Skip to main content
NCBI home page
RSC Advances logoLink to RSC Advances
. 2026 Jul 3;16(35):35418–35426. doi: 10.1039/d6ra02678h

Anhydrous proton conduction in cellulose nanofiber–heterocyclic molecule composite materials

Masanori Yamada a,✉, Hiroki Shibata a
PMCID: PMC13329784  PMID: 42405161

Abstract

TEMPO (2,2,6,6-tetramethylpiperidine-1-oxylradical)-oxidized cellulose nanofibers (CNFs) form rigid rod-like structures and possess carboxyl groups on the surface of the nanofibers. We prepared CNF–heterocyclic molecule composite materials by mixing the CNF with basic heterocyclic molecules (HMs), such as imidazole (Im), benzimidazole, and pyrazole. The resulting CNF–HM composite materials formed acid–base complexes on the surface of the nanofibers. Consequently, these CNF–HM composite materials exhibited high proton conductivity under anhydrous conditions. In particular, the CNF–Im composite achieved an anhydrous proton conductivity of 1.6 × 10−4 S cm−1 at 140 °C. The activation energy for proton conduction in the CNF–Im composite material was 0.41 eV, and this value was one order of magnitude higher than those of conductors using water as a medium. In contrast, the gluconic acid–Im composite material, which lacked a continuous proton conducting pathway, exhibited a maximum proton conductivity of 5.8 × 10−7 S cm−1, which was two orders of magnitude lower than that of the CNF–Im composite material. This is due to the anhydrous proton conducting pathway formed on the CNF surface. Finally, we investigated the relationship between pKa of the HM and anhydrous proton conduction. The anhydrous proton conductivity increased with pKa, reaching a maximum anhydrous proton conductivity at a pKa value of 6.5. However, at pKa values exceeding 6.5, the anhydrous proton conductivity decreased abruptly. These results suggest that the pKa difference (ΔpKa) between acidic and basic molecules is a critical factor governing anhydrous proton conduction.

TEMPO (2,2,6,6-tetramethylpiperidine-1-oxylradical)-oxidized cellulose nanofibers (CNFs) form rigid rod‑like structures and generate an anhydrous proton‑conducting pathway on their surfaces.graphic file with name d6ra02678h-ga.jpg

1. Introduction

Polymer electrolyte fuel cells (PEFCs) have garnered significant attention as environmentally benign energy conversion systems that emit no carbon dioxide. In particular, PEFCs operating at intermediate temperatures (100–200 °C) or under anhydrous or low-humidity conditions offer numerous advantages, including higher energy efficiency, improved carbon monoxide (CO) tolerance of the platinum electrode, simplified fuel management, and the potential for co-generation.1–3 However, conventional PEFCs rely on humidified perfluorinated sulfonic acid (PFSA) membranes as proton-conductors; consequently, their proton conductivity drops abruptly above the boiling point of water due to dehydration.1–6 Moreover, the synthesis method for PFSA membranes containing fluorine atoms is complicated and costly. Furthermore, PFSA membranes possess high chemical stability and are difficult to dispose of in the environment after use. Therefore, electrolyte membranes consisting of fluorine-free polymers, such as hydrocarbon polymers, organic–inorganic polymers, and biopolymers, have been reported to exhibit proton conduction at intermediate temperatures or under anhydrous or low-humidity conditions.7–11 These membranes typically comprise acidic or basic polymers that form acid–base composites through the addition of basic or acidic molecules. Generally, within the composite, basic groups accept protons from acidic groups, generating protonated basic species and deprotonated acidic species. Protons are then transferred from a protonated basic group to an adjacent non-protonated basic group or alternatively from a deprotonated acidic group to an adjacent non-deprotonated acidic group.

Recently, we have reported anhydrous proton conductors exhibiting defined proton conducting pathways, such as self-assembled nucleobases12 and acid–base composite materials with self-assembled structures.13 In these composite materials, the alignment of proton-conducting functional groups reduces the intermolecular distance. Consequently, anhydrous proton transfer is facilitated, occurring either from protonated to non-protonated basic molecules or from non-deprotonated to deprotonated phosphate groups. Thus, the establishment of well-defined proton-conducting pathways within composite materials is crucial for achieving efficient anhydrous proton conduction. Such pathways are typically generated through polymer phase separation into hydrophobic and hydrophilic domains,14 the formation of crystalline ion channels,15,16 or the self-assembly of surfactant molecules bearing long alkyl chains.17 In this context, nanofibers have emerged as promising proton-conductive components because their surface acidic or basic groups can serve as proton donors or acceptors, thereby facilitating anhydrous proton transport. A notable advantage of nanofibers is that their high aspect ratio and abundant surface functional groups enable the construction of continuous proton-conducting networks even at relatively low filler contents. However, their proton-conducting performance can be strongly influenced by surface chemistry, moisture sensitivity, and dispersion stability, which may limit their applicability depending on the composite environment. Considering both these advantages and limitations, we focused on cellulose nanofibers (CNFs) in this study.

