Light-Activated Room Temperature Surface Acoustic Wave H₂S Sensor Based on Bi₂S₃ Nanoribbons

Abstract

The expansion of the Internet of Things (IoT) has rendered wireless passive, highly stable, and room-temperature gas sensors indispensable for sensor applications. In this work, a room-temperature surface acoustic wave (SAW) H₂S sensor based on a thin film of nano-mesh woven with Bi₂S₃ nanoribbons was successfully designed and prepared. The impact of varying inorganic salts solution ligand substitution of long-chain organic ligands of Bi₂S₃ films on performance was assessed. Notably, the responses of the sensors following ligand substitution exhibited improvement to varying degrees. In particular, the Cu(NO₃)₂-treated sensor to 10 ppm H₂S was 203% of that of the untreated sensor. Furthermore, the impact of visible light activation on sensor performance was assessed. The results show the sensor has a high sensitivity to H₂S molecules under yellow light activation at room temperature, with excellent selectivity, fast response speed and low detection limit. The sensor exhibited a response to 10 ppm H₂S under yellow light activation that was approximately equal ~ two times greater than the response observed in a dark environment. This work provides a novel approach to enhance the performance of room-temperature SAW H₂S sensors.

1. Introduction

Hydrogen sulfide (H₂S) is a toxic, corrosive and irritating odor gas that is commonly found in human daily life and industrial production [1]. High concentrations of H₂S have been reported to cause damage to the animal/human nervous system, thus posing a threat to animal/human health [2]. In addition, H₂S is a potential marker for certain diseases in human exhaled breath [3]. The monitoring of trace amounts of H₂S in exhaled gas has been shown to provide useful diagnostic information for potential diseases (such as respiratory diseases, metabolic disorders) [4]. Consequently, there is an increasing demand for the detection of low H₂S concentrations at room temperature in many fields, including chemical, industrial and medical diagnostics.

Advances in technology have resulted in the development of a wide variety of gas sensors, including semiconductor [5], electrochemical [6], infrared optical [7], quartz crystal microbalance (QCM) [8] and surface acoustic wave (SAW)-type sensors [9]. Among them, surface acoustic wave (SAW) sensors have been extensively utilized for the detection of gases, humidity and biological substances due to their high stability and easy integration [10,11]. Furthermore, SAW devices have the capacity to be wireless and passive, with the possibility of supporting remote sensing, as well as stable operation under harsh conditions [12]. Consequently, SAW gas sensors are expected to provide fast and selective detection of H₂S sensors operating wirelessly at room temperature for the next generation of the Internet of Things (IoT).

For gas sensors, the selection of a suitable material is paramount to the effectiveness of the sensor [13]. The material must exhibit high selectivity and high adsorption capacity for the target gas to ensure an accurate and reliable sensor. Recent years have witnessed a rapid development in the use of various low-dimensional materials deposited on the surface of SAW devices for the purpose of gas detection, such as polymers [14], metal oxide nanomaterials [15] and various carbon-based nanomaterials [10,16]. However, the majority of these materials are incompatible with high sensing response and fast response at room temperature, thus limiting their practical application in real-time gas sensing.

The utilization of light activation energy to enhance the gas adsorption/desorption kinetics of SAW sensors can result in sensors with high sensitivity and expedited response/recovery speed [17]. For instance, Jakubik et al. reported a light-activated SAW sensor based on photoconductive polymer films for DMMP sensing prepared by spray deposition of polymers on the surface of SAW devices [18]. Pasupuleti et al. reported an SAW for NH₃ gas detection under UV irradiation at room temperature [19]. Under UV irradiation, the ZnO@MXene SAW sensor shows significant improvement in frequency response and selectivity. Zhang et al. reported a GO-MoS₂/SnO₂ composite film for the detection of NO₂ at room temperature under ultraviolet (UV) light irradiation [20]. The key sensing mechanism is mainly the generation of a large number of photoelectron–hole pairs under UV light irradiation, which accelerates the adsorption and reaction process of NO₂ molecules.

