1 Introduction

The rapid development of industry, accompanied by environmental pollution, has led to severe threats to human health, namely respiratory illnesses and diseases. Poor air quality and exposure to toxic pollutants are found to be responsible for thousands of deaths annually, especially in children and older adults. Inhaling various toxic gases, such as H2S, CO, NH3, C2H5OH, and (CH3)2CO, can cause instant and long-term threats to human health; moreover, pollution can cause environmental problems, such as acid rain, ozone layer depletion, global warming, and climate change [1, 2]. Therefore, real-time, reliable, and sensitive sensors are needed to detect toxic gases such as volatile organic compounds (VOCs) and hydrogen sulfide (H2S). Nanomaterial metal oxides have many advantages, such as compact size, low power consumption, low cost, and large surface area [3]. Metal oxide gas sensors are devices that can detect and sense the concentration of dangerous gases in the environment with great accuracy due to their semiconducting properties, allowing for alteration in electrical resistance [4,5,6,7]. Nanomaterials, with their exceptional physicochemical properties, offer potential groundbreaking solutions for fabricating highly efficient and miniaturized gas sensors [5, 8,9,10].

Developing noninvasive techniques for early disease detection and monitoring based on breath gas signatures can potentially diagnose a broad spectrum of diseases. For example, H2S as a biomarker for noninvasive, rapid, and precise evaluation of small intestinal bacterial overgrowth (SIBO) in irritable bowel syndrome (IBS) patients [11], breath ammonia as a biomarker for screening and detecting chronic kidney disease (CKD) [12], and ethanol as a biomarker for various diseases [13] all contribute to this hopeful outlook. Furthermore, breath ethanol is associated with hepatic steatosis [14]. Acetone biomarkers in human breath help diagnose various diseases and disorders, such as starvation, diabetes mellitus, and lung tumors [15]. Despite the significant progress made in breathomics, several challenges remain before it can be fully integrated into routine clinical practice.

Additionally, extensive clinical validation studies are essential to establish the sensitivity, specificity, and clinical utility of breathomics for various diseases [16] and the fast and precise monitoring of sub-ppm biomarker gases in human breath. Numerous studies have explored using metal oxide materials to detect trace concentrations of breath biomarker gases. For instance, Umesh et al. [17] developed a ZnO nanocomposite that employed sense H2S at 1 ppm with a response rate of up to 40%. This composite also exhibited high selectivity toward other gases like CH4 and NH3 breath sensors to detect lung cancer and for food and environmental monitoring [11, 18]. A. Raghu et al. examined VOx/TiO2 nanoflakes for oxygen sensing of VO2. Carbon nanotubes, graphene, metal oxide nanoparticles (NPs), and other nanomaterials have demonstrated remarkable sensitivity and selectivity towards various gases, enabling the detection of trace levels of VOCs [12]. Similarly, Yin et al. [13] reviewed recent advances in nanomaterial-based sensors for detecting hazardous organic compounds in exhaled breath, emphasizing the potential of these sensors for breathomics applications. Zhang et al. [19] They have been found to play a crucial role [20].

Tungsten oxide (WO3) with different morphologies [21] is one of the most used metal oxides that reveal high responses to many biomarkers that exist in the breath, such as acetone, hydrogen sulfide, ammonia, and nitric oxide [22,23,24] due to its propensity to interact with H2S molecules and its low bandgap (2.5-3.0 eV) [25,26,27]. However, it has limitations regarding low concentration sensitivity and high operating temperatures [28]. Loading of other conductive metals such as gold (Au), lead (Pb), and cobalt (Co) enhances the sensor properties. Silver (Ag) has been used to enhance the WO3 sensing properties in many studies to enhance the sensitivity and the sensor's selectivity abilities [29, 30]. Z. Yang et al. examined nanoflowers of SnO2/MoSe2. They reported decent Carbon Monoxide (CO) sensing properties with anti-humidity interference assigned to forming heterojunctions and surface defects among SnO2 and MoSe2 [31]. Ag is well-known as a material that increases the catalytic decomposition of reducing gasses like H2S. [5, 32, 33]. ZnO-cobalt oxide nanocomposite produced via co-electrospinning was investigated for H2S gas sensing, achieving a remarkable detection limit of 23 ppb for H2S. Additionally, it was reported that sensing material based on Ag-doped SnO2 can detect hydrogen sulfide at 0.1 ppm with a 27% response [34]. As far as we know, few studies have reported the synthesis of Ag-doped WO3 (Ag/WO3) as a sub-ppm gas sensing material.

