Introduction

Structural control of highly ion-conductive channels is a notable strategy for energy conversion system, water treatment, and biotechnology [1,2,3,4,5]. Nanostructured liquid crystals that self-organize into dimensionally ordered states represent a promising approach to the development of structural control of ion-conductive materials [6, 7]. Thermotropic and lyotropic liquid crystal properties can drive organized structure of various kinds. Kato et al. demonstrated switchable ionic conductivities induced by a rectangular–hexagonal phase transition in wedge-shaped liquid-crystalline (LC) ammonium salts [8]. Several groups have also reported relationships between ionic conductivity and LC structure with thermotropic and lyotropic liquid crystal properties [9,10,11,12].

Designing phase-separated conductive channels composed of hydrophobic backbones and hydrophilic parts is a fundamentally important strategy to create highly proton-conductive materials [13,14,15,16]. Typical proton exchange membranes (PEMs) have sulfonic acid groups attached to the hydrophobic polymer as a side chain [17]. State-of-the-art membranes are based on perfluorosulfonic acid (PFSA) ionomers such as Nafion [13, 18, 19]. They exhibit high levels of proton conductivity, water transport, and durability. Protons are transported through water-swollen conductive channels. Understanding the relationship between the structure and proton transport property is fundamentally important, but such attempts have often been hampered in many highly proton-conductive polymers because less structural information can be derived from an amorphous or amorphous-like nature.

Our recent studies have elucidated that proton conductivity enhancement originates from the improvement of molecular ordering and the main chain orientation of LC domains in alkyl-sulfonated polyimide (ASPI) thin films [20,21,22,23]. In fact, ASPI thin films with rigid backbones exhibit lamellar-organized structure by water uptake using a lyotropic LC property. However, no reports have described the influence of the semialiphatic backbones on the organized structure and proton conductivity of alkyl-sulfonated polyimide thin films.

This study elucidated the relationship between the organized structure and proton conductivity by introduction of a semialiphatic structure in the main chain. Lower planarity of backbones can change the organized molecular structure and ordering in the lyotropic LC structure. For polyimides without sulfonic acid groups, Ando and coworkers [24] reported the intermolecular aggregation structures in thin films of fully aromatic polyimides and semialiphatic polyimides by grazing incidence X-ray scattering measurements. In our work, the lyotropic LC structure and proton conductivity were assessed in two alkyl-sulfonated semialiphatic polyimide (ASSPI) films with different molecular weights by in situ studies of grazing incidence small-angle X-ray scattering (GISAXS), quartz crystal microbalance (QCM), Fourier transform infrared spectra (FT-IR) and impedance analyses. Based on those results, we conducted systematic evaluations of the molecular structure, proton conductivity, water uptake, and dissociation state of protons from sulfonic acid groups in thin films. Two ASSPI thin films exhibited different orientations of the lamellar structure and proton conductivity. In contrast, water uptake isotherms and the proton dissociation properties from sulfonic acid groups were almost identical in the two thin films. Therefore, the orientation of the lyotropic lamellar structure largely contributed to the proton conductivity enhancement. Our results indicate that not only main chain planarity but also the increment of molecular weight affects the lyotropic LC structure and, moreover, the oriented lamellar structure enhanced proton conductivity.

Materials and methods

Materials

The 3,3′-dihydroxybenzidine, 1,3−propanesultone, 9,10-dihydro-9-oxa-10-phosphaphenanthrene 10-oxide (DOPO) and 1,2,4,5-cyclohexanetetracarboxylic dianhydride (HPMDA) were purchased from Tokyo Chemical Industry Co. Ltd., Japan and were used without further purification. The 3,3′-bis(sulfopropoxy)-4,4′-diaminobiphenyl (3,3′-BSPA) was synthesized from 3,3’-dihydroxybendizine and 1,3−propanesultone as described in the literature [22]. Acetone and m-cresol were purchased from Wako Pure Chemical Ind. Ltd., Japan. Triethylamine (TEA) was used as received from Kanto Chemical Co. Inc., Japan.