However, the majority of reports on light-activated SAW gas sensors have focused on NO₂ and NH₃ gases. There are no other reports on light-activated SAW H₂S sensors operating at RT using visible illumination. Bismuth sulfide (Bi₂S₃) is a semiconductor material with safe and non-toxic properties, high absorption coefficient, low band gap (~1.3 eV), abundant reserves and layered structure, which is widely available [21,22,23]. This material has emerged as an attractive material for supercapacitor, photodetectors and gas sensors [22,23,24]. Consequently, the integration of Bi₂S₃ with SAW devices under photoactivation conditions offers the potential for the development of a highly sensitive, rapid H₂S sensor operating at room temperature.

In light of the aforementioned considerations, one-dimensional (1D) Bi₂S₃ nanoribbon structures were synthesized in this work through a straightforward solvothermal method. A light-activated SAW H₂S sensor based on Bi₂S₃ nanoribbons was successfully designed and prepared. It is noteworthy that the Bi₂S₃ nanoribbons overlap with each other to form a reticulated film, which brings a large number of adsorption sites for the adsorption of H₂S molecules. The removal of long-chain organic ligands from the Bi₂S₃ surface by ligand exchange facilitates the contact of H₂S molecules with Bi₂S₃ nanoribbons, improving the sensor response. Furthermore, light activation has been shown to generate additional carriers of the Bi₂S₃ nanoribbons, which subsequently combine with oxygen present in the air to form adsorbed oxygen. This additional adsorbed oxygen is also involved in the gas-sensitive reaction, which improves the response of the sensor and reduces the response/recovery time. The sensor exhibited a response to 10 ppm H₂S under yellow light activation that was approximately equal ~ two times greater than the response observed in a dark environment. The sensor response/recovery time is only 27 s/21 s for 10 ppm H₂S under light activation. In addition, the sensor also provides excellent selectivity and reliable repeatability. The efficacy of the light-activated SAW H₂S sensor based on Bi₂S₃ design in enhancing the response/recovery speed and sensitivity of gas sensors is well documented.

2. Materials and Methods

2.1. Chemicals and Instruments

Triphenyl bismuth (C₁₈H₁₅Bi, 98%), dibenzyl disulfide (C₆H₅CH₂SSCH₂C₆H₅, 98%) and oleyl amine (OLA) were purchased from Shanghai Aladdin Biochemical Technology Co., Ltd. (Shanghai, China). Polyvinyl pyrrolidone ((C₆H₉NO)_n, PVP, MW ≈ 29,000) was purchased from Macklin Biochemical Technology Co., Ltd. (Shanghai, China). The hydrogen gas, nitrogen dioxide, hydrogen sulfide gases and ammonia gases with 2% concentration were provided by Gold Valley Gas Co., Ltd. (Shenzhen, China).

The microscopic surface morphology of Bi₂S₃ films were characterized using field emission scanning electron microscopy (FE-SEM, Supra 55 Sapphire, Zeiss, Baden-Württemberg, Germany). The crystal structures of the Bi₂S₃ nanoribbons were investigated using a high-resolution transmission electron microscope (HR-TEM, Tecnai G2 F20, Hillsboro, OR, USA) and an X-ray diffraction (XRD, X’pert pro, PANalytical, Almere, The Netherlands) with Cu Kα radiation source (40 kV and 40 mA). An X-ray photoelectron spectroscopy (XPS, Escalab 250Xi, Thermo Fisher, Dartford, UK) was used to investigate the chemical state of the Bi₂S₃-sensitive films.

