Hydrogen is the most common element in the universe, and it has gathered considerable attention as one of the most promising green and renewable energy sources for various industrial applications [1,2]. In particular, it can be used as a clean alternative fuel, making it a promising renewable energy option. However, hydrogen is highly explosive and has a low flammability point (4 vol%) in air [3]. Furthermore, H₂ is odorless, tasteless, and colorless, making it difficult to detect in the atmosphere in everyday life [4]. Therefore, a rapid and precise leakage detection and monitoring system is necessary for safe H₂ production, storage, and use. H₂ is detected using various sensors, including catalytic, electrochemical, mechanical, optical, acoustic, thermal-conductivity, and chemiresistive sensors. Among these, chemiresistive sensors with metal oxide semiconductors, such as SnO₂, ZnO, TiO₂, and WO₃ doped with noble metals, are considerably effective in sensing H₂ [5]. WO₃ is widely used in H₂ sensors because of its high oxygen vacancy diffusion coefficient and good chemical stability [6]. Pristine WO₃ with Pd-functionalized or Pt-functionalized nanostructures can effectively enhance the sensitivity and response of sensors for hydrogen detection [7,8]. However, the operating temperature of such sensors is quite high (150−300 °C), which limits their practical application [9]. This issue is addressed by employing two methods to reduce the operating temperature, i.e., the in-situ functionalization of noble metals on the surface of WO₃ and the development of a porous WO₃ architecture. In-situ functionalization results in strong orbital interaction with WO₃, which enhances electron-transport pathways. The porous architecture of WO₃ nanoparticles enhances the surfaceto- volume ratio, thereby increasing the surface area. This results in efficient physisorption and chemisorption of H₂ and reduces the operating temperature. In this study, pure WO₃ and Pd/WO₃ nanocomposites [with various amounts (wt%) of Pd] were synthesized in situ using a high temperature solvothermal process. The morphologies, structures, and H₂-sensing properties of the as-synthesized nanocomposites were systematically investigated. This study developed a simple method for synthesizing sensing devices that can be expanded to create other metal-oxide-sensing devices.

2.1. Materials and method

An in-situ high-temperature solvothermal growth reaction was used to synthesize Pd/WO₃ nanocomposites. First, 0.5 g of terephthalic acid (98 %, Sigma-Aldrich) and 0.3 g of D-glucose (≥ 99.9 %, Sigma- Aldrich) were dissolved in 40 mL ethanol and 10 mL dimethylformamide followed by the addition of an appropriate amount of WCl₆ (≥ 99.9 %, Sigma-Aldrich) such that the final concentration of WCl₆ was 0.016 mM. Then, a PdCl₂ solution (with a Pd/W ratio of 0.5, 1.0, and 1.5) was added, and the mixture was transferred to a 50 mL autoclave for 24 h at 180 °C. The final precipitates were filtered and washed with acetone and dried in air at 60 °C for 24 h. Thereafter, all the samples were calcined in air at 500 °C for 2 h to obtain the final sensing product. Pure WO₃ was prepared using the same procedure but without the addition of the PdCl₂ solution.

2.2. Fabrication of the gas sensor device and measurement

All the dried powder samples were mixed with ethanol to create a homogeneous thick solution, which was dropped onto a ceramic substrate with Pt interdigital electrodes using a micropipette to form a uniform gas-sensing layer. The obtained gas sensor was aged at 50 °C for 24 h to improve thermal stability. The devices were then placed in a Nextron 100 cc microprobe station connected to a Keithley 2400 multimeter that controlled the voltage with a temperature controller. The mass flow controller maintained a constant total flow of the gas mixture (100 sccm). The gas response (R_s) was calculated using Eq. (1).

where R_a and R_g are the sensor resistances in air and the target gas, respectively.

2.3. Characterizations

The crystal structures of WO₃ and Pd/WO₃ were examined using an X-ray diffraction (XRD) (Rigaku SmartLab) analyzer with Cu Kα radiation (40 kV, 100 mA) at a scanning speed of 0.5°/min. The morphologies and microstructures of the samples were investigated using high-resolution transmission electron microscopy (HRTEM) (Tecnai G2 F20, FEI, USA).