Cellulose, the most abundant biomass on Earth, consists of glucose units linked by β(1 → 4) glycosidic bonds.18,19 These cellulose polymers exhibit distinct properties, including light weight, chemical stability, physical strength, enzymatic resistance, and biodegradability. Furthermore, cellulose-based materials are highly advantageous from the perspective of the Sustainable Development Goals (SDGs). Consequently, they are widely utilized in paper, textile products, food additives, etc.18,19 In particular, cellulose nanofibers (CNFs), cellulose fibrillated to the nanoscale, have garnered attention as novel cellulose-based materials for applications such as reinforcement agents, thickening, and water retention.20–22 These CNFs are synthesized using a variety of methods, and a prominent approach is the TEMPO-mediated method using 2,2,6,6-tetramethylpiperidine-1-oxylradical (TEMPO).23,24 The CNFs produced by the TEMPO method possess –COOH group on their surface.23,24 Crucially, these –COOH groups are linearly aligned along the CNF surface. Therefore, CNFs with such linearly aligned –COOH groups are promising candidates for constructing anhydrous proton-conducting pathways.

In this study, we prepared CNF–heterocyclic molecule composite materials by mixing the CNF with basic heterocyclic molecules (HMs), such as imidazole (Im), benzimidazole (BnIm), and pyrazole (Py). The CNF–Im composite exhibited an anhydrous proton conductivity of 1.6 × 10−4 S cm−1 at 140 °C. The activation energy (Ea) for proton conduction in the CNF–Im composite material was 0.41 eV; this Ea value was approximately one order of magnitude higher than that of the conductor using water as a medium. In contrast, the proton conductivity of a gluconic acid (GA)–Im composite material, which lacked a defined proton conducting pathway, was two orders of magnitude lower than that of CNF–Im composite material. This significant difference is attributed to the formation of effective anhydrous proton conducting pathways on the CNF surface. Furthermore, we investigated the relationship between the pKa value of various HMs and anhydrous proton conductivity. The anhydrous proton conductivity increased with the pKa value, reaching a maximum at pKa of 6.5, after which it decreased abruptly. These results suggest that the difference in pKa (ΔpKa) between acidic and basic molecules is a critical factor governing anhydrous proton conduction.

2. Experimental section

2.1. Material

2,2,6,6-Tetramethylpiperidine-1-oxylradical (TEMPO), sodium bromide, sodium hypochlorite, sulfamic acid, sodium hydroxide, methanol, and urea were purchased from Fujifilm Wako Pure Chemical Industries, Ltd, Osaka, Japan or Nacalai Tesque Inc., Kyoto, Japan. Cellulose powder (100–200 mesh) was obtained from Advantec Toyo Kaisha, Ltd, Tokyo, Japan. Imidazole (Im), pyrazole (Py), triazole (Tr), 2-methylimidazole (2MeIm), 4-methylimidazole (4MeIm), 1-methylimidazole (1MeIm), 2-ethylimidazole (2EtIm), 2-nonyl benzimidazole (NBnIm), and benzimidazole (BnIm) was purchased from Fujifilm Wako Pure Chemical Industries, Nacalai Tesque Inc., or Tokyo Chemical Industry Co., Ltd, Tokyo, Japan. The molecular structures of these heterocyclic molecules (HMs) are shown in Scheme 1. The ion exchange resin Amberlite IR 120(H) was obtained from Supelco, Inc., Bellefonte, PA. Solvents of analytical grade were used in all the experiments. Ultra-pure water (Merck KGaA, Darmstadt, Germany) was used in this research.

Scheme 1. Molecular structures of various HMs.

Scheme 1

Open in a new tab

2.2. Synthesis of the CNF

The CNF was synthesized as previously described in the literature.23,24 The synthetic procedure is shown in Scheme 2. The cellulose powder (2 g) was suspended in water (150 mL) containing TEMPO (0.025 g, 0.16 mmol) and sodium bromide (0.25 g, 2.4 mmol). The aqueous sodium hypochlorite solution was slowly added to the cellulose slurry and the pH of the mixture was maintained to be 10.5 at room temperature by adding a 0.5 M NaOH aqueous solution using a pH meter for 0.5–14 hours. The amount of available chlorine was 10 mmol per gram of cellulose. The reacted solution was neutralized by 1 M hydrochloric acid. The water-insoluble fractions were collected by centrifugation and washed with ultra-pure water by filtration. The obtained substrate was suspended in water (200 mL) and treated for 5 min in an ultrasonic bath of 28 kHz (W-113A tabletop ultrasonic cleaner, Honda electronics Co., Aichi, Japan). The precipitation was removed by the centrifugation and the resulting solution was changed from the Na+ form to H+ form by an ion exchange column containing an ion exchange resin. The ion exchange of CNF was confirmed by the pH measurement. The amount of carboxyl groups (the amount of H+) in the CNF was estimated by titration.

Scheme 2. Synthetic scheme of CNF.

Scheme 2

Open in a new tab

2.3. Preparation of the CNF–HM composite material

The CNF–HM composite material was prepared as follows: CNF and HM were dissolved in water or methanol/water (1 : 1, v/v) and the CNF–HM mixed solution was stirred for 24 hours at room temperature. This mixed solution was cast on a polytetrafluoroethylene (PTFE) plate and dried overnight at room temperature. The molar ratio R of the HM in the CNF–HM composite material was defined by eqn (1):

2.3. 1

where [–COOH group in CNF] and [HM] are the molar concentrations of the –COOH group in the CNF and HM, respectively.