2.2. Finite Element Analysis and Preparation of SAW Devices

A simplified three-dimensional finite element (FEA) model of the SAW delay line device was successfully achieved using COMSOL Multiphysics 6.0 software. The model is comprised three modules: Piezoelectric Effect module, Solid Mechanics module and Electrostatics modules [25,26]. A frequency-based solver is used to solve the total displacement [27]. As shown in Figure S1, the computational domain of the model consists of a Y-128° cut LiNbO₃ piezoelectric material and a pair of interdigital transducers (IDT) electrodes and a perfect match layer (PML). The width of the SAW model (x-axis direction) is equal to the device wavelength (λ = 16 μm). The electrodes are designed with a width and spacing of 4 μm, and the electrodes are 80 nm in thickness. A 25 μm PML is applied to the bottom boundary of the model, while all boundaries are left free. Considering the computational volume of the model, the depth of the piezoelectric substrate and electrodes in the y-axis direction was set to 1/8 of the device wavelength (2 μm). A mapping approach is used to minimize computational complexity while ensuring simulation accuracy. As shown in Figure S2, the two-dimensional computing domain of S₂₁ signal of SAW model consists of a Y-128° cut LiNbO₃ (12,080 μm × 120 μm) coated with a 20 μm PML on the lower, left and right edges of the mode. All the boundary conditions are free boundary conditions. On the upper surface of the piezoelectric substrate, the left end is the input IDTs, and the right end is the output IDTs. The input-output IDTs consist of 20 pairs of fingers with an electrode width and spacing of 4 μm and a thickness of 80 nm. Input IDT plus 1 V sinusoidal signal, output IDT in floating potential condition.

SAW devices were fabricated using conventional photolithography and lift-off processes, based on our previous work [28]. The SAW device consists of 50 pairs of input and output fingers. The wavelength of the SAW device is 16 µm and the ratio of metal electrode width to slit width is 1:1.

2.3. Preparation of Bi₂S₃ Nanoribbons

A simple and straightforward solvothermal method was utilized to prepare one-dimensional Bi₂S₃ nanoribbons [29]. As demonstrated in Figure 1a, the triphenyl bismuth (C₁₈H₁₅Bi, 1.2 mM) and dibenzyl disulfide (C₆H₅CH₂SSCH₂C₆H₅, 1.2 mM) were dissolved in 8 mL of OLA to yield solution A. Subsequently, 0.8 g of polyvinylpyrrolidone (PVP) was ultrasonically dissolved in 52 mL of anhydrous ethanol to obtain solution B. Finally, the above A and B solutions were stirred and mixed, and then transferred to a Teflon-lined autoclave reactor and reacted for 8 h at 200 °C. Subsequent to the conclusion of the reaction, the product was repeatedly washed three times using anhydrous ethanol. The final product was uniformly dispersed in ethanol solvent to obtain a 50 mg/mL solution of Bi₂S₃ nanoribbons.

2.4. Fabrication and Measurement of SAW H₂S Sensors

An amount of 50 μL of Bi₂S₃ nanoribbon solution was covered the central sensitive region (the delay region between input IDT and output IDT) of the SAW delay line device, ensuring complete coverage. Then, Bi₂S₃ was uniformly deposited on the surface of the SAW device by spin-coating at a rotational velocity of 1000 rpm for a duration of 90 s. The above steps were repeated to spin-coat two layers of Bi₂S₃ nanoribbon-sensitive films. Subsequently, a solution of Pb(NO₃)₂ in methanol (10 mg/mL) and pure methanol was deposited on the SAW sensor in an alternating manner, followed by a spin-coating at 1000 rpm for 100 s. This process was repeated thrice to ensure the removal of ligands that had been capping the surface of the Bi₂S₃ films to obtain a Pb(NO₃)₂-treated SAW H₂S sensor. Similarly, Mg(NO₃)₂-treated, NaNO₂-treated and Cu(NO₃)₂-treated SAW H₂S sensors can be prepared separately.

The gas test system is illustrated in Figure 1b. The SAW sensor is interconnected with a network analyzer via coaxial RF cable. The computer collects the SAW sensor signals and outputs them in real time through a test program. The test data of SAW devices are collected by network analyzer, transmitted and recorded in real time through GPIB bus and computer test system. The frequency change before and after adsorption of H₂S is directly output through the computer. The test chamber is equipped with LED light sources of varying wavelengths (395 nm, 465 nm, 520 nm, 590 nm and 625 nm), which are powered by a signal generator. During the light-activated gas-sensitive test, the optical power density was 20 μW/cm². The gas sensitivity test is conducted in a static mode with a chamber volume of 18 L. For example, 9 mL of 2% H₂S gas is injected into the chamber through a syringe, and a 10 ppm H₂S test environment is obtained. The entire process is conducted within a fume hood, with temperature of 24 ± 1 °C and a relative humidity of 60%RH.

