares similarities with gasoline, allowing for the utilization of the existing infrastructure used for petroleum transportation.^32,33 This makes it an efficient LOHC choice and more suitable for commercialization and widespread application compared to other LOHC candidates.

Conventionally, toluene undergoes hydrogenation to form MCH in the presence of gaseous hydrogen over transition metals including Ni,³⁴ Pt,³⁵ Ru,^36,37 and Rh.^38,39 This conventional process typically follows the hydrogen production process, such as water electrolysis. However, recent literature suggests that toluene can be effectively hydrogenated through a single-step electrochemical process, bypassing the conventional two-stage process.^40,41 In this process, the direct electro-hydrogenation of toluene (DEHT) at the cathode occurs concurrently with the oxygen evolution reaction at the anode, allowing for direct hydrogenation of toluene at a thermodynamic potential of 1.08 V at 25 °C and 1 atm.⁴² At the cathode, toluene is directly reduced to methyl cyclohexane with protons generated and transferred from the anode (reaction (1)), where water is electrolyzed into oxygen and protons (reaction (2)).

C₆H₅CH₃ (l) + 6H⁺ + 6e⁻ → C₆H₁₁CH₃ (l) (1)

3H₂O (l) → 3/2O₂ (g) + 6H⁺ + 6e⁻ (2)

Similar to the water electrolysis cells, electrochemical cells for DEHT employ a proton exchange membrane (PEM) to effectively conduct protons while preventing the crossover of reactants and products to the other electrode, thereby enhancing the cell efficiencies. However, toluene crossover from the cathode to the anode poses a significant problem by affecting anode catalytic activity, which ultimately impacts Faradaic efficiency. Nagai et al. investigated the oxygen evolution reaction (OER) current density with and without saturated toluene in 1 M H₂SO₄.⁴³ Although they demonstrated that a composite OER catalyst of Ir–Ta–Zr is less affected by toluene contamination compared to an IrO₂ catalyst, they found that the OER current density decreased when the catalysts were contaminated with toluene. This suggests that toluene permeation from the cathode side to the OER catalysts hinders toluene hydrogenation at the cathode side, as protons are provided by the OER, as depicted in reaction (2). From this perspective, the toluene barrier property of the PEM is a critical factor in DEHT application. However, no studies have yet focused on enhancing DEHT performance by reducing toluene permeability of the PEM. Nagasawa et al. investigated the toluene permeability of perfluorinated sulfonic acid (PFSA) membranes (Nafion and Aquivion),⁴⁴ but their primary focus was on the toluene permeation effects of PFSA ionomers in the catalyst layers rather than in PEMs. Moreover, the long-term DEHT performance with respect to the PEM type remains unexplored.

Here, to reduce the toluene permeability of PEMs, we tailored the hydrophilic channel size of a hydrocarbon-based PEM to approximately 2.1 nm, which is about half that of the commercial Nafion PEM (4.5 nm), using synthesized sulfonated poly(arylene ether sulfone) (SPAES). This approach resulted in lower toluene diffusivity within the hydrophilic domains and led to a 60% reduction in both in situ and ex situ toluene permeability for the SPAES PEM. We demonstrated the advantages of this hydrocarbon-based PEM with low toluene permeability for toluene electro-conversion applications by evaluating Faradaic cell efficiency and degradation rates during long-term operation at 150 mA cm⁻². These findings suggest that hydrocarbon-based PEMs represent a promising alternative to conventional PEMs in toluene electro-conversion systems, potentially offering improvements in efficiency and durability.

Materials and methods

Materials

4,4′-Difluorodiphenyl sulfone (DFDPS, ≥99.9%) and 4,4′-dihydroxybiphenyl (BP, ≥99.9%) were purchased from Richem. Sulfonated DFDPS (SDFDPS, ≥99.9%) was synthesized using 65% fuming sulfuric acid (Merck).⁴⁵ Potassium carbonate (K₂CO₃, ≥99.9%), N-methyl-2-pyrrolidinone (NMP, anhydrous ≥99.9%), toluene (anhydrous ≥99.8%), and methyl cyclohexane (anhydrous ≥99%) were purchased from Sigma-Aldrich. Iridium oxide (IrO₂, ≥99.9%) and PtRu/C (Pt: 31 wt%, Ru: 19 wt%, and C: 50 wt%) were supplied by Boyaz Energy and KORENS RTX, respectively. A sulfuric acid solution was prepared by mixing sulfuric acid (Sigma-Aldrich, 95–97%) with deionized water (18.2 MΩ cm at 25 °C).

Synthesis of SPAES

SPAES was synthesized through step-growth polymerization of the monomers, BP, DFDPS, and SDFDPS as shown in Fig. 1. The molar ratio of BP : (DFDPS + SDFDPS) : K₂CO₃ was fixed at 1 : 1 : 1.21. To adjust the degree of sulfonation in SPAES to 40, 50, and 60%, which corresponds to the values of m and n in SPAES copolymers, the input molar ratios of DFSPS : SDFDPS were varied from 6 : 4 to 5 : 5 and 4 : 6, respectively. The synthesis process proceeded as follows: initially, BP (0.0369 mol) and K₂CO₃ (0.0446 mol) were added to a 100 mL round-bottom flask equipped with a mechanical stirrer containing NMP (40 mL) and toluene (40 mL) under a nitrogen atmosphere. The temperature was gradually raised from room temperature to 155 °C over 2 hours and maintained for 3.5 hours to remove water via reflux. Subsequently, the temperature was increased to 185 °C over 2 hours to eliminate toluene. Afterward, DFDPS and SDFDP (0.0369 mol in total), along with NMP (40 mL), were introduced into the mixture, and the temperature was further elevated to 195 °C. The reaction was continued at 195 °C for 20 hours until a viscous solution was obtained. Finally, the resulting viscous solution was precipitated in water, washed several times with isopropanol, and the SPAES ionomers were dried at 80 °C for 24 hours.