2.4. Characterization of the CNF–HM composite material

The molecular structure of the CNF–HM composite material was determined by the KBr method using an FT/IR-4700 infrared spectrophotometer (JASCO Corporation, Tokyo, Japan). The IR spectrum was recorded at a resolution of 4 cm−1. The thermal stability of the CNF–HM composite material was analyzed using a DTG-60 thermogravimetric – differential thermal analyser (Shimadzu Corp., Kyoto, Japan). The measurement was done at a heating rate of 10 °C min−1 under flowing dry-nitrogen. The sample weight was normalized at 1 mg.

2.5. pKa measurement of HM

The pKa measurement was performed as follows:25 HM was dissolved in a 0.015 M methanol solution. This solution (30 mL) was mixed with 0.05 M hydrochloric acid (10 mL). A 0.05 M NaOH aqueous solution was added to the mixed solution while stirring and the pH was measured using a pH meter. These measurements were performed under nitrogen bubbling. The pKa value was calculated from the approximate curve.

2.6. Proton conductivity measurements of the CNF–HM composite material

The proton conduction of the CNF–HM composite material was measured as follows: the CNF–HM composite film was sandwiched between two gold electrodes (5 mm diameter).12,13 The electrode samples were placed in a stainless-steel vessel, heated at 140 °C for 2 hours under flowing dry-nitrogen to evaporate the water and volatile components, and then cooled to room temperature under flowing dry-nitrogen. The proton conductivity of the CNF–HM composite material was measured by the a.c. impedance method using a 3532-80 chemical impedance analyzer (Hioki Co., Nagano, Japan) from room temperature to 140 °C under flowing dry-nitrogen. The frequency range in the proton conductivity measurement was from 4 Hz to 1 MHz.12,13 The resistance and proton conductivity of the CNF–HM composite material were determined through the analysis of the Nyquist plots (Cole–Cole plots).26

3. Results and discussion

3.1. Synthesis of the CNF

The TEMPO-oxidized cellulose nanofibers (CNF) were prepared from the cellulose powder, TEMPO, NaBr, and NaClO.23,24 The amount of –COOH group (the amount of H+) in the CNF was determined by titration. Fig. 1 shows the amount of –COOH groups as a function of reaction time. The amount of –COOH groups in the CNF increased with the reaction time and reached a constant value at 8 hours. Consequently, the reaction time was set to 8 hours, and the CNF obtained at this duration was used for subsequent experiments. Under these conditions, the amount of –COOH group was 1.5 mmol g−1, which is consistent with the previously reported value.23,24 The molar ratio of glucose units to glucuronic acid units was calculated to be 3 : 1. Generally, on the microfibril surface, glucose units with C2-exposed secondary hydroxyl groups and those C6-exposed primary hydroxyl groups were arranged alternately, due to the trans conformation of the glycosidic linkages along the cellulose chain.17,18 Therefore, assuming that TEMPO oxidation occurred exclusively on the surface of the cellulose nanofiber, the theoretical ratio of glucose units to glucuronic acid units is expected to be 1 : 1.23,24

Fig. 1. Amount of –COOH groups in the CNF as a function of reaction time. Data are presented as the mean of three independent determinations ± standard deviation.

Fig. 1

Open in a new tab

Fig. 2 shows a photograph of the CNF membrane. The CNF membrane was prepared by casting an aqueous CNF dispersion followed by drying. The thickness of the CNF membrane is approximately 80 µm. The resulting free-standing membrane is transparent and flexible, similar to those reported in previous studies.27

Fig. 2. Photograph of the CNF membrane. Scale bar = 10 mm.

Fig. 2

Open in a new tab

3.2. pKa value of HM

The pKa values of the HMs were determined from their respective titration curves.25Fig. 3 shows the titration curve of Im. The pH value increased abruptly with the addition of NaOH solution below 1.0 mL, increased gradually in the range of 1.0–2.7 mL, and then increased abruptly again above 2.7 mL. The pKa value was calculated from the fitting curve. As a result, the pKa value of Im was determined to be 6.32. This pKa value is approximately 0.5 lower than that the reported value,28 which is attributed to the use of a 0.015 M methanol solution. Similar experiments were performed for the other HMs. Fig. S1(a)–(h) show the titration curves of 1MeIm, 2MeIm, 4MeIm, EtIm, BnIm, NBnIm, Py, and Tr, respectively (see the SI). The pKa values of these HMs were also calculated from the fitting curve. Table 1 summarizes the calculated pKa values for all HMs. In this study, we employed HMs with pKa values ranging from 2.67 to 7.42.

Fig. 3. Titration curve of Im. The pKa value of Im was determined from the corresponding titration derivative. Similar results were obtained for triplicate experiments.

Fig. 3

Open in a new tab

Table 1. pKa values, maximum proton conductivities, and activation energy (Ea) for the proton conduction of the CNF–HM composite materials.