This paper presents Ag/WO3 for detecting toxic gases in the sub-ppm concentrations, which could be used for air quality monitoring or as a disease diagnostic tool. Sensitivity, response duration, selectivity, and stability of the sensing material qualities toward the targeted gases are thoroughly investigated. A comprehensive study of the material attributes regarding morphology, crystalline structure, and material content is also presented. The results show that it is highly sensitive toward various toxic gases and highly selective toward H2S at low temperatures and concentrations in dry and humid conditions. As a result, such sensors appear to be a viable option for useful practical applications.

2 Experimental

2.1 Materials

Sigma-Aldrich provided silver nitrate (AgNO3), tungsten oxide powder (WO3), and sodium hydroxide (NaOH). All synthesis procedures were carried out with double distilled water.

2.2 Synthesis of nanoparticles and experimental setup

The WO3-containing Ag nanoparticles are synthesized using deposition precipitation techniques [35]. 1.0 g of WO3 was dissolved in 10 mL of distilled water (DI). The solution was continuously stirred in a 50 mL beaker. Gradually, 1.0 mL of NaOH was introduced dropwise until a brown color changed. Next, the solution was placed in a microwave chemical reactor (MCR-3) with microwave power limited to one-third of the machine's rated output (800W). The solution underwent microwave radiation for 5-10 minutes, with 30-second cycles, until reaching the boiling point. After cooling, the solution precipitated at room temperature for 24 hours. The recovered particles were filtered and washed with ethanol and distilled water four times. Finally, 0.05 g of AgNO3 was added to achieve a 3 wt. % Ag concentration promotes Ag growth on the WO3 surface and produces a WO3 precursor containing Ag particles.

2.3 Materials Characterization

In the N2 sorption analysis conducted at 77 K using an accelerated surface area (Micromeritics ASAP2020) and porosity system (Micrometrics, Norcross, GA, USA), the Brunauer–Emmett–Teller (BET) method measured the specific surface area, while the Barrett–Joyner–Halenda (BJH) method determined the pore size distribution. X-ray diffraction (XRD) that employed Cu–Kα emission line (λ = 0.15418 nm) within the range of 10.0° to 99.9° provided insights into material structure, crystallography, and phase analysis. Additionally, a scanning electron microscopy (SEM) model FEI Quanta 650 FEG and a transmission electron microscopy (TEM) model FEI TECNAI G2 and TF20 microscope that allowed examination of the sensing material's morphology and elemental mapping. High-resolution transmission electron microscopy (HRTEM) (FEI Talos 200X) and X-ray photoelectron spectroscopy (XPS) (Thermo Fisher Scientific, USA) were employed to assess the chemical state, concentration, and surface conditions of silver nanoparticles. Data analysis and peak fitting utilized Casa XPS software, with XPS spectra calibrated at 284.6 eV to the C 1s feature. Finally, the prepared samples' UV–Vis absorption spectra were evaluated using a UV–Vis split-beam spectrophotometer (YK Scientific, UV1810/UV1810S, China).

2.4 Sensor Fabrication and Testing

Permeation tubes test sensors at extremely low concentrations, producing highly accurate results as low as 50 ppb. The sensor is put in a testing chamber with a heating plate for accurate temperature adjustment. A gas generator (Dynacalibrator VICI 340) uses a permeation tube filled with concentrated gas (in liquid form, such as H2S) to create exact gas concentrations. Temperature changes and dilution gas flow from a synthetic air cylinder regulate concentration. Heating the permeation tube produces gas vapor; changing the dilution airflow allows concentrations to be adjusted. Permeation tubes serve several functions, including air quality monitoring and gas sensor calibration. Comparable research shows an illustration of the experimental setting [10]. Our experiment focused on how the synthesized sensors responded to various H2S temperatures, concentrations, and humidity levels. We investigated how different circumstances affect Ag-doped WO3 (Ag/WO3) nanoparticles and their gas-sensing properties. The sensor response factor, or sensitivity, is defined by the difference in resistance (ΔR) between the sensor's resistance in air (Rair) and the resistance in test gas (Rgas). Because WO3 and Ag/WO3 are n-type semiconductors, the sensor's reaction to reducing gases (such as hydrogen sulfide, ammonia, and ethanol) can be calculated using equation (1). However, the response to an oxidizing gas like acetone follows a different equation (2) [31].

