As the renewable energy sources, fuel cells have been received much attention in transportation and stationary power generation [1–3]. Fuel cell is an electrochemical powering device that converts the chemical energy directly into the electrical energy by using hydrogen gas as a fuel. Hydrogen gas is being also widely used in many chemical processes and other industrial applications including aerospace, energy, medical and biomass processing [4–6]. However, the use of hydrogen may raise many safety concerns, and thus the detection of potentially hazardous hydrogen is receiving increased attention. In this respect, the development of the hydrogen sensors has been an important research topic in the field of process control and hydrogen detection for the sake of safety. In view of their application, the sensors should fulfill the requirements such as high sensitivity, low cost, long-term stability, high reliability, simple construction and easy operation. Further, the ambient temperature operation is an important criterion to achieve safe and reliable performance [7,8]. For this purpose, the electrochemical gas sensors have been widely used because of their high sensitivity, simplicity, low power consumption, easy operation and low cost compared to other types of sensors [9–13]. The electrochemical gas sensors are gas detectors that measure the concentration of a target gas by oxidizing or reducing the target gas at a sensing (working) electrode and measuring the resulting current. The electrochemical hydrogen sensor typically consists of proton-conducting liquid electrolyte and three electrodes (working, reference and counter electrodes) [14–18]. Unfortunately, the electrochemical sensors employing liquid electrolyte suffer from some drawbacks such as limited stability and durability, which are mainly attributed to the leakage and evaporation of the electrolyte. These problems could be overcome by replacing the liquid electrolyte by solid electrolyte. Compared to liquid electrolyte, solid electrolyte can be used in the form of thin film or pellet without any separator, which permits the fabrication of compact device without electrolyte leakage and avoids deterioration of the sensing property. However, the main disadvantage of solid electrolytes is that they exhibit low ionic conductivities at room temperature, which precludes their practical applications in the electrochemical sensors operating at ambient conditions.

In this study, we prepared the highly conductive quasi-solid-state hybrid electrolytes by chemical cross-linking reaction of reactive SiO₂ particles and tetra (ethylene glycol) diacrylate (TEGDA) in aqueous electrolyte. As the reactive SiO₂ particles, we synthesized mesoporous SiO₂ particles containing methacrylate groups on their surface to induce the free-radical reaction with TEGDA [19]. The methacrylate-functionalized SiO₂ (MA-SiO₂) particles dispersed in aqueous electrolyte reacted with TEGDA, resulting in the formation of a three-dimensional network, as schematically demonstrated in Fig. 1. The electrochemical hydrogen sensor was fabricated by employing an optimized quasi-solid-state hybrid electrolyte and Pt-based working electrode. The H₂ sensing performance was then evaluated and compared with that obtained in the electrochemical gas sensor assembled with sulfuric acid.

FE-SEM image of the mesoporous MA-SiO₂ particles is presented in Fig. 3(a). The MA-SiO₂ particles have spherical shape with an average diameter of around 30 nm. From the TEM image shown in Fig. 3(b), it is found that that the mesoporous MA-SiO₂ particles have inner-pore channels that can provide pathways for H⁺ ions to move through the mesoporous SiO₂ particles. The BET surface area and pore volume of the MA-SiO₂ particles were measured to be 292.3 m² g⁻¹ and 1.14 cm³ g⁻¹, respectively.