CNF–HM pKa value of HM Maximum proton conductivity/S cm−1 Activation energy (Ea)/eV
CNF–Tr 2.67 2.34 × 10−8 0.54
CNF–Py 2.70 1.20 × 10−7 0.54
CNF–BnIm 4.81 2.00 × 10−6 1.39
CNF–NBnIm 5.58 2.40 × 10−5 0.63
CNF–1MeIm 6.30 8.94 × 10−8 0.74
CNF–Im 6.32 1.55 × 10−4 0.41
CNF–4MeIm 6.95 1.81 × 10−4 0.46
CNF–EtIm 7.23 6.62 × 10−5 0.60
CNF–2MeIm 7.42 5.33 × 10−6 0.68
Open in a new tab

3.3. Molecular structure of the CNF–HM composite material

The molecular structure of the CNF–Im composite material was evaluated using an IR spectrometer by the KBr method. Fig. 4 shows the IR spectra of (a) CNF material without the composite, (b) R = 1 composite material, (c) R = 2 composite material, (d) R = 4 composite material, (e) R = 16 composite material, and (f) Im material without the composite, respectively. The CNF material without the composite exhibited an absorption band at 1733 cm−1, assigned to the stretching vibration of C Created by potrace 1.16, written by Peter Selinger 2001-2019 O in –COOH29–32 on the CNF surface. Upon the addition of Im, this absorption band disappeared and a new absorption band emerged at 1604 cm−1, which corresponds to the asymmetric stretching vibration of –COO−,29–32 this indicates that the –COOH groups on the CNF surface were deprotonated by the composite with a basic Im molecule. In contrast, the absorption band at 1670 cm−1, attributed to the aromatic C Created by potrace 1.16, written by Peter Selinger 2001-2019 N stretching in the Im molecule,29 decreased in intensity with the addition of acidic CNF. This is due to the protonation to the –N Created by potrace 1.16, written by Peter Selinger 2001-2019 group of the Im molecule. These results suggest that the acidic CNF and basic Im molecules form an acid–base composite material through intermolecular electrostatic interactions between the –COO– groups on the CNF surface and the protonated Im molecules. Similar results have been reported for various acid–base composite materials.9,32,33

Fig. 4. IR spectra of the CNF–Im composite material with varying Im mixing ratios. (a) CNF material without the composite, (b) R = 1 composite material, (c) R = 2 composite material, (d) R = 4 composite material, (e) R = 16 composite material, and (f) Im material without the composite. Similar results were obtained for triplicate experiments.

Fig. 4

Open in a new tab

Fig. S2(A)–(H) show the IR spectra of CNF–1MeIm, CNF–2MeIm, CNF–4MeIm, CNF–EtIm, CNF–BnIm, CNF–NBnIm, CNF–Py, and CNF–Tr composite materials, respectively (see the SI). Specifically, spectra (a)–(e) in Fig. S2(A)–(H) display the results for various R values of HMs. All composite materials exhibited deprotonation of the –COOH groups on the CNF surface. Additionally, protonation of the –N Created by potrace 1.16, written by Peter Selinger 2001-2019 groups was observed upon CNF addition. These results confirm that, analogous to the CNF–Im system, the acidic CNF and basic HMs form acid–base composite materials via intermolecular electrostatic interactions.

3.4. Thermal property of the CNF–HM composite material

The thermal property of the CNF–Im composite materials was evaluated using thermogravimetric (TG) and differential thermal analysis (DTA). These measurements were performed from room temperature to 300 °C at a heating rate of 10 °C min−1 under flowing dry-nitrogen. Fig. 5(a) and (b) show the TG and DTA curves of (1) CNF material without the composite, (2) R = 1 composite material, (3) R = 4 composite material, (4) R = 8 composite material, (5) R = 16 composite material, and (6) Im material without the composite, respectively. The CNF material without the composite exhibited a small endo-thermic peak below 100 °C, which is attributed to the evaporation of adsorbed water. Furthermore, the CNF material did not show exo- and endo-thermic peaks between 100 and 300 °C. Consequently, the CNF material without the composite indicated thermal stability up to 300 °C. In contrast, the Im material without the composite showed two endothermic peaks at approximately 90 °C and 140 °C, corresponding to the melting and evaporation of Im, respectively. Similar thermal behavior has been reported for other composite materials containing the Im molecule.9 Then, we prepared the CNF–Im composite material by mixing CNF and Im and measured the thermal stability. Upon the formation of the CNF–Im composite material, the intensities of these two endo-thermic peaks decreased significantly. This phenomenon suggests thermal stabilization driven by electrostatic interactions between the acidic and basic moieties, as previously reported in similar systems.9 These results indicate that the CNF–Im composite materials are thermally stable up to 140 °C.

Fig. 5. (a) TG curve and (b) DTA curves of the CNF−Im composite material. (1) CNF material without the composite, (2) R = 1 composite material, (3) R = 4 composite material, (4) R = 8 composite material, (5) R = 16 composite material, and (6) Im material without the composite. Similar results were obtained for triplicate experiments.