Synthesis of ASSPI

The ASSPIs of different molecular weights were synthesized using a similar polymerization scheme. Scheme 1 represents the synthesis of the higher-molecular-weight (HMw) ASSPIs. For HMw ASSPIs, DOPO was used as the catalyst, whereas it was not used in the case of lower-molecular-weight (LMw) ASSPIs. Polymerization was conducted using 3,3′-BSPA (0.46 g), HPMDA (0.22 g), DOPO (0.86 g), m-cresol (5 ml), and TEA (0.3 ml) in a 50-mL three-necked round-bottomed flask with a mechanical stirrer under an argon atmosphere. m-Cresol was used as the solvent. After reaction for 25 h at 180 °C, the polymerized mixture was poured dropwise into cooled acetone and washed by centrifugation. The final product was dried under a vacuum and again subjected to an ion-exchange process using Amberlyst. According to the previous literature [25], molecular weight and polydispersity were examined by gel permeation chromatography (GPC), and the results are shown in Supplementary Table S1. In this scheme, we obtained the ASSPI polymer with a high-molecular-weight (HMw, 4.0 × 104). We also obtained the ASSPI polymer with a lower-molecular weight (LMw, 2.5 × 104), which was polymerized according to the same scheme without DOPO (Scheme S1). The obtained product was characterized based on 1H NMR and FT-IR spectra (Supplementary Figs. S1 and S2). We selected it for comparison. (Supplementary Fig. S3 shows the GPC results). The degree of sulfonation for the final product was estimated from the 1H NMR data as more than 98%. The calculated ion-exchange capacity (IEC) was 2.9 meq g-1.

Scheme 1
scheme 1

Synthesis of higher-molecular-weight alkyl-sulfonated semialiphatic polyimides

Full size image

Preparation of thin films

ASSPI thin films onto Si, SiO2 substrates and SiO2-coated 9 MHz quartz crystals (Seiko EG&G Co. Ltd.) were prepared from 5 wt% of ASSPI solution with a mixed solvent of Milli-Q water and tetrahydrofuran using a spincast method with a spin-coater (ACT−200D; Active Co. Ltd.). Before film deposition, the substrates were washed by soaking in 2-propanol and plasma treatment with a vacuum plasma system (Cute-MP; Femto Science, Korea) to improve the surface hydrophilicity. The thicknesses of the thin films were measured using a white interference microscope (BW-S506; Nikon Corp.) and an atomic force microscope (AFM, VN-8000; Keyence Co.).

Grazing incidence small-angle X-ray scattering (GISAXS)

GISAXS measurements were taken with a FR-E X-ray diffractometer equipped with R-AXIS IV two-dimensional (2D) detector (FR-E; Rigaku Corp.). Thin-film samples were placed into a humidity-controlled cell with X-ray transparent polyester film (Lumirror) windows. To control the humidity, nitrogen carrier gas was used as received from the gas cylinder without further dehumidification. A voltage of 45 kV, current of 45 mA, and irradiation time of 1 h were applied to create copper Cu Kα radiation (λ = 0.1542 nm) with a beam size of approximately 300 × 300 μm. The camera length was 300 mm. X-ray scattering patterns were recorded on an imaging plate (Fujifilm Corp.). The incident angle was chosen as 0.20°–0.22°. For 1D out-of-plane and in-plane patterns, the integrated regions were taken, respectively, between –0.5° to + 0.5° as 2θ from the center (0°) and a width of 1° as 2θ, respectively.

Thin-film conductivity measurements

For conductivity measurements of ASSPI thin films, impedance spectroscopy measurements were performed using a two-probe method to determine the proton conductivity parallel to the film surface with a frequency response analyzer and high-frequency dielectric interface (SI1260 and SI1296; Solartron Analytical). Gold paste was used as thin-film electrodes for the conductivity measurements. The relative humidity (RH) and temperature were monitored by a computer-controlled environmental test chamber (SH-221; Espec Corp.). Impedance data were collected by application of an alternating potential of 50 mV over frequencies ranging from 10 MHz to 1 Hz. The thin-film conductivity (σ) was calculated from the resistance value (R) obtained directly from the impedance measurements using

$$\sigma = \frac{d}{{Rlt}}$$

where d represents the space between the gold electrodes, t stands for the film thickness, and l expresses the contact electrode length.