3. Results

3.1. FEA Simulations of SAW Devices

As illustrated in Figure 2b, the simulated S₂₁ signal from the SAW device exhibits an optimal work frequency of 243.1 MHz, accompanied by an insertion loss of −17.52 dB. The displacement field distribution at the work frequency (Figure 2a) confirms the Rayleigh wave propagation pattern, which is characterized by elliptical particle motions that are primarily concentrated near the device surface. This indicates that the main displacement of the surface particles occurs in the z-axis and the device is sensitive to surface perturbations, which facilitates the achievement of highly sensitive detection of small gas molecules. The insertion loss profile of the SAW device is illustrated in Figure 2c. The center frequency was found to be 241.0 MHz, and the insertion loss was −16.3 dB. The minor discrepancies observed between the simulated and experimental outcomes can be primarily attributed to the inaccuracies in the thickness of the deposited electrodes during the preparation stage and the simplification of the finite element model. The insertion loss of the device was then compared before and after deposition of Bi₂S₃ (Figure 2c). After Bi₂S₃ deposition, the insertion loss changed from −16.3 dB to −24.7 dB (~0.84 dB), and the center frequency decreases from 241.0 MHz to 240.6 MHz, which is mainly due to the increase in the surface mass loading of the device.

3.2. Characterization of Bi₂S₃ Nanoribbons

Figure 3a shows the X-ray diffraction pattern of Bi₂S₃ nanoribbons, where the peaks with 2θ of 15.80°, 17.58°, 22.39°, 24.93°, 28.60°, 35.58°, 39.89°, 43. 12°, 47.33°, 48.27°, 62.68° and 66.90° correspond to the peaks at (020), (120), (130), (211), (240), (141), (250), (530), (470) and (470) of Bi₂S₃, respectively. Furthermore, the XRD patterns of the samples exhibited a strong resemblance to the characteristic peaks of the standard PDF#17-0320 [22], and there was an absence of any other impurity peaks in the results. This finding suggests that the prepared Bi₂S₃ nanoribbons are not only highly crystalline but also extremely pure. As demonstrated in Figure 3b, Bi₂S₃ nanoribbons exhibit pronounced absorption in the visible light spectrum, ranging from 300 to 700 nm. Therefore, these nanoribbons are anticipated to facilitate the development of highly sensitive and light-activated gas sensors within the visible light band.

To further investigate the chemical valence states and corresponding chemical compositions of the Bi₂S₃ nanoribbon, the samples were analyzed by XPS. The characteristic peaks of Bi 5d, S 2s, C 1s, Bi 4d, O 1s and Bi 4p can be clearly observed from the full XPS spectrum of the Bi₂S₃ sample in Figure 3c. In addition, the peaks of Bi 4f overlapped to some extent with the peak positions of S 2p, which was also confirmed by the high-resolution Bi 4f and S 2p XPS spectra (Figure 3d). In particular, the S 2p spectrum could be divided into two characteristic peaks of S 2p_1/2 and S 2p_3/2 with positions at 161.6 eV and 160.5 eV, respectively, which are attributed to the S element in Bi₂S₃ materials [23]. Meanwhile, the two characteristic peaks in the Bi 4f profile with positions at 163.0 eV and 157.8 eV are attributed to Bi 4f_5/2 and Bi 4f_7/2, respectively, suggesting that the +3 valent Bi in the Bi₂S₃ nanoribbons dominates the synthesized material.

SEM images of Bi₂S₃ nanoribbons are shown in Figure 4a–c. It can be clearly observed that the prepared Bi₂S₃ exhibits a typical 1D nanoribbon structure. The 1D Bi₂S₃ nanoribbons possess a large specific surface area, and the interlocking Bi₂S₃ nanoribbons form a reticulated film. These features enable a large number of adsorption sites for gas molecules, which is favorable for the adsorption of gas molecules to increase the gas-sensitive response. The tight binding of their overlapping is also conducive to the effective transfer of carriers between the materials, which is conducive to the rapid adsorption and desorption of gases. As illustrated in Figure 4h, the EDS mapping images of the Cu(NO₃)₂-treated Bi₂S₃ film demonstrates that Bi and S are distributed evenly, with a small amount of Cu resulting from the Cu(NO₃)₂ ligand substitution.