Fig. 5

Open in a new tab

Similar measurements were performed on the other CNF–HM composite materials. Fig. S3 displays TG (upper) and DTA (lower) curves for (a) CNF–1MeIm, (b) CNF–2MeIm, (c) CNF–4MeIm, (d) CNF–EtIm, (e) CNF–BnIm, (f) CNF–NBnIm, (g) CNF–Py, and (h) CNF–Tr composite materials, respectively (see the SI). With the exception of the 1MeIm (as 1MeIm is a liquid at room temperature), all CNF–HM composite materials exhibited enhanced thermal stabilization due to acid–base complex formation via intermolecular electrostatic interactions. Thus, it is concluded that these CNF–HM composite materials possess thermal stability up to 140 °C.

3.5. Proton conductivity of the CNF–Im composite material

The proton conductivity measurements of CNF–HM composite materials were demonstrated under flowing dry-nitrogen at a temperature up to 140 °C. Prior to the proton conductivity measurement, the samples were preheated in a stainless-steel vessel at 140 °C for 2 hours under flowing dry-nitrogen to remove adsorbed water and other volatile components. Consequently, the samples were considered free of ionic species other than protons. The resistance and proton conductivity of the CNF–HM composite material were determined through the analysis of the Nyquist plots (Cole–Cole plots).26

The proton conductivity measurements of the CNF–Im composite materials were performed under flowing dry-nitrogen. Fig. 6 shows the temperature dependence of the proton conductivity for (□) R = 1 composite material, (△) R = 4 composite material, (■) R = 8 composite material, (●) R = 16 composite material, and (▲) R = 24 composite material, respectively. Based on the results of the TG–DTA measurements, the CNF–Im composite was found to be stable up to 140 °C; therefore, proton conductivity measurements were conducted up to this temperature. The CNF without the composite did not exhibit proton conduction, and the Im molecule without the composite similarly showed no measurable proton conductivity. Similar results have been reported for various Im based composite materials.9 In contrast, the CNF with the composite of Im exhibited proton conduction, with conductivity increasing in correlation with the R values. Furthermore, the proton conductivity increased with the temperature. The CNF–Im composite material showed the 1.6 × 10−4 S cm−1 of maximum proton conductivity at R = 16. Given the linear alignment of –COOH groups on the CNF surface, it is hypothesized that the Im molecules interacting with these –COOH groups also adopt a similar linear arrangement.

Fig. 6. Proton conductivity of the CNF–Im composite material with different Im mixing ratios. (□) R = 1 composite material, (△) R = 4 composite material, (■) R = 8 composite material, (●) R = 16 composite material, and (▲) R = 24 composite material. (◯) shows the proton conductivity of the GA–Im (R = 16) composite material. Similar results were obtained for triplicate experiments.

Fig. 6

Open in a new tab

Then, to investigate the proton conduction mechanism, we prepared a control composite material in which the Im molecules cannot be linearly aligned. In our synthesized CNF sample, the molar ratio of glucose units to glucuronic acid (GA) units – where the C6 position of the glucose molecule is substituted with a carboxyl group – on the CNF surface is 3 : 1. Accordingly, a reference GA mixture was prepared by mixing glucose and glucuronic acid molecules at a 3 : 1 molar ratio. We then prepared the GA–Im composite material by mixing this GA mixture with Im molecules and measured its proton conductivity under flowing dry-nitrogen. The (◯) in Fig. 6 represents the proton conductivity of the GA–Im composite material at R = 16. The GA–Im composite material, which lacks the linear arrangement of –COOH groups, exhibited a maximum proton conductivity of 5.8 × 10−7 S cm−1 at 140 °C. This conductive value is two orders of magnitude lower than that of the CNF–Im (R = 16) composite material. These results suggest that the one-dimensional linear structure of CNFs is necessary for the high proton conductivity observed in the CNF–Im composite material. Similar phenomena, where linear alignment facilitates high proton conductivity, have been reported in the system as self-assembled nucleobases12 and the protamine–monododecyl phosphate (MDP) composite material.13

Fig. 7 shows the Arrhenius plots of the proton conduction of (□) R = 1 composite material, (△) R = 4 composite material, (■) R = 8 composite material, (●) R = 16 composite material, and (▲) R = 24 composite material. (◯) shows the Arrhenius plots of the proton conductivity of the GA–Im composite material at R = 16. Since the CNF–Im composite materials exhibited a linear relationship within this temperature range, the proton conduction mechanism is considered to follow Arrhenius behavior, indicating that the mechanism remains consistent across the measured temperatures. The activation energy (Ea) for proton conduction was calculated from the slop of these linear fits. Table 1 summarized the maximum proton conductivity and Ea values for the proton conduction of the CNF–Im composite material. The Ea value for the R = 16 composite material was 0.41 eV, which was significantly one order of magnitude higher than that of the humidified proton conducting material, such as humidified Nafion® and sulfonated poly(ether ether ketone).34–36 Given the absence of ionic species other than protons in the CNF–Im composite material, these results suggest that proton transport in the CNF–Im composite material occurs without the mediation of mobile water molecules, such as hydronium ion. Thus, the CNF–Im composite material functions as an anhydrous proton conductor. Similarly, the Ea value of proton conduction for the GA–Im composite material was 0.50 eV, indicating that this composite material also operated as an anhydrous proton conductor.

Fig. 7. Arrhenius plots of the proton conductivity of the CNF–Im composite material with various Im mixing ratios. (□) R = 1 composite material, (△) R = 4 composite material, (■) R = 8 composite material, (●) R = 16 composite material, and (▲) R = 24 composite material. (◯) shows the proton conductivity of GA–Im (R = 16) composite material. Similar results were obtained for triplicate experiments.