$$\textrm{S}\ \left(\%\right)=\left|\frac{R_{air}-{R}_{gas}}{R_{gas}}\right|\times 100\%$$
(1)
$$\textrm{S}\ \left(\%\right)=\left|\frac{R_{air}-{R}_{gas}}{R_{air}}\right|\times 100\%$$
(2)

The sensor's selectivity to one gas (gas (a)) relative to another interfering gas (gas (b)) can be expressed as the percentage of the sensor's best response to its second-best response or as the percentage of the sensor's response to gas (a) relative to gas (b)[36].

3 Results and discussion

3.1 Characterization of Ag/WO3 nanoparticles

The BET technique calculates the surface area of the samples. The adsorption-desorption experiments of N2 demonstrate the surface area and pore structure of Ag/WO3 nanostructures. In particular, the BET surface area of Ag/WO3 nanostructures is 17.3 m2/g, indicating a substantial surface area. This large surface area makes the material appealing for gas-sensing applications, providing more gas-sensing sites.

XRD was used to analyze the crystal structure of the Ag/WO3 and WO3 phases in the presence of varying silver loadings. The resulting XRD spectrum [37], as seen in Fig. 1a, displays a characteristic pattern corresponding to the tricyclic phase of WO3. Notably, several peaks representing WO3 crystals and others are observed at 22.90, 25.00, and 35.00 for the (002), (020), and (200) refraction indices of the monocyclic crystal structure, respectively. These peaks fit well with the diffraction pattern of the monocyclic tungsten oxide crystal (JCPDS 43-1035).

Fig. 1
figure 1

XRD of tungsten oxide (WO3) doped with silver (Ag) loadings (a) pure WO3 nanoparticles and (b) WO3 doped with 3 wt.% Ag NPs

Full size image

However, the XRD of Ag/WO3 showed that the intensity of the peak WO3 decreases upon doping with Ag, indicating that Ag was substituted in WO3 as shown in Fig. 1b, which illustrates Bragg diffraction peaks at 2θ values of 27.81° and 32.16° which are indexed to (210) and (122) planes, respectively, determined for the pure silver using the face-centered cubic structure (JCPDS, file No. 04-0783) [38] and Ag (JCPDS, file No. 97-005-3759) [39], indicating the presence of 3 wt. % Ag. According to the information provided by (JCPDS card No. 41-1104); the peaks in Fig. 1b refer to the face-centered phase of the silver cubic structure [40]. The XRD patterns of WO3 are its monoclinic structure. The Debye-Scherer equation formula determines the size of pure WO3 and Ag/WO3 crystals [41]. Analyzing the (200) and (020) peaks, the average crystallite size of pure WO3 is 35 ± 10 nm. Similarly, the average crystallite size of Ag is 9 ± 2 nm, examining the peaks at (122) and (210).

SEM was used to study the morphology of the samples. The synthesized Ag/WO3 nanostructured morphology is characterized in Fig. 2. From the SEM micrograph, it is observed that the Ag/WO3 resembles plate-like morphology and is composed of large particles with average sizes of 100 nm–1μm and small particles with approximate average sizes of 50–300. Furthermore, SEM-EDX results showed that the overall sample contains WO3 and Ag/WO3 regions and confirmed that the obtained material comprises W, O, and Ag. [5]

Fig. 2
figure 2

Scanning electron microscope (SEM) of Ag/WO3 at different magnifications

Full size image

To study the morphology further, TEM was used to investigate the morphology, content, and size of the produced Ag/WO3 nanoparticles. Fig. 3 shows an HRTEM image of Ag/WO3 nanoparticles. The results clearly show an irregular shape and size WO3 sheet-like morphology with a size of about 150-200 nm. However, no Ag nanoparticles are observed. Analysis of HRTEM images (Fig. 3b and d) confirms that the interplanar spacing of WO3 is 3.86 nm for the crystal plane [020] [42]. Moreover, selected area electron diffraction (referred to as SAED) (Fig. 3c) clearly showed that the existence of the spots is mainly attributed to well-crystalline underlying WO3. The SAED pattern also indicates that the prepared material exhibits well-crystallized nanoparticles grown along the crystal plane [002] [43], [002] [44], and crystal plane [020] directions [42].