The MA-SiO₂ particles have reactive C=C double bonds on their surface, and thus they can serve as inorganic cross-linking sites through free radical reaction with TEGDA in the aqueous electrolyte. In order to confirm the chemical cross-linking reaction between mesoporous MA-SiO₂ particles and TEGDA, FT-IR spectra were analyzed before and after thermal cross-linking reaction. The FT-IR spectrum of MA-SiO₂ particles in Fig. 4(a) showed symmetrical stretching vibration of siloxane (Si-O-Si) at 766 cm⁻¹ and asymmetrical stretching vibrations at 1190 and 1082 cm⁻¹. The spectrum also exhibited a small peak at 1636 cm⁻¹, which is a characteristic peak of C=C double bond in the methacrylate group on the surface of MA-SiO₂ particles [24,25], indicating the MA-SiO₂ particles have reactive groups to permit free-radical reaction. Fig. 4(b) presents the FT-IR spectrum of the mixture of MA-SiO₂ particles and TEGDA before thermal cross-linking reaction. In addition to peak of C=C double bond in the MA-SiO₂ particles, the peak corresponding to C=C double bonds in TEGDA could be also observed at 1618 cm⁻¹. The FT-IR spectrum of the quasi-solid-state hybrid electrolyte obtained after thermal cross-linking reaction revealed that the peaks corresponding to C=C double bonds in the MA-SiO₂ particles (1636 cm⁻¹) and TEGDA (1618 cm⁻¹) disappeared after thermal cross-linking, as shown in Fig. 4(c). From these results, it was confirmed that methacrylate groups on the surface of MA-SiO₂ particles reacted with TEGDA through free radical reaction to form the three-dimensional cross-linked network, as demonstrated in Fig. 1.

Fig. 5(a) presents photographs of the cross-linked hybrid electrolytes cured by different amounts of MA-SiO₂ particles. After cross-linking reaction using MA-SiO₂ particles and TEGDA, the electrolyte became highly viscous gel and finally non-fluidic due to the formation of three-dimensional cross-linked networks. It should be noted that the complete solidification without fluidity was not occurred when the MA-SiO₂ content was less than 6 wt.%, which is attributed to low degree of cross-linking at low MA-SiO₂ content. This result suggests that the content of the MA-SiO₂ particles should be higher than 6 wt.% in order to prepare the dimensionally stable quasi-solid-state electrolyte. Fig. 5(b) shows the ionic conductivities of the quasi-solid-state hybrid electrolytes as a function of MA-SiO₂ content at room temperature. For comparison, the ionic conductivity of the liquid electrolyte is also shown in the figure. As shown, the ionic conductivity of the quasi-solid-state electrolyte was decreased with increasing MA-SiO₂ content. The cross-linking reaction causes an increase in the resistance for ion migration due to the formation of three-dimensional networks, which results in a decrease of the ionic conductivity. For the electrolyte systems under study, the optimum MA-SiO₂ content was determined to be 6 wt.% when considering both ionic conductivity and dimensional stability. Accordingly, the quasi-solid-state hybrid electrolyte synthesized by 6 wt.% MA-SiO₂ was used to in assembling the electrochemical hydrogen sensor. The ionic conductivity of the optimized electrolyte was 177 mS cm⁻¹ at room temperature, which is a sufficiently high ionic conductivity for sensor applications.

The electrochemical sensor was assembled using a carbon as a reference electrode. To confirm the capability of carbon as a quasi-reference electrode, we performed cyclic voltammetry of the electrochemical sensor with carbon reference electrode in 3.0 M H₂SO₄ aqueous electrolyte and observed the potential shift as a function of cycle. Fig. 6 presents CVs of the electrochemical sensor as a function of cycle, which were obtained in the presence of H₂ gas. The anodic peak corresponding to the hydrogen oxidation is observed around −0.74 V and the cathodic peak for oxygen reduction appeared at 0.51 V. During the subsequent 5 cycles, the CVs show very stable currents without any signs of drifting or shifting of the peak potential. This result indicates that the carbon quasi-reference electrode exhibits a notable stability and reliability with low potential drift in the aqueous electrolyte. Lee et al. also demonstrated that the porous carbon electrode could be used as reliable quasi-reference electrode in the acid aqueous electrolyte [26].