Fig. 1 Synthetic scheme of SPAES copolymers.

SPAES PEM preparation

The Na⁺-form SPAES powders were dissolved in NMP at a concentration of 15 wt%. The resulting solution was poured onto a clean glass plate and spread using a doctor blade. The cast solution was dried in a convection oven at 80 °C for 12 hours to ensure complete evaporation of NMP. Subsequently, the dried membrane was immersed in deionized water to facilitate its removal from the glass plate. The detached SPAES membrane was then soaked in 1.5 M sulfuric acid for 24 hours to exchange Na⁺ ions with H⁺ ions. Following this ion exchange process, the membrane was thoroughly rinsed multiple times with deionized water to remove any residual sulfuric acid. Finally, the PEMs were dried under vacuum at 80 °C for at least 24 hours before use.

Proton conductivity and solvent uptake of PEMs

The proton conductivity (σ) of PEMs was measured by electrochemical impedance spectroscopy (EIS, Solartron 1280 AC) at 25 °C in water using a 4-point probe cell with platinum electrodes. A 10 mV voltage was applied over 0.1 Hz to 20 kHz, and resistance (r) values were used to calculate the proton conductivity. The proton conductivity was calculated using the formula: σ = d/(w × l × r), where d, w and l represent the distance between the electrodes, width of the sample and thickness of the sample, respectively.

The solvent uptake (ΔU) of the membrane was evaluated using the masses of the dried membrane (m_d) and the swollen membrane (m_w) according to the following equation: ΔU = (m_w − m_d)/m_d.

Toluene permeability of PEMs

The permeability of toluene through a PEM was assessed using a diffusion cell with two chambers, separated by a PEM having an effective area of 2.84 cm². One chamber was filled with 100 mL of 1 M H₂SO₄, while the other contained 100 mL of pure toluene. Both chambers were continuously stirred at 25 °C using a magnetic stirrer. The toluene concentration was determined using a UV-VIS spectroscopy (Cary 8454 UV-Vis, Agilent Technologies), monitoring the absorption peak of toluene at 200 nm. The peak intensity was converted to concentration using a calibration curve correlating absorbance with toluene concentration in a 1 M H₂SO₄ solution.

Pulsed-field gradient nuclear magnetic resonance (PFG-NMR) for toluene diffusion

The diffusivity of toluene in the PEMs was measured using an NMR spectrometer (Bruker AVANCE III 500 MHz) equipped with a 5 mm triple resonance broad band probe and a temperature controller. The PEMs (40 × 3 mm²) were immersed in toluene for at least 24 hours. Subsequently, any excess toluene on the swollen PEMs was carefully wiped away before placing them into an NMR tube. A flame-sealed cylindrical glass tube containing deuterated dimethyl sulfoxide (DMSO)-d₆ was then introduced into the NMR tube. Each NMR tube was meticulously sealed to prevent toluene evaporation from the PEMs, and all NMR measurements were conducted at 25 °C.

PFG-NMR analyses were performed on the methyl proton peaks of toluene (1.4–1.7 ppm) in the PEMs using the diffusion-ordered spectroscopy (DOSY) method, employing the ‘ledbpgp2s’ pulse sequence (longitudinal eddy current delay experiment utilizing bipolar gradients acquired in 2D mode). The pulse sequence involved the application of two gradient pulses with a strength (G) varying from 2% to 98% of the maximum gradient strength (0.47 T m⁻¹) for a duration (δ). The attenuation intensity (I) of the proton peaks from toluene was observed during the diffusion time (Δ), and data processing was performed using the Top Spin software package to derive the translation diffusion coefficient (D) of toluene in the PEMs. The attenuated I compared to the intensity at G = 0 (I₀) follows as:

I/I₀ = exp(−γ²δ²G²D(Δ − δ/3)), (3)

where γ represents the proton gyromagnetic ratio (42.6 MHz T⁻¹). The D value can be determined from the slope of the linear graph of ln(I/I₀) versus (−γ²δ²G²(Δ − δ/3)).

Cross-sectional PEM morphology

The cross-sectional PEM morphology was examined using a TALOS F200X transmission electron microscope (TEM). To prepare the membrane for imaging, it was soaked in a 1.0 M aqueous solution of lead(II) acetate for 24 hours to stain the hydrophilic regions. After staining, the membrane was rinsed with deionized water and then soaked in deionized water for an additional 24 hours. Following vacuum drying of the membrane at 80 °C for 24 hours, it was embedded in epoxy resin and sectioned to a thickness of 80 nm using an ultramicrotome diamond knife (LEIGA EM UC6). The sections were then mounted on copper grids for TEM images.