Fig. 3
figure 3

(a, b, d) HRTEM image of Ag/WO3 nanoparticles, (c) SAED pattern of Ag/WO3 nanoparticles

Full size image

To investigate the presence of the silver nanoparticles in the sample, TEM-EDS mapping and HAADF (High-angle annular dark-field imaging) were performed on the sample. Fig. 4 shows the TEM-EDS mapping and HAADF. The results reveal that the silver nanoparticles are small (< 3 nm) and well dispersed within the WO3 nanosheets [45]. These results are consistent with previous research on the same material using various synthesis techniques [40]. These results confirm the XRD results.

Fig. 4
figure 4

EDS mapping of TEM and HAADF images of particle dispersion in Ag/WO3 nanoparticles

Full size image

XPS investigated surface chemical states and compositions of Ag/WO3 nanoparticles. The measurements can elucidate surface structure and active species down to a depth of less than 10 nm, which is significant for gas detection. This requires ejecting surface elements upon exposure to X-rays and observing the energy patterns of the photoelectrons. Figure 5 shows the XPS study of Ag/WO3. Scanning of the XPS spectrum of the Ag-unloaded WO3 sample showed Ag, W, and O. The presence of C and a small amount of Na suggested possible contamination.

Fig. 5
figure 5

XPS analysis of fresh Ag/WO3 sample and the peak fitting to analyze the data. (a) shows the XPS spectrum survey of the fresh sample, while Figures (b-d) depict high-resolution scans of W, Ag, and O, along with the conforming deconvoluted peaks

Full size image

XPS spectra of W 4f, Ag 3d, and O 1s were deconvoluted and plotted separately in Fig. 5b, c, and d. In the W 4f spectra, two helical doublets corresponding to the valence states of W 4f7/2 and W 4f5/2 were detected with cutoff energies of 34.9 eV and 37.1 eV, and the doublet separation was 2.2 eV. Deconvolution of the W 4f5/2 and W 4f7/2 spectra revealed that most of the analyzed tungsten oxide exists as W5+ with cutoff energies of 34.9 eV and 37 eV. A small concentration of W6+ was observed at voltages of 34.9 eV and 38.1 eV. This study shows that W5+ is the transcendental valence state of tungsten in the tungsten oxide assay [46]. Analysis of the silver nail from the sample shows that it consists mainly of silver oxide (Ag2O) and a small amount of metallic silver (Ag). Ag 3d XPS spectra showed two spin-orbit doublets corresponding to Ag 3d5/2 and Ag 3d3/2 binding energies of 367.5 eV and 373.5 eV, respectively, and a doublet splitting of 6.0 eV [47]. The significant splitting of the Ag 3d doublet indicates the presence of Ag2O. Deconvolution of Ag 3d5/2 and Ag 3d3/2 spectra showed peaks corresponding to Ag1+ (Ag2O) binding energies of 367.4 eV and 373.4 eV, respectively, and a small percentage of Ag0 (metallic Ag) was attributed to binding energies of 368.2 eV and 374.0 eV. A binding energy peak at 529.8 eV was observed in the O 1s spectrum, which was deconvolved into three peaks at 529.8 eV, 531.4 eV, and 532.7 eV. These peaks can be attributed to oxygen bound to W6+ in the stoichiometric WO3 phase [48] to oxygen vacancies in the non-stoichiometric WO3-x phase with higher surface oxygen [47], and to the Surface oxygen content in tungsten hydroxide (W-OH) hydroxyl groups [49]. These observations confirm the presence of silver oxide (Ag2O) and a small amount of metallic silver (Ag) in the nanoparticles, confirming the TEM analysis results.