Fig. 7

Open in a new tab

Fig. 8 illustrates the proposed mechanism for anhydrous proton transfer in the CNF–Im composite material. The –COOH groups are linearly aligned on the CNF surface. Due to the intermolecular acid–base interactions with these –COOH groups, the Im molecules also adopt a linear arrangement along the CNF surface. In this system, the CNF–Im material is formed through electrostatic interactions between the –COOH groups of the CNF and the –N Created by potrace 1.16, written by Peter Selinger 2001-2019 sites of the Im molecules. Specifically, protons from the –COOH groups interact strongly with the –N Created by potrace 1.16, written by Peter Selinger 2001-2019 sites of Im, leading to the formation of protonated Im molecules (indicated by the blue arrow). At high R values, the Im molecules are densely packed on the CNF surface. Consequently, anhydrous proton conduction occurs via proton hopping from the protonated Im molecules to adjacent neutral Im molecules (indicated by the red arrow). This proton transfer is facilitated by the one-dimensional pathway established by the linearly aligned Im molecules on the CNF surface. In contrast, the GA–Im composite material, prepared from a mixture of glucuronic acid (GA) and Im, fails to exhibit high proton conductivity because it cannot form such an ordered proton conduction pathway. Similar enhancement in proton conductivity via the formation of structured pathways, have been reported for self-assembled nucleobases12 and the protamine–MDP composite material,13 both of which utilize the organized two-dimensional proton conducting network in the composite material.

Fig. 8. Schematic of the proposed anhydrous proton transfer mechanism in the CNF–Im composite material. Blue and red arrows indicate the proton transfer from the –COOH group to the Im molecule and between adjacent Im molecules, respectively.

Fig. 8

Open in a new tab

3.6. Relationship between the proton conductivity and the pKa value of various HMs

Fig. S4(a)–(h) show the temperature dependence of the proton conduction for the CNF–1MeIm, CNF–2MeIm, CNF–4MeIm, CNF–EtIm, CNF–BnIm, CNF–NBnIm, CNF–Py, and CNF–Tr composite materials, respectively (see the SI). In all cases, the proton conductivity increased with the HM mixing ratio. Furthermore, the proton conductivity increased with temperature, reaching a maximum at 140 °C. The maximum proton conductivities for these CNF–HM composite materials are summarized in Table 1. The maximum proton conductivities of these composite materials ranged from 2.34 × 10−8 to 1.81 × 10−4 S cm−1 at 140 °C, with CNF–4MeIm exhibiting the highest proton conductivity among the tested composite materials. Fig. S5(a)–(h) present the Arrhenius plots of the proton conduction for the corresponding CNF–HM composite materials (see the SI). Consistent with the CNF–Im composite material, all CNF–HM composite materials displayed linear relationships, indicating that the proton conducting mechanism follows Arrhenius behavior within this temperature range. The Ea values of proton conduction, calculated from the slopes, ranged from 0.41 to 1.39 eV (Table 1). These values are more than one order of magnitude higher than those of typical humidified proton conducting materials.34–36 These results suggest that all the synthesized CNF–HM composite materials function as anhydrous proton conductors.

Fig. 9 shows the relationship between the maximum proton conductivity at 140 °C and the pKa value of various HMs. The maximum proton conductivity values were derived from Fig. 6 and S4, while the pKa values were obtained from Fig. 3 and S1. The proton conductivity of CNF–HM composite materials increased with the pKa value, reaching a peak of 1.8 × 10−4 S cm−1, and then decreased abruptly. The pKa value of the maximum proton conductivity was 6.5. These results suggested that the proton conduction mechanism is strongly dependent on the pKa values of the HMs. Generally, mixing acidic and basic molecules results in the formation of an acid–base composite material, where the intermolecular interaction strength increases with the difference in pKa (ΔpKa) value between the components. A higher ΔpKa promotes the formation of protonated HM species, which initially facilitates anhydrous proton transfer to adjacent neutral HM molecules. However, when the ΔpKa value is excessively high, the interaction between the –COOH groups and HM molecules becomes too strong, forming an overly stable complex. In such cases, the protonated HM molecules are stabilized to the extent that proton dissociation and subsequent transfer to neighboring molecules are hindered. Consequently, the anhydrous proton conductivity decreased abruptly for HM molecules with a pKa values exceeding 6.5.

Fig. 9. Relationship between the maximum proton conductivity at 140 °C and the pKa values of various HMs. The maximum proton conductivity values were obtained from Fig. 6 and S4, while the pKa values were obtained from Fig. 3 and S1.