Figure 1 shows the representative XRD patterns of pure WO₃ and the doped Pd/WO₃ nanocomposites. All the diffraction peaks have the same fundamental characteristics as WO₃ with a monoclinic structure (JCPDS No.43-1035) [10]. The diffraction peaks at 2θ = 23.1, 23.7, and 24.2° correspond to the (002), (020), and (200) monoclinic phases of WO₃ crystal planes, respectively. These peaks are quite sharp, revealing the crystalline character of the nanocomposites [5]. No discernible diffraction peaks of Pd or PdO are observed, which may be due to the low Pd concentration [6]. The diffraction peak intensity of the doped Pd/WO₃ composite films is slightly lower than that of pure WO₃, indicating successful functionalization of Pd atoms on the surface of the WO₃ composite with high crystallinity. No other impurity peaks are observed under these experimental conditions, suggesting that the as-synthesized product is highly pure.

Figure 1. XRD patterns of WO³ and Pd/WO₃ nanocomposites with different amounts (wt%) of Pd.

Figure 2 depicts the HRTEM images of the 1.0 wt% Pd/WO₃ nanocomposite. As shown in Fig. 2(a), WO₃ particles with sizes of 50−100 nm are arbitrarily arranged, and they cluster to form a WO₃ structure. Figure 2(b) depicts the selected area electron diffraction (SAED) pattern, which clearly shows that the multiple rings correspond to crystalline WO₃ [10,11]. A high-magnification image of the 1.0 wt% Pd/WO₃ nanocomposite shows the presence of Pd functionalized on the surface of WO₃. The HRTEM images [Figs. 2(c) and 2(d)] show lattice fringes, indicating the high crystallinity of Pd/WO₃. The lattice fringe spacing values are 0.364 and 0.238 nm, which match the d-spacing values of the monoclinic WO₃ (020) and Pd (111) facets, respectively [5,11].

Figure 2. HRTEM image of 1.0 wt% Pd/WO₃: (a) low magnification, (b) SAED pattern, and (c and d) high magnification with d-spacing of WO³ and Pd.

X-ray photoelectron spectroscopy (XPS) was used to determine the surface compositions and chemical states of the as-synthesized pure WO₃ and doped Pd/WO₃ nanocomposites, as shown in Fig. 3. The presence of W, O, and Pd was confirmed by the full-range XPS spectra of all nanocomposites [Fig. 3(a)], which show highly pure nanocomposites. This is supported by the XRD data (Fig. 1). The deconvoluted W 4f spectrum [Fig. 3(b)] consists of peaks at 34.96 and 37.10 eV for pure WO₃, which correspond to W 4f_7/2 and W 4f_5/2, respectively (pink solid line), and indicate the existence of W⁶⁺ in the WO₃ crystal lattice [12]. The addition of Pd to the WO₃ nanocomposite shifts the W 4f peaks to higher binding energy sites (cyan dotted line). The W 4f_7/2 and W 4f_5/2 peaks in all the Pd/WO₃ nanocomposites are located at approximately 35.14 and 37.27 eV, respectively [11]. The shift of the W 4f peak to a higher binding energy confirms that Pd is doped into the WO₃ lattice. Figure 3(c) shows the deconvoluted Pd 3d core spectrum of the 1.0 wt% Pd/WO₃ nanocomposite. The peaks at 336.67 and 342.30 eV correspond to Pd 3d_5/2 and Pd 3d_3/2, respectively [8,9]. These peaks correspond to metallic Pd (Pd⁰). The small peak at 337.69 eV, corresponds to the Pd²⁺ state, which is due to the partial oxidation of Pd during calcination. Figure 3(d) shows the deconvolution of the O 1s peak of the 1.0 wt% Pd/WO₃ nanocomposite into two peaks: one corresponding to the lattice oxygen (O_L) in WO₃ (530.06 eV) and the other at approximately 531.22 eV corresponding to the surface chemisorbed oxygen (O_A) [10–12]. O_A interacts with target gas molecules on the surfaces of metal oxides; therefore, it is essential for improving the efficiency of gas sensing. The area percentages of O_L and O_A for all samples are shown in the inset of Fig. 3(d). As Pd increases, O_A increases and reaches the maximum at 1.0 wt% Pd/WO₃. Then, O_A decreases as Pd increases further owing to the catalytic poisoning caused by excess Pd.

Figure 3. XPS spectra of pure WO³ and doped Pd/WO₃ nanocomposites: (a) full range XPS spectra, (b) W 4f spectra of all the samples, (c) Pd 3d spectra, and (d) O 1s spectra of 1 wt% Pd/WO₃ nanocomposite. Inset shows the corresponding variations in the composition of various oxygen species for all the samples.