Fabrication of catalyst inks and the electrochemical cell

For the anode catalyst ink, 35.3 wt% IrO₂, 21.3 wt% deionized water, 23.7 wt% n-propanol, and 19.7 wt% Nafion ionomer solution (D2021, Chemours) were mixed and stirred for 6 hours. Similarly, the cathode catalyst ink was prepared by mixing 3.8 wt% PtRu/C (Pt: 31.0 wt%, Ru: 19.3 wt%, C: 50.0 wt%), 39.2 wt% n-propanol, 32.1 wt% deionized water, and 24.9 wt% Nafion ionomer solution (D521, Chemours) for 6 hours. Note that all catalyst inks were sonicated to achieve uniformity. For the anode catalyst, the cast and dried ink was decal-transferred to a PEM using a hot-press machine at 58.8 MPa and 130 °C for 10 min. For the cathode catalyst, the ink was sprayed onto a gas diffusion layer (GDL), specifically carbon paper (39BC, SIGRACET^®), using a sprayer to prepare the cathode catalyst-coated GDL. The loading amount of transferred IrO₂ was 2 mg cm⁻², and that of PtRu was 1.4 mg cm⁻². The Pt-coated Ti-porous transport layer (Ti-PTL, LT Metal Ltd) on the anode side, IrO₂-coated PEM, and PtRu/C-coated GDL were placed between a pair of flow plates, current collectors, and end plates to form an electrochemical cell for the direct toluene hydrogenation.

Electrochemical cell performance and impedance

The anode and cathode of the electrolysis cell were supplied with 1 M H₂SO₄ and pure toluene at a flow rate of 30 mL min⁻¹, respectively. The cell performance was measured by scanning the voltage from 1.3 to 2.0 V under ambient pressure, with the voltage being increased incrementally by 0.05 V every 30 seconds. The current data measured in response to the applied voltage were recorded every 10 seconds. The cell temperature was maintained at 25 °C. EIS was performed at a voltage of 1.5 V over a frequency range of 30 kHz to 50 MHz, with an amplitude of 5 mV.

Results and discussion

As the input molar ratio of hydrophobic to hydrophilic monomers was varied from 6 : 4 to 5 : 5 and 4 : 6, the resulting SPAES ionomers exhibited degrees of sulfonation (DS) of 37.3%, 48.3%, and 57.2%, respectively, as evaluated by ¹H NMR (Fig. S1, see the ESI† for details on DS calculation). Based on these DS values, the SPAES PEMs were named SPAES40, SPAES50, and SPAES60, respectively. All SPAES series ionomers exhibited higher molecular weight over 140 kg mol⁻¹ (Table S1†). Note that all membrane thicknesses were approximately 50 μm, similar to that of commercial Nafion 212 (48.3 μm); specifically, the thicknesses of SPAES40, SPAES50, and SPAES60 were 48.5, 45.0, and 46.4 μm, respectively.

For the toluene crossover measurements, three SPAES PEMs and Nafion 212 were positioned between the chambers containing pure toluene and a 1 M H₂SO₄ aqueous solution in a diffusion cell, as schematically illustrated in Fig. 2a. The transient concentration of toluene on the toluene-deficient side was tracked by monitoring its characteristic absorbance peak at 200 nm and compared to that of the commercial Nafion. The calibration curve was obtained at the saturated concentration of toluene in water at 25 °C (Fig. S2†), measured as 6.95 mmol L⁻¹, which is similar to literature values.^44,46 Note that the saturated concentration of MCH in water at 25 °C was measured as 0.47 mmol L⁻¹.⁴⁷ As illustrated in Fig. 2b, toluene permeated through the PEMs, resulting in a continuous increase in concentration over time. Notably, Nafion 212 exhibited a higher rate of concentration increase compared to the other SPAES PEMs. For example, after 7 hours, the toluene concentrations permeated through Nafion 212, SPAES40, SPAES50 and SPAES60 were 4.8, 2.2, 2.8 and 4.0 mmol L⁻¹, respectively, as shown in the UV spectra (Fig. S3†).

Fig. 2 (a) Schematic illustration of a two-chamber diffusion cell for toluene permeation through the PEM. (b) The changes in toluene concentration over time in the toluene deficient side (1 M sulfuric acid solution). (c) Toluene permeability in a two-chamber diffusion cell with PEMs. (d) Toluene uptake of PEMs measured after soaking in toluene for 24 hours. (e) Intensity decay curves of toluene within the PEMs with variation of the field gradient strength G for different diffusion times Δ and durations δ.

Given that the membrane thicknesses of the PEMs were similar, approximately 50 μm, the permeability values are expected to correlate with the rates of concentration increase mentioned earlier. Assuming that the partitioning of toluene into the PEMs (toluene solubility in the membrane on both the toluene-rich and toluene-deficient sides) does not vary significantly, the toluene permeability (P) through the PEMs can be evaluated using the following equation (see the ESI† for details):

C_S = C_T (1 − exp (−(P × A)/(l × V) × t)), (4)

where C_S and C_T represent the toluene concentrations in the H₂SO₄ solution and toluene side, respectively; A is the effective area for toluene permeation; l is the membrane thickness; V is the volume of the H₂SO₄ solution; and t is the permeation time. By fitting the transient concentration data (Fig. 2b) to eqn (4), the P values were obtained. As expected, Nafion 212 exhibited the highest P value, P = 2.0 × 10⁻⁷ m² s⁻¹, while the SPAES PEMs showed lower P values, as summarized in Fig. 2c. Specifically, SPAES50 had a P value of 0.8 × 10⁻⁷ m² s⁻¹, approximately 2.5 times lower than that of Nafion 212. Notably, no permeated toluene was detected across the PTFE film (Fig. S4†), which has the same chemical structure as the hydrophobic backbone of Nafion. This observation indicates that, despite toluene being non-polar,^48,49 it primarily permeates through the hydrophilic domains rather than the hydrophobic domains. Supporting this conclusion, the P values and toluene uptake for the SPAES PEMs increased with the DS (Fig. 2c and d), suggesting that greater sulfonation enhances the hydrophilic character and toluene permeability of the membrane.