The optical properties of Ag-loaded WO3 samples were investigated using UV-Vis absorption spectra at 310–950 nm wavelengths. The optical absorption spectra of Ag/WO3 are shown in Fig. 6. The nanoparticle's size, shape, dielectric properties, and the dielectric permittivity of the surrounding medium affect the surface plasmon resonance (SPR). In the visible light spectrum, a peak is observed at a wavelength of 350 nm. Fig. 6b illustrates the tauc plots used to calculate the bandwidth energy, showing the relationship between (αhυ)1/2 and the optical photon energy (eV), where α is defined as the absorption and hυ is the incident photon energy. In our experiment, we choose where the tangent line is drawn at the Tauc plot curve's inflection point, the tangent line's intersection point (eV value), and the horizontal axes are band gap values. However, small direct band gap energies exist due to the presence of Ag/WO3 nanoparticles and the dispersion in Ag nanoparticles heterojunctions and two indirect band gap energies related to WO3-x domains due to the presence of oxygen vacancies in WO3-x, which XPS data confirmed.

Fig. 6
figure 6

(a) UV-VIS absorption spectrum results of WO3 nanoparticles with 3% Ag loading, (b) Tauc plot of Ag/WO3 nanoparticles is shown, and the orange line represents the linear fit in the bandgap calculation

Full size image

The figures indicate a band gap of 2.55 eV with an Ag content of 3 wt. %. Silver reduces the energy required to move electrons from the valence to the conduction band.

3.2 Gas Sensing Performance for H2S, NH3, C2H5OH, and (CH3)2CO Detection

Dynamic change of resistance and gas-sensing response are used to evaluate the gas-sensing characteristics. Initially, the Ag/WO3 sensors were exposed to each gas in dry air. To find the optimum conditions to ensure stable and well-reproducible resistance changes representing sensor signal. Fig. 7 shows a dynamic response of the Ag/WO3 sensor to the periodical change between gas concentrations of 0, 0.25, 0.5, and 1 ppm of the tested gases in dry air at room temperature. All gases showed a typical response of n-type oxides towards a reducing gas such as hydrogen sulfide, ammonia, ethanol, or an oxidizing gas such as acetone: the resistance decreases upon exposure to reducing gases or increases upon exposure to oxidizing gases and returns in dry air. All gases showed stable and reproducible response signals.

Fig. 7
figure 7

Measurement of Ag/WO3 response to (a) hydrogen sulfide, (b) ethanol, (c) ammonia, and (d) acetone

Full size image

Fig. 8 shows the maximum response of the target gases measured at room temperature at 1 ppm. In particular, H2S has the most significant response at all concentrations, while ethanol, ammonia, and acetone have the lowest responses. In addition, the reaction shows direct proportionality with the gas concentration, except for Fig. 8d at 0.25 ppm.

Fig. 8
figure 8

Room temperature Ag/WO3 sensor response to different gasses against ppm concentration

Full size image

At 25 °C and in dry air, the sensor based on Ag/WO3 thin film responds dynamically to different H2S concentrations (0.25, 0.5, and 1 ppm). Upon replacing the H2S gas with air, the reaction approaches the maximum equilibrium resistance value and returns to the base resistance. Specifically, the sensor records 24.8%, 58.4%, and 89.8% response rates for H2S concentrations of 0.25, 0.5, and 1 ppm, respectively. The results show that the response of the gas sensor decreases with increasing gas concentration, confirming that the presence of H2S reduces the n-type WO3 nanomaterial. Thus, this material can effectively quantitatively analyze atmospheric hydrogen sulfide.

Sensor selectivity is a critical property for target gas detection. The Ag/WO3 thin film sensor response was evaluated for three other gases (acetone, ethanol, and ammonia) at a fixed concentration of 1 ppm to evaluate the selectivity at room temperature. As shown in Figure 9, the response rates of the sensor were 89% for H2S, 24.9% for acetone, 21% for ethanol, and 11% for ammonia. In particular, the sensor's selectivity for H2S, calculated as the ratio of its response to H2S gas and ethanol, is 8.09, the highest selectivity despite its low concentration compared to other gases. This indicates that the material can be effectively used in multi-gas scenarios to detect H2S accurately.