Figure 4 shows the gas-sensing analysis of the as-synthesized nanocomposites. In gas-sensing measurements, the operating temperature is a critical parameter that must be tuned for all sensing materials. All of the sensors in this study were evaluated by increasing the operating temperature from 50 to 150 °C at an interval of 25 °C. Figure 4(a) depicts the temperature-dependent response of all sensors to 100 ppm H₂ for different amounts of Pd. All the nanocomposites show a volcano-shaped response. The optimal operating temperature is 100 °C for all the Pd-functionalized samples [12] and 125 °C for pure WO₃. This proves that Pd functionalization has positive effects on the operating temperature. At low temperatures, the sensing processes are slow owing to a lack of energy. The number of free electrons increases with the temperature; this entices surface oxygen molecules and enhances sensing reactions [9]. At temperatures higher than the optimum range, the rate of oxygen molecule desorption becomes more than that of adsorption, which reduces the sensing response of the surface [10]. The highest response value is 12.25 for the 1.0 wt% Pd/WO₃ nanocomposite at 100 °C. Figure 4(b) shows the dynamic transient curve at 100 °C for different concentrations of H₂ (5−100 ppm). As WO₃ is an n-type semiconductor gas sensor, the 1.0 wt% Pd/WO₃ sensor resistance decreases and increases when it is exposed to H₂ (reducing gas) and air, respectively. Figure 4(c) shows the corresponding calibration curves, which depict a linear relationship for all concentrations. The linear relationship curve is extremely useful for calibrating and determining hydrogen concentration in real-world applications [9].

Figure 4. Gas-sensing analysis: (a) response as a function of the operating temperature of all the sensors, (b) dynamic response, and (c) calibration curves for various concentrations of H² for 1.0 wt% Pd/WO₃.

The times required for the total resistance to change by 90 % after exposure to H₂ and air are defined as the response time (T_res) and recovery time (T_rec), respectively. Figure 5(a) shows that the response and recovery times of the 1.0 wt% Pd/WO₃ nanocomposite for 50 ppm H₂ at 100 °C are 73 and 85 s, respectively. An as-fabricated gas-sensing device must exhibit high repeatability to be implemented in practical applications. The dynamic response–recovery curves of the 1.0 wt% Pd/WO₃ sensor were obtained for the detection of 50 ppm H₂ for six cycles at an optimum operating temperature of 100 °C, as shown in Fig. 5(b). The sensor response remains unchanged, which is extremely important for practical applications. Sensors must achieve good selectivity in real-world applications. Therefore, all the sensors were tested at 100 °C in contact with typical interfering gases, i.e., NH₃ (50 ppm) and NO₂ (50 ppm). The 1.0 wt% Pd/WO₃ sensor strongly responds to H₂ compared to the other interfering gases [Fig. 5(c)]. This may be because H₂ has a relatively lower bond dissociation energy and smaller kinetic diameter compared to the other gases, making diffusion reactions easier [6,13]. The strong response at the low operating temperature can be attributed to Pd, which acts as a catalyst and can dissociate more oxygen and hydrogen because of the spillover effect on the surface of WO₃ [14,15]. The XPS results prove that the 1.0 wt% Pd/WO₃ sensor contains the highest chemisorbed oxygen species on the surface. Therefore, a strong interaction occurs between H₂ and the chemisorbed oxygen, which improves the sensing response. Additionally, as the porous Pd/WO₃ nanostructure provides more reaction-active sites on its surface, more adsorbed oxygen can react with hydrogen, thereby improving the sensor response.

Figure 5. (a) Transient response–recovery curve, (b) repeatability test of 1.0 wt% Pd/WO₃ for 50 ppm H2 at 100 °C, and (c) selectivity test of all the sensors.

Pd/WO₃ nanocomposites with different amounts of Pd (wt%) were successfully synthesized using a high-temperature solvothermal method, and their H₂-sensing properties were measured. Compared to pure WO₃, the 1.0 wt% Pd/WO₃ sensor exhibited a fairly high response of 12.25 to 100 ppm hydrogen at an optimal operating temperature of 100 °C. Furthermore, it exhibited satisfactory selectivity for H₂ compared to NH₃ and NO₂. The sensor showed good stability, which demonstrated its potential for practical applications. The improvement in the H₂-sensing property was attributed to the successful functionalization of Pd on the surface of WO₃, which effectively catalyzed the spillover phenomenon of Pd nanoparticles, and to the porous structure of Pd/WO₃, which provided active sites. The results of this study show that Pd/WO₃ nanocomposites exhibit improved hydrogen-sensing properties, and they can potentially be used for detecting hydrogen.

This work was financially supported by the Korea Institute of Science and Technology (KIST) institutional program.

The authors declare no conflicts of interest.