Water uptake measurements of thin films

Water uptake of ASSPI thin films was measured by an in situ QCM system. The relative humidity (RH) was controlled by a dry N2 gas and humidified streams using a humidity controller (BEL Flow; BEL Japan Inc.). QCM substrates were connected to an oscillation circuit with a DC power supply and frequency counter (53131 A; Agilent Technologies Japan Ltd.). The QCM substrate was placed in an in-house constructed humidity chamber with a high-resolution RH sensor. The frequencies found before and after spin-coating of the QCM substrate were confirmed at the dry N2 stream to determine the mass of dry film by the Sauerbrey equation as

$$\Delta m = \frac{{S \times \sqrt {\rho \mu } }}{{2 \times F^2}} \times ( - \Delta F)$$

where S denotes the electrode surface area, ρ stands for the quartz density, µ expresses the shear modulus of quartz, and F signifies the fundamental frequency of the QCM substrate. The water content λ, the number of water molecules per sulfonic acid, was calculated as shown below:

$$\lambda = \left(\frac{m}{{m_0}} - 1\right) \times \frac{{{\mathrm{E }}{\mathrm{W }}}}{{M_{{\mathrm{H }}_2{\mathrm{O }}}}}$$

where m represents the film mass at each RH, m0 stands for the film mass at 0% RH,\(M_{H_2O}\) is the molecular mass of the water molecular, and EW expresses the equivalent of each ASSPI.

In situ FT-IR

The dissociation state of protons from sulfonic acid groups was examined by in situ FT-IR measurements. ASSPI thin films on Si wafers were placed in a homemade cell. CaF2 windows were used for the humidity-controlled cell. Transmission in situ FT-IR measurements were then collected using an FT-IR spectrometer (Nicolet 6700; Thermo Fisher Scientific Inc.) equipped with a deuterium triglycine sulfate (DTGS) detector. The relative humidity (RH) was controlled by a humidity generator (me-40DP-2PW; Micro equipment). The thicknesses of the films prepared on oxidized Si substrates for this measurement were approximately 500 nm.

Results and discussion

In situ GISAXS

GISAXS is a powerful tool to reveal molecular packings and molecular orderings in molecular organized thin films [26,27,28]. Matsui and coworkers [29] showed anisotropic proton conductivity in a polymer multilayer thin film with a well-defined lamellar structure that was confirmed based on GISAXS measurements. Proton conductivity was enhanced by the formation of 2D hydrogen-bonding networks in multilayer nanosheets [30, 31]. To investigate the influence of the semialiphatic main chain on the lyotropic organized structure, RH-dependent in situ GISAXS measurements were conducted in ASSPI thin films. The 2D scattering images are shown in Fig. 1a–d and Fig. 2a–d for HMw and LMw ASSPI thin films, respectively. A series of in-plane and out-of-plane 1D profiles for HMw and LMw thin films with various RH values are depicted in Fig. 1e, f and Fig. 2e, f, respectively. In the in situ GISAXS measurements of the HMw thin film, scattering peaks in the out-of-plane position were observed around the small-angle region of 2θ = 4–5° (d = 2.1–2.7 nm) at more than 70% RH (Fig. 1c, d). The out-of-plane peaks in the HMw thin film implied that the repeating ordered structure was formed perpendicularly to the substrate plane in the thin film. Our earlier work demonstrated that the out-of-plane scatterings were attributed to the lamellar structure in the in-plane stacked hydrophobic polyimide backbone with hydrophilic sulfonated side chains [21]. Under high RH conditions, this out-of-plane scattering peak was enhanced and shifted to the smaller angle region (Fig. 1f). At 95% RH, the out-of-plane peak reached 3° (d = 2.7 nm); it was markedly enhanced relative to the X-ray specular peak. These results reflect that the lamellar structure is constructed by containing water through the thin film, which is incorporated selectively into the interlamellar space. The scattering peak shift and intensity enhancement with humidity represented the lamellar expansion and ordering of the amphiphilic lamellar structure to the out-of-plane direction, respectively. Molecular ordering is enhanced as the lamellar spacing expands to the out-of-plane direction as humidity further increases. In contrast, the LMw thin film exhibited semicircle scattering under conditions of high humidity of more than 70% RH (Fig. 2d–f), unlike the anisotropic scattering in the HMw thin film. In the in-plane profile, the scattering shifted to the small-angle region as humidity increased (Fig. 2e), indicating the formation of the same lamellar structure as the HMw thin film. In the out-of-plane direction, however, peak shifts were not clearly observed due to the weak scattering intensity. The semicircle scattering suggested an isotropic structure, and, therefore, the LMw thin film formed a nonoriented lamellar structure.