To figure out the reduction in toluene permeability of the SPAES PEMs compared to Nafion 212, the toluene uptake of the PEMs and diffusivity in the PEMs-key factors determining toluene permeation according to the solution-diffusion model were investigated.^50,51 In contrast to the permeability values, Nafion 212 absorbed less toluene than the SPAES PEMs, as shown in Fig. 2d. The toluene uptake of Nafion 212 was 2.0 wt%, while SPAES40, SPAES50 and SPAES60 exhibited uptakes of 2.3, 2.6 and 6.3 wt%, respectively. Therefore, the lower toluene permeability through the SPAES PEMs compared to Nafion 212 is primarily attributed to reduced toluene diffusivity rather than solubility. Indeed, the toluene diffusivity in SPAES50 was D = 3.1 × 10⁻¹² m² s⁻¹, which is 19.6 times lower than that of Nafion 212 (D = 6.13 × 10⁻¹¹ m² s⁻¹), as obtained from the slopes of the attenuated PFG-NMR intensities using eqn (3) (Fig. 2e). As mentioned above, toluene primarily permeates through the hydrophilic domains, which means that the tortuosity and width of the hydrophilic domains are important factors determining the effective diffusivity.⁵² As shown in the cross-sectional TEM images (Fig. S5†), SPAES50 has narrower hydrophilic channels compared to Nafion 212. In the TEM images, the dark regions correspond to hydrophilic domains, while the bright regions represent hydrophobic domains. SPAES50 exhibits a hydrophilic domain width of 2.1 ± 0.4 nm, which is smaller than the 4.5 ± 0.3 nm width of Nafion 212. Additionally, SPAES50 shows more tortuous hydrophilic channels, which further limit the permeation of toluene. Therefore, the narrower and more tortuous hydrophilic domains in SPAES50 contribute significantly to the reduced toluene diffusivity.

On the other hand, the permeability differences observed among the SPAES40, SPAES50, and SPAES60 membranes can be attributed to the preferential permeation of toluene through the hydrophilic rather than the hydrophobic domains, as discussed in the case of Nafion 212 and PTFE (Fig. S3a and S4†), as well as to the geometric factors of the hydrophilic channels. As the DS increases, the volume fraction of the hydrophilic component in the copolymer also increases, which is reflected in the water uptake (Fig. S6†), and consequently, the toluene uptake also increases (Fig. 2d). SPAES40, which has the lowest DS, exhibited the lowest water and toluene uptake among the SPAES series. In terms of morphology, as the DS value increases, the size of the hydrophilic domains decreases (1.8, 2.1 and 3.5 nm for SPAES40, SPAES50 and SPAES60, respectively), as shown in the cross-sectional TEM images (Fig. S5†), thereby reducing the available pathways for toluene diffusion. Consequently, SPAES membranes with higher DS values exhibit higher toluene permeability (Fig. 2c).

Based on proton selectivity to toluene, defined as the ratio of proton conductivity (σ) to toluene permeability (P), a criterion for evaluating suitable PEMs for DEHT, SPAES50 was selected to compare the DEHT cell performance with the commercial Nafion 212. The measured proton conductivities of SPAES40, SPAES50 and SPAES60 at 25 °C in water were 64, 102 and 125 mS cm⁻¹, respectively. As the DS increases, the hydrated channels become wider and more continuous, facilitating proton movement and resulting in an increase in proton conductivity (Table S1†). Among the SPAES series, SPAES50 exhibited the highest selectivity value of 12.6 × 10⁴ mS s cm⁻³, whereas SPAES40 and SPAES60 displayed values of 9.9 × 10⁴ and 8.7 × 10⁴ mS s cm⁻³, respectively. Note that the selectivity of Nafion 212 was 3.4 × 10⁴ mS s cm⁻³, which is much lower than that of the SPAES series, with σ = 68 mS cm⁻¹.

Therefore, to assess the effect of toluene permeability on DEHT performance, cells with Nafion 212 and SPAES50 were prepared and compared using polarization curves and cell voltages during long-term operation (48 hours) at a constant current density of 150 mA cm⁻², where the transported protons are involved in DEHT rather than the hydrogen evolution reaction. As illustrated in Fig. 3a, toluene molecules permeated through the PEM from the cathode to the IrO₂-based anode and adsorbed on it,⁵³ diminishing the OER activity, as evidenced by a reduction in the double-layer capacitance (C_dl) of IrO₂ in the presence of toluene.⁴³ It should be noted that C_dl is proportional to the electrochemical active surface area, which represents the surface area of the electrode that participates in electrochemical reactions.⁵⁴ As shown in Fig. 3b, at the beginning of the test (BOT), Nafion 212 and SPAES50 show similar polarization curves. At 2 V, both the PEMs exhibited about 2.00 A cm⁻². However, after the constant current operation, referred to as the end of the test (EOT) in Fig. 3b, a more noticeable reduction in current density was observed in Nafion 212 compared to SPAES50. At EOT, while the current density for SPAES50 remained at 1.96 A cm⁻², that of Nafion 212 decreased from 2.00 to 1.45 A cm⁻² at 2 V.