Fig. 9
figure 9

Selectivity of Ag/WO3 towards 1ppm concentration of biomarker gasses at room temperature (25°C)

Full size image

3.2.1 Sensor's response and recovery time

Sensor response and recovery time are critical to assessing the sensor's ability to detect the presence of gas. The Ag/WO3 thin film's response and recovery times were evaluated at different gas concentrations under dry conditions. Specifically, the response time represents when the sensor reaches 90% of the highest response value. At the same time, the recovery time represents when the response reaches 10% of the minimum value (the signal in Fig. 7 at low concentrations is narrow, yet zooming out reveals that the 90% of saturation is reached). Higher gas concentrations do not increase the measured reaction time (Fig. 10). H2S has a longer response time than other gases because it takes longer to reach a 90% peak. NH3 exhibits the shortest response and recovery times. Overall, the sensor response and recovery times were highest for H2S, followed by ethanol, ammonia, and acetone. These response and recovery times are longer than those reported for NH3 [50] and H2S [7, 51].

Fig. 10
figure 10

Response and recovery time of the produced thin film sensor for selected gasses at a concentration of 1 ppm measured at room temperature (25°C)

Full size image

3.2.2 Effect of temperature on H2S detection on Ag/WO3 sensor

We investigated the response of the fabricated sensors to H2S at different temperatures. The sensor was tested under dry conditions at various temperatures to determine its optimal operating temperature. Fig. 11 illustrates the dynamic response of Ag/WO3 in dry air at different operating temperatures. In Fig. 11a, the sensor's response in the gas test chamber at 25 °C, 80 °C, and 150 °C with an H2S concentration of 0.5 ppm shows that response at room temperature is almost the same as in the test chamber. The optimum reaction temperature is 150 °C (60%). This result shows that the material remains sensitive to H2S gas even at low temperatures, which reduces the current consumption of the sensing device compared to the optimal operating temperatures of ethanol and ammonia [52,53,54] and acetone, which was 180 °C 200°C, and 350°C respectively. The longer reaction and recovery times observed at lower concentrations, as shown in Fig. 11b, may be due to the additional time required to cover the material's surface due to the low gas content in contact with the sensor's surface [28]. As the concentration increases, the flow rate of gas molecules covering the surface increases, which shortens the reaction time. On the contrary, at higher concentrations, the recovery is longer because more gas molecules are adsorbed on the surface of the thin film, which requires a longer desorption time. Fig. 11b shows the variation in response time with H2S concentration in the test chamber. At 80°C, the sensor shows the fastest response and recovery time compared to other temperatures. However, the reaction time remains almost the same, and the effect of temperature is minimal. These results can be explained by the presence of pre-absorbed water on the surface of the sensing layer, which influences the sensor response and recovery time at different temperatures. The adsorption of water molecules on the surface of the Ag/WO3 sensing material is highly dependent on the synthesis conditions and can be attributed to the inhomogeneity of the sensing material, which usually consists of grains and vacancies; these vacancies in the structure provide direct ducts for water molecules to be adsorbed during synthesis and flow in from the environment. Water molecules' adsorption on the surface occurs via a dissociative chemisorption process. First, water molecules adsorbed on the grain surface react with the oxygen from the lattice to produce oxygen vacancy and double-ionized oxygen (O2-), which reacts with the H+ from the water molecules to form a hydroxyl group (OH-) via dissociation. The pre-adsorbed water molecules on the surface of the sensing material will not donate electrons to the sensing layer, which reduces the sensitivity of the metal oxide sensors [55]. Since WO3 is an n-type semiconductor with a low conduction band level that is more positive than the given potential for the single electron reduction of O2, the multi-electron reduction of O2 will move electrons accumulated at Ag nanoparticles to surface-adsorbed oxygen molecules [56]. Furthermore, the adsorption of water molecules reduces the chemisorption of oxygen species on the Ag/WO3 surface due to the decreased surface area responsible for the sensor response. The pre-adsorbed water molecules also act as a barrier against H2S adsorption, slowing the migration of the H2S molecules on the Ag/WO3 surface. This results in a decrease in the sensitivity of the gas sensors and an increase in their response and recovery times. The Ag/WO3 sensor's response is highly dependent on the gas environment, where pre-adsorbed water at the surface of the sensing material results in the formation of terminal hydroxyl groups. The effect of H2S exposure in the presence of H2O is increased sensor resistance due to the competitive reaction between water vapors with H2S on the sensing layer where Ag (surface) presents an Ag/WO3 site on the surface, the consumed hole h+, Ag (surface)-OH the formed terminal hydroxyl groups, and an available site on the surface S. As a result, the sensor sensitivity to H2S increases, influencing the sensor response and recovery time. However, such an effect was found to be highly dependent on the reaction temperature [55]. At room temperature, the oxygen and water pre-adsorption is increased, and the adsorption/desorption kinetics of H2S at the surface of the sensing material are decreased. The sensing layer responds to H2S, and its desorption at room temperature is reduced. This may be attributed to the reaction of adsorptive H2S with chemisorptive oxygen when exposed to the sensor surface. These reactions likely compete with the dominant sulphurization process. H2S reacts directly with Ag/WO3 to produce metallic Ag and tungsten sulfide. The reactants of both processes discussed above are recovered at room temperature by removing H2S and exposing the Ag/WO3 sensing layer to the ambient conditions. Upon increasing the temperature, the adsorption/desorption kinetics, sensor response, and recovery time increase while the response time decreases. Most pre-adsorbed water evaporates at 150 0C, which explains the recovery time increase compared to room temperature.