Fig. 9

Open in a new tab

4. Conclusions

We developed CNF–heterocyclic molecule composite materials by incorporating various basic heterocyclic molecules (HMs) into a cellulose nanofiber (CNF) matrix. The CNF–Im composite exhibited an anhydrous proton conductivity of 1.6 × 10−4 S cm−1 at 140 °C. In contrast, the anhydrous proton conductivity of a gluconic acid (GA)–Im composite material – which lacked a defined structural pathway – was two orders of magnitude lower. This significant disparity was attributed to the formation of highly effective, one-dimensional anhydrous proton conducting pathways along the CNF surface. Furthermore, our investigation into the relationship between the pKa value of various HMs and their conductivity revealed a peak at a pKa value of 6.5, suggesting that the pKa difference (ΔpKa) between the acidic and basic moieties is a critical factor governing anhydrous proton transport. As cellulose is an environmentally benign and sustainable resource, these anhydrous proton conducting CNF–HM composite materials hold great promise for application in next-generation electronic devices within the biomedical, bioengineering, and environmental sectors.

Author contributions

Masanori Yamada: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, supervision, validation, visualization, writing – original draft, writing – review and editing. Hiroki Shibata: formal analysis, software.

Conflicts of interest

There are no conflicts to declare.

Supplementary Material

RA-016-D6RA02678H-s001

Acknowledgments

This work was supported by JSPS KAKENHI Grant Numbers JP22K12451 and JP25K15509.

Data availability

The data that support the findings of this study are available in the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d6ra02678h.