Fig. 3 (a) Schematic illustration of the OER suppression of IrO₂ catalyst due to toluene crossover. (b) The polarization curves of Nafion and SPAES50 PEMs at the beginning of test (BOT) and end of test (EOT) at a constant current density of 150 mA cm⁻² for 48 hours. (c) The voltage change profiles and evaluated degradation rates of Nafion and SPAES50 PEMs at a constant current density of 150 mA cm⁻² for 48 hours. (d) The Nyquist plots of the EIS measurements (30 mHz–50 kHz) for Nafion and SPAES50 before and after applying a constant current density of 150 mA cm⁻² for 48 hours. (e) Ohmic resistance (R_Ω) and charge transfer resistance (R_anode and R_cathode) of Nafion and SPAES50 PEMs evaluated based on an equivalent circuit model. (f and g) UV absorbance peaks caused by toluene concentration changes in the sulfuric acid compartment of the DEHT cell with Nafion and SPAES50 PEMs during electrochemical reactions at a constant current density of 150 mA cm⁻² for 48 hours. (h) Quantification of toluene concentration changes in the sulfuric acid compartment of the DEHT cell with Nafion and SPAES50 PEMs over time based on UV absorbance peaks.

The performance of SPAES50, the first hydrocarbon-based PEM adopted for DEHT, at BOT and EOT was compared to that of PFSA-based PEMs (Nafion and Aquivion) reported in previous literature.^44,55,56 Current densities at 1.65 V, obtained from these studies, were used for comparison. Only results with 100% toluene feed for the cathode were included for consistency. As shown in Fig. S7,† SPAES50 at BOT demonstrated comparable performance to Nafion 212 in this study and outperformed the PFSA PEMs reported in the literature. Importantly, this is the first study to evaluate the long-term performance of PEMs in DEHT. SPAES50 exhibited better stability and performance than Nafion 212, even at EOT, with minimal loss in current density during extended operation. Considering that all other components were the same except for the PEM, the significant decrease in cell performance with Nafion 212 can be attributed to its poor toluene barrier properties. In the voltage profile during the long-term test, the degradation rate, defined as the rate of voltage change (dV/dt), was 1.7 times higher for Nafion 212 (1270 μV h⁻¹) than for SPAES50 (728 μV h⁻¹), as shown in Fig. 3c.

To evaluate the stability of SPAES50 during long-term DEHT operation, its chemical structure and DS were assessed by ¹H NMR, and its molecular weight was measured by gel permeation chromatography (GPC). As summarized in Fig. S8 and Table S2,† no significant changes were observed in the chemical structure or molecular weight, confirming the chemical stability of SPAES50 during long-term DEHT operation.

In general, the voltage increase during the long-term test is closely associated with the overpotentials arising from ohmic,^57,58 charge transfer,^59,60 and mass transport resistances.^61,62 To investigate the detailed contributions of these resistances to performance degradation during DEHT, EIS measurements were conducted at BOT and EOT to obtain Nyquist plots. The high-frequency intercept of the Nyquist plot indicates the ohmic resistance (R_Ω), while the subsequent arc in the lower frequency region reflects charge transfer resistance arising from oxidation and reduction reactions.^63,64 However, as shown in Fig. 3b and d, no clear contribution of mass transport resistance to the polarization curves, particularly in the high current density region, and to the Nyquist plots was observed. After the long-term test, R_Ω of Nafion 212 and SPAES50 remained nearly unchanged. However, a significant increase in the arc size was observed for Nafion 212 compared to SPAES50, indicating a more pronounced reduction in catalytic activity for Nafion 212. As mentioned in the introduction, the negative effects of toluene crossover from the cathode to the anode may explain this observed reduction in activity. Toluene permeation from the cathode contaminates the anodic IrO₂ catalyst, leading to catalyst poisoning, which impedes long-term operation. Furthermore, toluene inhibits the OER at the anode, disrupting proton supply to the cathode and reducing the efficiency of toluene hydrogenation.⁴³

To differentiate the contributions of the anode (R_anode) and cathode (R_cathode) charge transfer resistances, Nyquist plots were fitted to an equivalent circuit model,⁶³ allowing for the evaluation of each resistance component: R_Ω, R_anode, and R_cathode. In Fig. 3e, these resistance components are summarized, and the values for Nafion 212 and SPAES50 are compared at BOT and EOT. While R_Ω and R_cathode showed no significant changes for either PEM, R_anode increased more significantly for the highly toluene-permeable Nafion 212 than for SPAES50. Specifically, R_anode for Nafion 212 increased from 713.2 to 1391.2 mΩ cm², whereas for SPAES50, R_anode showed only a modest increase from 571.3 to 672.4 mΩ cm². Based on the results, the main factor contributing to the performance reduction in DEHT appears to be an increase in kinetic overpotential, likely caused by anode catalyst poisoning from toluene crossover.

To confirm that toluene crossover is the primary cause of performance degradation, the toluene concentration on the anode side was monitored during the long-term cell performance testing. As shown in Fig. 3f and g, the absorbance peak of toluene continuously increased over time for both PEMs. Notably, on the anode side, no characteristic peaks of benzaldehyde or benzyl alcohol, potential oxidation products of toluene at the anode, were observed, aside from those of permeated toluene and MCH in the ¹H NMR spectra (Fig. S9†). Based on these results, it can be concluded that no reaction occurred between the penetrated toluene (or MCH) and IrO₂ at the anode side. By EOT, the toluene concentration reached 2.3 mmol L⁻¹ for Nafion 212 and 1.1 mmol L⁻¹ for SPAES50 (Fig. 3h), indicating that Nafion 212 exhibited higher toluene permeability compared to SPAES50, consistent with the ex situ permeation test.