Fig. 11
figure 11

The H2S response curve at 1 ppm concentration and 25, 80 and 150 °C temperatures. (b) The response and recovery times of the Ag/WO3 sensor for H2S at 1 ppm and at room temperature, 80 and 150°C

Full size image

3.2.3 Effect of humidity on H2S detection on Ag/WO3 sensor

Water vapor significantly impacts the design and expansion of chemically resistant gas sensors using metal oxides. The presence of moisture molecules in the surrounding environment can significantly change the electrical properties of the sensor material, affecting its response to target gases. The sensor response was investigated. The sensor's response measurements were performed using 0.5 ppm H2S at room temperature (25°C) in dry conditions and then at 60% relative humidity (RH). Fig. 12a illustrates that the response in dry conditions (58.2%) is higher than in humid conditions (36.6%). Remarkably, the material still provides an adequate response to H2S even at high humidity levels. The behavior was observed for ethanol, ammonia, and acetone, where the response decreased gradually in humid conditions [52,53,54]. The response and recovery times are reduced under humid conditions, as shown in Fig. 12b. However, the H2S reactivity of Ag/WO3 nanoparticles decreases by 40% in a humid environment, indicating competition for active sites between water molecules (forming hydroxyl groups) and H2S. However, the response and recovery times decrease due to the lower response signal, indicating that the sensor responds and recovers faster in humid conditions. The effect of humidity on sensor sensitivity correlates well with the relative surface distribution of hydroxyl groups and oxygen species [57]. Notably, the sensor response to H2S recovers when humidity is removed, demonstrating the reasonable sensing stability of the produced material. It should be noted that arrays of multiple sensors should be integrated to produce commercial sensors that provide a fingerprint of the levels of the most relevant volatile biomarkers.

Fig. 12
figure 12

H2S response curve at 0.5 ppm concentration for dry and humid ambient conditions

Full size image

4 Conclusion

We successfully synthesized Ag/WO3 nanoparticles using a microwave-assisted route method and subsequent deposition techniques to enhance toxic gas sensing performance. A systematic study of the sensitivity of Ag/WO3 fabricated sensors to sub-ppm gas concentrations was performed. The material showed an optimal response even at room temperature (RT). Ag/WO3 nanoparticles achieved a remarkable sensor response of 89% at 1ppm H2S, with fast recovery and gas sensitivity at low concentrations and temperatures. In addition, the Ag/WO3-based sensor showed excellent selectivity for ammonia, ethanol, and acetone gases. Our study confirms that the Ag/WO3 gas sensor effectively monitors low concentrations of hydrogen sulfide, ethanol, ammonia, and acetone gases at room temperature, which can help indoor air quality and health monitoring.