Notes and references

  1. Segale M. Seadira T. Sigwadi R. Mokrani T. Summers G. A new frontier towards the development of efficient SPEEK polymer membranes for PEM fuel cell applications: a review. Mater. Adv. 2024;5:7979–8006. doi: 10.1039/D4MA00628C. [DOI] [Google Scholar]
  2. Shah M. Hakim N. U. D. Advances in nanocellulose proton conductivity and applications in polymer electrolyte membrane fuel cells. Next Mater. 2025;6:100484. doi: 10.1016/j.nxmate.2025.100484. [DOI] [Google Scholar]
  3. Kahraman H. Akin Y. Recent studies on proton exchange membrane fuel cell components, review of the literature. Energy Convers. Manage. 2024;304:118244. doi: 10.1016/j.enconman.2024.118244. [DOI] [Google Scholar]
  4. Xiao T. Wang R. Chang Z. Fang Z. Zhu Z. Xu C. Electrolyte membranes for intermediate temperature proton exchange membrane fuel cell. Prog. Nat. Sci. 2020;30:743–750. doi: 10.1016/j.pnsc.2020.08.014. [DOI] [Google Scholar]
  5. Escorihuela J. Olvera-Mancilla J. Alexandrova L. del Castillo L. F. Compañ V. Recent progress in the development of composite membranes based on polybenzimidazole for high temperature proton exchange membrane (PEM) fuel cell applications. Polymers. 2020;12(41):1861. doi: 10.3390/polym12091861. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Zhou Z. Zholobko O. Wu X. F. Aulich T. Thakare J. Hurley J. Polybenzimidazole-based polymer electrolyte membranes for high-temperature fuel cells: Current status and prospects. Energies. 2021;14:135. doi: 10.3390/en14010135. [DOI] [Google Scholar]
  7. Esmaeili N. Gray E. M. Webb C. J. Non-fluorinated polymer composite proton exchange membranes for fuel cell applications – A review. ChemPhysChem. 2019;20:2016–2053. doi: 10.1002/cphc.201900191. [DOI] [PubMed] [Google Scholar]
  8. Quartarone E. Angioni S. Mustarelli P. Polymer and composite membranes for proton-conducting, high-temperature fuel cells: A critical review. Materials. 2017;10:687. doi: 10.3390/ma10070687. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Honma I. Yamada M. Bio-inspired membranes for advanced polymer electrolyte fuel cells. Anhydrous proton-conducting membrane via molecular self-assembly. Bull. Chem. Soc. Jpn. 2007;80:2110–2123. doi: 10.1246/bcsj.80.2110. [DOI] [Google Scholar]
  10. Yamada M. Honma I. Anhydrous proton conducting polymer electrolytes based on poly(vinylphosphonic acid)-heterocycle composite material. Polymer. 2005;46:2986–2992. doi: 10.1016/j.polymer.2005.02.056. [DOI] [Google Scholar]
  11. Yamada M. Goto A. Proton conduction of DNA-imidazole composite material under anhydrous condition. Polym. J. 2012;44:415–420. doi: 10.1038/pj.2012.5. [DOI] [Google Scholar]
  12. Yamada M. Tanoue K. Synthesis of self-assembled nucleobase and its anhydrous proton conductivity. RSC Adv. 2019;9:36416–36423. doi: 10.1039/C9RA06841D. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Yamada M. Yoshihara N. Anhydrous proton conductor consisting of protamine–monododecyl phosphate composite with self-assembled structure. RSC Adv. 2023;13:34877–34883. doi: 10.1039/D3RA07191J. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Chopade S. A. So S. Hillmyer M. A. Lodge T. P. Self-supporting, hydrophobic, ionic liquid-based reference electrodes prepared by polymerization-induced microphase separation. ACS Appl. Mater. Interfaces. 2016;8:6200–6210. doi: 10.1021/acsami.5b12366. [DOI] [PubMed] [Google Scholar]
  15. Cao L. He X. Jiang Z. Li X. Li Y. Ren Y. Yang L. Wu H. Channel-facilitated molecule and ion transport across polymer composite membranes. Chem. Soc. Rev. 2017;46:6725–6745. doi: 10.1039/C5CS00906E. [DOI] [PubMed] [Google Scholar]
  16. Kim O. Kim K. Choi U. H. Park M. J. Tuning anhydrous proton conduction in single-ion polymers by crystalline ion channels. Nat. Commun. 2018;9:5029. doi: 10.1038/s41467-018-07503-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Yamada M. Honma I. Anhydrous protonic conductivity of a self-assembled acid-base composite material. J. Phys. Chem. B. 2004;108:5522–5526. doi: 10.1021/jp030767o. [DOI] [Google Scholar]
  18. Padinjakkara A., Thankappan A., Souza F. G. and Thomas S., Biopolymers and Biomaterials, Apple Academic Press, New York, 2019 [Google Scholar]
  19. Young R. A. and Rowell R. M., Cellulose: Structure, Modification and Hydrolysis, Wiley, New York, 1986 [Google Scholar]
  20. Isogai A. Wood nanocelluloses: fundamentals and applications as new bio-based nanomaterials. J. Wood Sci. 2013;59:449–459. doi: 10.1007/s10086-013-1365-z. [DOI] [Google Scholar]
  21. Menon M. P. Selvakumar R. Kumar P. S. Ramakrishna S. Extraction and modification of cellulose nanofibers derived from biomass for environmental application. RSC Adv. 2017;7:42750–42773. doi: 10.1039/C7RA06713E. [DOI] [Google Scholar]
  22. Dufresne A., Nanocellulose: Potential Reinforcement in Composites, The Royal Society of Chemistry, London, 2012. pp. 978–1-84973-531-5 [Google Scholar]
  23. Isogai A. Saito T. Fukuzumi H. TEMPO-oxidized cellulose nanofibers. Nanoscale. 2011;3:71–85. doi: 10.1039/C0NR00583E. [DOI] [PubMed] [Google Scholar]
  24. Saito T. Nishiyama Y. Putaux J. L. Vignon M. Isogai A. Homogeneous suspensions of individualized microfibrils from TEMPO-catalyzed oxidation of native cellulose. Biomacromolecules. 2006;7:1687–1691. doi: 10.1021/bm060154s. [DOI] [PubMed] [Google Scholar]
  25. Nakamura S. Application of pKa values for structural studies. J. Synth. Org. Chem., Jpn. 1966;24:923–933. doi: 10.5059/yukigoseikyokaishi.24.923. [DOI] [Google Scholar]
  26. Bard A. J. and Faulkner L. R., Electrochemical Methods: Fundamentals and Applications, 2nd edn, Wiley, New York, 2001 [Google Scholar]
  27. Zhu H. Fang Z. Preston C. Li Y. Hu L. Transparent paper: fabrications, properties, and device applications. Energy Environ. Sci. 2014;7:269–287. doi: 10.1039/C3EE43024C. [DOI] [Google Scholar]
  28. Acheson R. M., An Introduction to the Chemistry of Heterocyclic Compounds, 3rd edn, Wiley, Canada, 1976 [Google Scholar]
  29. Silverstein R. M. and Webster F. X., Spectrometric Identification of Organic Compounds, John Wiley & Sons, New York, 1998 [Google Scholar]
  30. Ptak S. H. Sanchez L. Fretté X. Kurouski D. Complementarity of Raman and infrared spectroscopy for rapid characterization of fucoidan extracts. Plant Methods. 2021;17:130. doi: 10.1186/s13007-021-00830-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Shi L. Gunasekaran S. Preparation of pectin–ZnO nanocomposite. Nanoscale Res. Lett. 2008;3:491–495. doi: 10.1007/s11671-008-9185-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Günday Ş. T. Bozkurt A. Preparation and proton conductivity of polymer electrolytes based on alginic acid and 1,2,4-triazole. Polym. J. 2008;40:104–108. doi: 10.1295/polymj.PJ2007152. [DOI] [Google Scholar]
  33. Nakayama M. Sugiura Y. Hayakawa T. Nogami M. A novel proton conductor of imidazole–aluminium phosphate hybrids in the solid state. Phys. Chem. Chem. Phys. 2011;13:9439–9444. doi: 10.1039/C0CP02789H. [DOI] [PubMed] [Google Scholar]
  34. Kreuer K. D. Proton conductivity: Materials and applications. Chem. Mater. 1996;8:610–641. doi: 10.1021/cm950192a. [DOI] [Google Scholar]
  35. Shao Z. G. Joghee P. Hsing I. M. Preparation and characterization of hybrid Nafion–silica membrane doped with phosphotungstic acid for high temperature operation of proton exchange membrane fuel cells. J. Membr. Sci. 2004;229:43–51. doi: 10.1016/j.memsci.2003.09.014. [DOI] [Google Scholar]
  36. Kim J. D. Ohira A. Crosslinked sulfonated polyphenylsulfone (CSPPSU) membranes for elevated-temperature PEM water electrolysis. Membr. 2021;11:861. doi: 10.3390/membranes11110861. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

RA-016-D6RA02678H-s001
RA-016-D6RA02678H-s001.pdf (612.5KB, pdf)

Data Availability Statement

The data that support the findings of this study are available in the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d6ra02678h.

Articles from RSC Advances are provided here courtesy of Royal Society of Chemistry

RESOURCES