Another challenge in DEHT, in addition to toluene crossover, is the low Faraday efficiency due to the competitive nature between the toluene hydrogenation reaction and the hydrogen evolution reaction (HER), as schematically depicted in Fig. 4a.^40,41 This issue is particularly pronounced at higher current densities, where protons preferentially combine to generate H₂ rather than facilitate toluene hydrogenation. According to the report by Nagasawa et al.,⁴⁰ the predominant factor reducing Faraday efficiency in the presence of high-concentration toluene is confirmed to be the presence of hydrogen gas. Furthermore, in a recent study by Shigemasa et al.,⁴¹ a camera was installed on the cathode side of DEHT cells to visualize hydrogen gas generated from the PtRu/C catalyst layer, allowing for quantification of H₂ production and evaluation of Faraday efficiency. In this study, a gas discharge line was installed in the cathode supply and toluene storage container to collect H₂. At various current densities (50, 100, 200, 400, 600 and 800 mA cm⁻²), the produced H₂ was quantified using gas chromatography (GC, YL6500 GC, Young In) equipped with a thermal conductivity detector. To ensure sufficient H₂ for GC analysis, the cell was operated for at least 1 to 4 hours at each current density. The corresponding voltage profiles for the applied current densities are shown in Fig. S10.† Notably, during this rate experiment, SPAES50 exhibited a lower voltage increase compared to Nafion 212, despite both having similar initial cell voltages of approximately 1.52 V. Since no other byproducts (such as methylcyclohexadiene and methylcyclohexene, formed by partial hydrogenation of toluene) were detected apart from H₂, as confirmed from the ¹H NMR spectra in Fig. 4b, similar to the results of a recent study,⁶⁵ the Faraday efficiency (FE) can be defined as:

FE (%) = [1 − (ṅ/ṅ_HER)] × 100 (%), (5)

where ṅ represents the molar hydrogen flux measured by GC, and ṅ_HER represents the molar hydrogen flux assuming that all electrons are used for the HER. Fig. 4c shows the hydrogen molar flux at the cathode as a function of current density. For SPAES50, no hydrogen was detected as the current density increased up to 200 mA cm⁻². In contrast, for Nafion 212, hydrogen production was detected starting at 200 mA cm⁻², as quantified by GC (Fig. S11†). At lower current densities, the hydrogenation reaction of toluene dominated; however, as the current density increased, the HER, as a side reaction, became more prominent, leading to an increase in hydrogen production over 200 mA cm⁻². Consequently, as shown in Fig. 4d, below 200 mA cm⁻², both Nafion 212 and SPAES50 PEMs exhibited nearly identical FEs, approaching 100%. This result with high FEs of over 95% at current densities below 200 mA cm⁻² is consistent with previous studies.^40,41,44 As expected, with increasing current density, the hydrogen production rate increased significantly, leading to a sharp decline in FE. However, at higher current densities, SPAES50 PEM exhibited approximately 3–5% higher efficiency compared to Nafion 212. When FE was evaluated using the ¹H NMR spectra (Fig. 4b), it was calculated as:

FE (%) = (N × n × F)/Q_total × 100 (%), (6)

where N is the number of moles of MCH formed during DEHT, n is the number of moles of electrons required for 1 mol of toluene conversion to MCH, F is the Faraday constant, and Q_total is the total charge supplied during the test. The FEs based on the ¹H NMR spectra are shown in Fig. S12.† Consistent with Fig. 4d, SPAES50 demonstrated higher FE than Nafion 212, particularly at higher current densities. The reduction in FE due to toluene permeation through the PEM is evident, as a lower fraction of toluene participates in DEHT when a highly permeable PEM is used. Therefore, the higher FE observed with SPAES50 compared to Nafion 212 suggests that, during the rate performance test (16.5 hours), the increased toluene crossover through Nafion 212 may have contributed to catalytic deactivation at the anode. This, in turn, hindered proton generation and reduced the hydrogenation efficiency at the cathode, thereby potentially affecting FE.

Fig. 4 (a) Schematic diagram illustrating the inhibition of toluene hydrogenation by hydrogen gas generation as a side reaction in the PtRu/C catalyst layer inside the operational DEHT cell. (b) ¹H NMR analysis of the solvent collected from cathode supply during the current density change test. (c) Hydrogen molar flux produced as a side reaction at the cathode as a function of applied current density. The molar flux of hydrogen in the electrolysis cell without the DEHT reaction was calculated as (the applied current density)/(2 × (Faraday constant)). (d) The calculated MCH Faraday efficiency as a function of applied current density.

Conclusion

In summary, for the direct hydrogenation of toluene to MCH via an electrochemical process, we tailored a hydrocarbon-based PEM with narrower hydrophilic domains compared to those in a commercial PFSA-based PEM to reduce toluene permeation while maintaining proton conductivity. From the ex situ toluene permeability tests, it was determined that hydrophobic toluene molecules predominantly permeate through the hydrophilic domains. Consequently, the narrower and more tortuous hydrophilic channels of SPAES PEMs resulted in lower toluene diffusivity and permeability. To investigate the effect of PEM toluene permeability on the DEHT process, we conducted, to the best of our knowledge, the first long-term constant current density test. The highly toluene-permeable Nafion 212 exhibited a noticeable increase in anodic charge transfer resistance, likely due to significant contamination by toluene that permeated from the cathode side. In contrast, the relatively toluene-impermeable SPAES50 demonstrated stable performance without considerable changes in cell voltage and resistance. Additionally, during the rate performance test, toluene permeation through the PEM was found to negatively impact Faraday efficiency. The strategy of tailoring hydrophilic domains of PEM represents a promising approach for the DEHT process, which is potentially attractive for LOHCs, enhancing both efficiency and durability.