Fig. 1
figure 1

GISAXS results for the HMw film. Panels a, b, c, and d, respectively, show the 2D GISAXS patterns at 5% RH, 40% RH, 70% RH, and 95% RH. The humidity-dependent 1D GISAXS profiles in the e in-plane and f out-of-plane directions of the oriented lamellar thin film are also shown. The scattering arcs at the positions of 16 and 17 deg resulted from diffraction of Lumirror films used as the windows for humidity-controlled GISAXS cells

Full size image
Fig. 2
figure 2

GISAXS results for the LMw film. Panels a, b, c, and d, respectively, depict 2D GISAXS patterns at 10% RH, 40% RH, 70% RH, and 95% RH. The humidity-dependent 1D GISAXS profiles in the e in-plane and f out-of-plane directions of the Nonoriented lamellar thin film are also shown. The scattering arcs at positions of 16 and 17 deg resulted from diffraction of the Lumirror film used as windows for the humidity-controlled GISAXS cell

Full size image

Ando and coworkers [24] have discussed the details of main chain aggregates by ch-pack as intermolecular main chain packing and π-stacks as aromatic ring stacking in both aromatic and semialiphatic polyimides with no sulfonated alkyl side chains. These polyimides consisting of diaminocyclohexylmethane and pyromellitic dianhydride form a smectic LC-like ordered structure based on the main chain aggregations. Alkyl-sulfonated aromatic polyimides exhibit the scattering attributed to the periodic monomer unit length and ch-pack for the polyimide main chain in the in-plane and out-of-plane positions, respectively [20,21,22,23, 25]. The scatterings imply smectic LC ordering of the intermolecular aggregation in the lyotropic lamellar structure. However, in the present both cases of ASSPI thin films, scattering due to the main chain aggregations was not observed. The nonplanar aliphatic rings in the ASSPI thin films probably inhibited the intermolecular aggregation of the main chain in the lyotropic lamellar structure so that the main chain smectic ordering was not formed. Thereby, ASSPI thin films only exhibited scattering due to the lyotropic lamellar structure under high humidity conditions in comparison to aromatic polyimide thin films. In contrast, the HMw thin film formed an oriented lamellar structure, although the LMw thin film exhibited a nonoriented one. The reason for this structural difference might derive from the number of shorter main chains in ASSPI (Supplementary Fig. S3). Polarized optical microscopic observation clearly revealed the differences in the long range ordering and LC domain size between HMw and LMw thick films (Supplementary Fig. S4). Birefringence textures were clearly observed in the HMw film (Supplementary Fig. S4a, b), whereas the LMw film only exhibited an image close to the dark field (Supplementary Fig. S4c, d). These findings clearly indicated that the LMw film had a reduced long range order and small domain size. Hereinafter, we respectively designate HMw thin film as oriented lamellar thin films and LMw thin films as nonoriented lamellar thin films.

Proton conductivity

Figure 3 depicts RH-dependent proton conductivity plots for both oriented lamellar and nonoriented lamellar thin films. A linear increase in conductivity was observed by the increase in RH, which was similar for both the oriented lamellar and nonoriented lamellar thin films. However, the proton conductivity of the oriented lamellar film displayed more than half an order of magnitude higher value of 1.5 × 10-1 S cm-1 (at 25 °C and 95% RH) than the nonoriented lamellar thin film (3.0 × 10-2 S cm-1 at 25 °C and 95% RH). In the high RH region, this increased proton conductivity was probably due to improved molecular ordering and an oriented lamellar structure arising from an increased molecular weight. Generally, the IEC value is responsible for proton conductivity under high RH conditions [32, 33]. In the present work, however, the IEC value was fixed because of the same chemical structure of ASSPI. From the lamellar structure observed in the GISAXS measurement, the hydrophilic sulfonic acid side chains and hydrophobic polyimide main chains were segregated to form the lamellar structure parallel to the substrate. Under high humidity conditions, the hydrophilic sulfonic layers were selectively hydrated and expanded upon the adsorption of water molecules. Such continuously formed hydrophilic layers in the highly ordered structure promoted the marked enhancement of conductivity. To elucidate whether this increment of proton conductivity was a result of structural effects or water uptake, in situ QCM measurements were conducted.

Fig. 3
figure 3

RH-dependent proton conductivity plots for both oriented lamellar and nonoriented lamellar thin films at 298 K

Full size image

Water uptake

Water uptake for the PEM affects proton conductivity because water facilitates the transport of protons through the membrane [34]. Figure 4 shows RH-dependent water uptake plots for both the oriented lamellar and nonoriented lamellar thin films. The water uptake for both films increased concomitantly with increasing RH, reaching a value of λ = 16 in the oriented lamellar thin films and λ = 14 in the nonoriented films. Although no difference was detected in water uptake at low humidity (30–80% RH), the oriented lamellar thin film absorbed somewhat more water molecules than the nonoriented film. The slightly higher water uptake in the oriented film resulted from the anisotropic expansion  normal to the substrate plane under high RH conditions, as observed in the GISAXS results.