The electrochemical performance of the H₂ sensor was examined using amperometric measurements by sequentially exposing the sensor to 100 ppm hydrogen in nitrogen and then going to stop of gas flow. Fig. 7 shows the responses of the sensors assembled with liquid electrolyte and quasi-solid-state hybrid electrolyte to the repeated exposure of hydrogen gas. A fast response up to maximum current was observed within 12.0 sec. Upon stop of gas exposure, the current starts to decrease. It is well known that the electrochemical redox processes take place at the electrodes, i.e., oxidation at the working electrode and reduction at the counter electrode [27,28].

These redox reactions result in a flow of electrons as an external electrical current of the sensor, as depicted in Fig. 8. Both electrochemical reactions can only occur where the H⁺ ions, electron (or electrode) and gas phases are all in contact. It is thought that the solubility and diffusion coefficient of H₂ in the liquid electrolyte are larger than those in the quasi-solid-state electrolyte. The pH of 3.0 M H₂SO₄ solution (−0.68) is lower than that of quasi-solid-state electrolyte (−0.54), indicating higher concentration of H⁺ ions in the liquid electrolyte. Moreover, the ionic conductivity is much higher in the liquid electrolyte (0.78 S cm⁻¹) than quasi-solid-state hybrid electrolyte (0.18 S cm⁻¹). Accordingly, the sensor response was more rapid and the initial sensing current was higher in the liquid electrolyte-based sensor than sensor employing quasi-solid-state hybrid electrolyte. This result demonstrates that a type of electrolyte system plays an important role in determining the gas sensing performance of the electrochemical sensor [29,30]. It should be noted that the current response of the sensor with quasi-solid-state electrolyte was more stable during the repeated cycles, indicating the stable and reproducible redox processes at the electrodes. This result suggests that the hydrogen sensor employing a quasi-solid-state hybrid electrolyte can operate for much longer periods of time than the sensor assembled with liquid electrolyte.

In order to understand the electrochemical behavior of the electrochemical sensors with different electrolytes, their EIS analysis was performed. The resulting AC impedance spectra before and after amperometric measurements are shown in Fig. 9. The obtained spectra were analyzed using the equivalent circuit given in the inset of Fig. 9(a). In this circuit, R_E is the electrolyte resistance, which is corresponding to the real axis intercept. R_CT is the charge transfer resistance and CPE_dl denotes the constant phase element of the double layer capacitance to reflect the depressed semicircular shape, and Z_w is Warburg impedance. In the sensor with liquid electrolyte, the values of R_E and R_CT increased after test. The increase of R_E is mainly attributed to the gradual loss of liquid electrolyte during test, resulting from the evaporation of solvent under open H₂ atmosphere condition. At the same time, the electrolyte depletion retarded the charge transfer reaction at the electrolyte-electrode interface, resulting in increase of R_CT. In contrast, the sensor with quasi-solid-state electrolyte exhibited the almost same values in R_E and R_CT before and after test. This result can be ascribed to the effective encapsulation of liquid electrolyte in the three-dimensional cross-linked networks as well as the good interfacial contact between quasi-solid-state electrolyte and electrodes, which resulted in enhanced stability of sensing performance, as presented in Fig. 7(b).

The quasi-solid-state hybrid electrolytes were synthesized using reactive methacrylate-functionalized SiO₂ (MA-SiO₂) and tetra (ethylene glycol) diacrylate (TEGDA). A quasi-solid-state electrolyte synthesized with 6 wt.% MA-SiO₂ and 10 wt.% TEGDA formed the three-dimensional cross-linked networks without fluidity and exhibited a high ionic conductivity of 177 mS cm⁻¹ at room temperature. The electrochemical H₂ sensor was assembled with quasi-solid-state hybrid electrolyte, and its sensing performance was investigated. The quasi-solid-state electrochemical sensor showed relatively fast response and high sensitivity for hydrogen gas, and its sensing response was more stable than the sensor with liquid electrolyte. The simple, low cost, compact and robust construction of the electrochemical sensor and its high performance allowed the sensor to be used in various applications where hydrogen gas is utilized, especially fuel cell vehicles and stationary power generation systems.