Data availability

The datasets supporting this article have been uploaded as part of the ESI.†

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the KRICT research programs (BSF24-510, KS2422-20) and by the Nano Material Technology Development Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science and ICT (RS-2024-00406517).

References

1. D. Henner, REN21. 2017. Renewables 2017 Global Status Report, 2017.
2. A. Rahman, O. Farrok and M. M. Haque, Renew. Sustain. Energy Rev., 2022, 161, 112279.
3. A. Azarpour, S. Suhaimi, G. Zahedi and A. Bahadori, Arabian J. Sci. Eng., 2013, 38, 317–328.
4. T. Trainer, Energy Policy, 2017, 110, 386–393.
5. Z. Abdmouleh, A. Gastli, L. Ben-Brahim, M. Haouari and N. A. Al-Emadi, Renewable Energy, 2017, 113, 266–280.
6. I. Dincer and M. I. Aydin, Energy Convers. Manage., 2023, 283, 116950.
7. A. Rabiee, A. Keane and A. Soroudi, Int. J. Hydrogen Energy, 2021, 46, 19270–19284.
8. M. Dli, A. Baliabina and N. Drozdova, Altern. Energy Ecol., 2016, 37–41.
9. E. Rivard, M. Trudeau and K. Zaghib, Materials, 2019, 12.
10. Z. Yanxing, G. Maoqiong, Z. Yuan, D. Xueqiang and S. Jun, Int. J. Hydrogen Energy, 2019, 44, 16833–16840.
11. M. M. Rampai, C. B. Mtshali, N. S. Seroka and L. Khotseng, RSC Adv., 2024, 14, 6699–6718.
12. N. Tsogt, D. Q. Gbadago and S. Hwang, Energy Convers. Manage., 2024, 299, 117856.
13. T.-H. Le, N. Tran and H.-J. Lee, Int. J. Mol. Sci., 2024, 25, 1359.
14. A. Ali and M. N. Shaikh, Int. J. Hydrogen Energy, 2024, 78, 1–21.
15. P. M. Modisha, C. N. M. Ouma, R. Garidzirai, P. Wasserscheid and D. Bessarabov, Energy Fuels, 2019, 33, 2778–2796.
16. M. Niermann, A. Beckendorff, M. Kaltschmitt and K. Bonhoff, Int. J. Hydrogen Energy, 2019, 44, 6631–6654.
17. P. T. Aakko-Saksa, C. Cook, J. Kiviaho and T. Repo, J. Power Sources, 2018, 396, 803–823.
18. Z. Abdin, C. Tang, Y. Liu and K. Catchpole, iScience, 2021, 24, 102966.
19. P. C. Rao and M. Yoon, Energies, 2020, 13, 6040.
20. V. Tietze, S. Luhr and D. Stolten, in Hydrogen Science and Engineering : Materials, Processes, Systems and Technology, 2016, pp. 659–690.
21. E. Díaz, P. Rapado-Gallego and S. Ordóñez, Fuel, 2023, 353, 129168.
22. E. A. Martynenko, A. A. Pimerzin, A. A. Savinov, S. P. Verevkin and A. A. Pimerzin, Top. Catal., 2020, 63, 178–186.
23. E. A. Martynenko, S. V Vostrikov and A. A. Pimerzin, React. Kinet., Mech. Catal., 2021, 133, 713–728.
24. X. Li, P. Shen, X. Han, Y. Wang, Y. Zhu and Z. Wu, Appl. Surf. Sci., 2021, 543, 148769.
25. G. Zhou, W. Liu, Y. Zhao, X. Wang, S. Chen, A. Jia and H. Xie, J. Ind. Eng. Chem., 2024, 134, 561–573.
26. Y. Yue, X. Liu, M. Shakouri, Y. Hu, Y. Guo and Y. Wang, Catal. Today, 2024, 433, 114688.
27. D. Arora, M. Richards, Y. Zhu, I. Martinez-Botella, X. Wang, Z. Xie, J. Chiefari, S. Saubern and C. Hornung, Chem. Eng. Process., 2024, 201, 109822.
28. A. S. Mehr, A. D. Phillips, M. P. Brandon, M. T. Pryce and J. G. Carton, Int. J. Hydrogen Energy, 2024, 70, 786–815.
29. M. Markiewicz, Y.-Q. Zhang, M. T. Empl, M. Lykaki, J. Thöming, P. Steinberg and S. Stolte, Energy Environ. Sci., 2019, 12, 366–383.
30. Á. Bárkányi, B. L. Tarcsay, L. Lovas, T. Mérő, T. Chován and A. Egedy, Clean Technol. Environ. Policy, 2024, 26, 1521–1536.
31. J. Meng, F. Zhou, H. Ma, X. Yuan, Y. Wang and J. Zhang, Top. Catal., 2021, 64, 509–520.
32. Y. Yang, K. Seo, J. S.-I. Kwon and W. Won, ACS Sustain. Chem. Eng., 2024, 12, 11167–11180.
33. J.-S. Lee, A. Cherif, H.-J. Yoon, S.-K. Seo, J.-E. Bae, H.-J. Shin, C. Lee, H. Kwon and C.-J. Lee, Renew. Sustain. Energy Rev., 2022, 165, 112556.