Fig. 4
figure 4

RH-dependent water uptake plots for both oriented lamellar and nonoriented lamellar thin films

Full size image

To elucidate the relationship between conductivity and water uptake, plots of conductivity vs. λ are shown in Fig. 5. When the proton conductivity was compared at the same λ value, the oriented lamellar thin-film showed more than half an order of magnitude higher proton conductivity than the nonoriented films. This result clearly indicated that the water adsorption was not responsibility for the enhancement of proton conductivity. Hence, we inferred that the oriented lamellar structure facilitated proton conductivity and engendered higher proton conductivity in the high RH region.

Fig. 5
figure 5

Proton conductivity plots for both oriented lamellar and nonoriented lamellar thin films as a function of the λ value

Full size image

Dissociation state of protons from sulfonic acid groups using in situ FT-IR

To evaluate the dissociation of protons at sulfonic acid groups during the conductivity change, in situ FT-IR measurements under controlled humidity were conducted. Figure 6 shows the humidity-dependent FT-IR spectra for the oriented lamellar and nonoriented lamellar thin films. The peaks were clearly observed at bands of 1715, 1640, 1200, and 1030 cm-1 corresponding to the νas(C = O) symmetric stretching vibrations of the imide groups, δ(H–O–H), νas(SO3−), and νs(SO3−), respectively. With increases in humidity, the absorbance of the peaks was enhanced. The number of dissociated protons could be estimated by the absorbance due to the νs(SO3−) band. Figure 7 depicts the absorbance of the νs(SO3−) band at 1030 cm-1 as a function of the λ value. The absorbance of the νs(SO3-) bands increased with the λ value and was saturated for λ > 6 in both the oriented lamellar and nonoriented lamellar thin films. Saturation of the νs(SO3-) band absorbance suggested an almost complete dissociation of the protons at sulfonic acid groups with relatively low water uptake at approximately λ = 6. However, the proton conductivity increased considerably to more than λ = 6, as shown in Fig. 5. Miyatake and coworkers [33] reported a similar tendency to correlate the proton dissociation of sulfonic acid groups and proton conductivity during the hydration process in a sulfonated block poly(arylene ether sulfone ketone) membrane [35]. Additionally, both oriented and nonoriented films exhibited essentially the same behaviors. Therefore, the dissociation state of protons at sulfonic acid groups was also not responsible for the enhancement of proton conductivity in oriented lamellar thin film.

Fig. 6
figure 6

Humidity-dependent FT-IR spectra of a oriented lamellar thin film and b nonoriented lamellar thin film

Full size image
Fig. 7
figure 7

Change in the absorbance of νs(SO3-) of ASSPI thin films observed during hydration

Full size image

In our earlier work, the highly proton-conductive organized structure was obtained using the lyotropic LC property derived from rigid aromatic main chains. In the present work, the influence of the semialiphatic structure is discussed. The molecular ordering weakened as a result of the reduced planarity of the main chain due to suppressed (π-stack) aggregation of the main chain in the lyotropic lamellar structure. The degree of molecular ordering improved with increasing molecular weight, and an oriented lamellar-organized structure was obtained. Although the IEC values, water uptake and dissociation behavior of proton were identical in both films, the oriented lamellar film exhibited half an order of magnitude higher proton conductivity compared with the nonoriented film. Therefore, we inferred that the oriented lamellar structure enhanced proton conductivity more than the nonoriented lamellar structure.

Conclusion

As described herein, the influence of the semialiphatic backbone on molecular ordering and proton conductivity was discussed for sulfonated polyimide thin films. A new alkyl-sulfonated polyimide with less planarity for the main chain was synthesized. The lower-molecular weight of the sulfonated polyimide had a nonoriented lamellar structure under conditions of 70–95% RH. However, the higher-molecular weight enhanced the degree of molecular ordering and resulted in the oriented lamellar-organized structure based on the lyotropic LC property. The reason for the structural difference might derive from the number of shorter main chains in ASSPIs. Moreover, proton conductivity was enhanced in the case of the lamellar-organized structure with the same water amounts and dissociation states of protons. We concluded that higher proton conductivity was achieved with the oriented lamellar-organized structure.