34. R. Atsumi, K. Kobayashi, C. Xieli, T. Nanba, H. Matsumoto, K. Matsuda and T. Tsujimura, Appl. Catal., A, 2020, 590, 117374.
35. N. A. Dehm, X. Zhang and J. M. Buriak, Inorg. Chem., 2010, 49, 2706–2714.
36. A. Maximov, A. Zolotukhina, L. Kulikov, Y. Kardasheva and E. Karakhanov, React. Kinet., Mech. Catal., 2016, 117, 729–743.
37. M. Zahmakıran, Y. Tonbul and S. Özkar, J. Am. Chem. Soc., 2010, 132, 6541–6549.
38. D. J. M. Snelders, N. Yan, W. Gan, G. Laurenczy and P. J. Dyson, ACS Catal., 2012, 2, 201–207.
39. P. Giannoccaro, M. Gargano, A. Fanizzi, C. Ferragina and M. Aresta, Appl. Catal., A, 2005, 284, 77–83.
40. K. Nagasawa, Y. Sugita, A. Atienza-Márquez, Y. Kuroda and S. Mitsushima, J. Electroanal. Chem., 2023, 938, 117431.
41. K. Shigemasa, A. Atienza-Márquez, K. Inoue, S. Jang, F. I. R. Peña, T. Araki, T. Terao, K. Nagasawa and S. Mitsushima, J. Power Sources, 2023, 554, 232304.
42. K. Matsuoka, K. Miyoshi and Y. Sato, J. Power Sources, 2017, 343, 156–160.
43. K. Nagai, K. Nagasawa and S. Mitsushima, Electrocatalysis, 2016, 7, 441–444.
44. K. Nagasawa, K. Tanimoto, J. Koike, K. Ikegami and S. Mitsushima, J. Power Sources, 2019, 439, 227070.
45. K. B. Wiles, F. Wang and J. E. McGrath, J. Polym. Sci., Part A:Polym. Chem., 2005, 43, 2964–2976.
46. W. F. Claussen, J. Am. Chem. Soc., 1952, 74, 3937–3938.
47. C. Marche, C. Ferronato, J.-C. De Hemptinne and J. Jose, J. Chem. Eng. Data, 2006, 51, 355–359.
48. J. Chao, K. R. Hall and J.-M. Yao, Thermochim. Acta, 1984, 72, 323–334.
49. D. W. Scott, G. B. Guthrie, J. F. Messerly, S. S. Todd, W. T. Berg, I. A. Hossenlopp and J. P. McCullough, J. Phys. Chem., 1962, 66, 911–914.
50. G. S. Park, Synthetic Membranes: Science, Engineering and Applications, ed. P. M. Bungay, H. K. Lonsdale and M. N. de Pinho, Springer Netherlands, Dordrecht, 1986, pp. 57–107.
51. D. W. Urry, Int. J. Quantum Chem., 1975, 9, 221–235.
52. Q. Zhao, P. Majsztrik and J. Benziger, J. Phys. Chem. B, 2011, 115, 2717–2727.
53. Y. Inami, S. Iguchi, S. Nagamatsu, K. Asakura and I. Yamanaka, ACS Omega, 2020, 5, 1221–1228.
54. A. Karmakar and S. Kundu, Mater. Today Energy, 2023, 33, 101259.
55. S. Mitsushima, Y. Takakuwa, K. Nagasawa, Y. Sawaguchi, Y. Kohno, K. Matsuzawa, Z. Awaludin, A. Kato and Y. Nishiki, Electrocatalysis, 2015, 7, 127–131.
56. K. Nagasawa, A. Kato, Y. Nishiki, Y. Matsumura, M. Atobe and S. Mitsushima, Electrochim. Acta, 2017, 246, 459–465.
57. E. Padgett, G. Bender, A. Haug, K. Lewinski, F. Sun, H. Yu, D. A. Cullen, A. J. Steinbach and S. M. Alia, J. Electrochem. Soc., 2023, 170, 84512.
58. M. Suermann, B. Bensmann and R. Hanke-Rauschenbach, J. Electrochem. Soc., 2019, 166, F645.
59. H. Liu, H. B. Tao and B. Liu, J. Phys. Chem. Lett., 2022, 13, 6520–6531.
60. C. Liu, M. Shviro, G. Bender, A. S. Gago, T. Morawietz, M. J. Dzara, I. Biswas, P. Gazdzicki, Z. Kang, S. F. Zaccarine, S. Pylypenko, K. A. Friedrich, M. Carmo and W. Lehnert, J. Electrochem. Soc., 2023, 170, 34508.
61. M. Maier, K. Smith, J. Dodwell, G. Hinds, P. R. Shearing and D. J. L. Brett, Int. J. Hydrogen Energy, 2022, 47, 30–56.
62. G. Schmidt, M. Suermann, B. Bensmann, R. Hanke-Rauschenbach and I. Neuweiler, J. Electrochem. Soc., 2020, 167, 114511.
63. S. Siracusano, S. Trocino, N. Briguglio, V. Baglio and A. S. Aricò, Materials, 2018, 11, 1368.
64. C. Rozain and P. Millet, Electrochim. Acta, 2014, 131, 160–167.
65. K. Takano, H. Tateno, Y. Matsumura, A. Fukazawa, T. Kashiwagi, K. Nakabayashi, K. Nagasawa, S. Mitsushima and M. Atobe, Bull. Chem. Soc. Jpn., 2016, 89, 1178–1183.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ta06773h

---