The HR-TEM images of Bi₂S₃ nanoribbons are displayed in Figure 4d,e. The low-magnification HR-TEM image (Figure 4d) confirms that the 1D nanostructure of Bi₂S₃ nanoribbons is primarily attributable to the highly anisotropic orthorhombic phase of Bi₂S₃, which facilitates preferential growth along the nanoribbon axis. The high magnification HR-TEM image (Figure 4e) demonstrates the lattice striations of Bi₂S₃. The inset illustrates the high-quality, ordered lattice striations of Bi₂S₃, with the lattice spacing of 0.199 nm corresponding to the (002) crystallographic plane of Bi₂S₃. Furthermore, the selected area electron diffraction (SAED) (Figure 4f) demonstrates the highly crystalline nature of the Bi₂S₃, displaying clear diffraction points for the (020), (301) and (002) crystal planes. These results are in agreement with the XRD test results, suggesting that high-quality 1D Bi₂S₃ nanoribbons have been successfully prepared.

3.3. Sensing Performance of SAW H₂S Sensors

It is noteworthy that the surface of Bi₂S₃ nanoribbons synthesized via solvothermal method is covered with abundant long-chain surfactants (PVP and OLA), which play a crucial role in the formation of the nanoribbons and result in a well-dispersed solution with good processability [30]. Conversely, the organic long chains of OLA and PVP present on the surface of Bi₂S₃ nanoribbons have been shown to impede the adsorption of gas molecules, thereby hindering carrier transport within the material and resulting in suboptimal gas-sensitive properties [11]. The exchange of long-chain organic ligands with other short-chain ligands, such as inorganic salts, is a facile process. As demonstrated by the FT-IR measurements of Bi₂S₃ nanoribbon films treated with inorganic salts solution (see Figure 4g), the aliphatic C-H stretching vibrational peaks at 2856~2921 cm⁻¹ and 1747 cm⁻¹ in the untreated Bi₂S₃ nanoribbon films correspond to the organic long chains in OLA and PVP, respectively [29]. In contrast, these characteristic peaks were significantly attenuated after NaNO₂, Mg(NO₃)₂, Pb(NO₃)₂ and Cu(NO₃)₂ ligand substitution treatments, which significantly indicated that most of the long-chain ligands were successfully removed by ligand exchange treatments with the above inorganic salts. Removal of insulating organic ligands covering the Bi₂S₃ surface enhances target gas adsorption and rapid charge transfer.

In order to compare the improvement of sensor performance by ligand substitution of different inorganic salt solutions, Bi₂S₃ nanoribbon H₂S sensors treated with NaNO₂, Zn(NO₃)₂, Pb(NO₃)₂ and Cu(NO₃)₂ solutions were prepared, respectively. The responses of all sensors were then subjected to testing at varying concentrations of H₂S, ranging from 2 ppm to 50 ppm. The real-time response curves of the sensors without ligand substitution and after NaNO₂, Mg(NO₃)₂ and Pb(NO₃)₂ ligand substitution for different concentrations of H₂S are shown in Figure 5a–d.

The results show that the operating frequency of all sensors decreases after gas adsorption, which is mainly due to the change in mass loading caused by the gas adsorption process. In addition, all treated sensors showed good response recovery curves at both low and high H₂S gas concentrations, which can be attributed to the excellent performance of the Bi₂S₃ nanoribbon material. The absence of ligand exchange of the Bi₂S₃ nanoribbon SAW sensor resulted in a frequency shift of −5.8 kHz for 50 ppm H₂S gas and a response of only about −1 kHz for 2 ppm H₂S. This is mainly due to the use of organic solvents during the synthesis process, resulting in the attachment of long-chain groups to the Bi₂S₃ nanoribbons, which hinders the adsorption of the target gases and leads to a weakened response. As a result, the responses of the ligand exchange-treated sensors were all higher than those of the untreated Bi₂S₃ sensor. This phenomenon can be attributed primarily to the ligand exchange which induced the sensitive films to be able to contact the H₂S gas molecules more efficiently, which increased the adsorption of H₂S to improve the response of the sensors. The responses of the SAW sensors were −6.26 kHz, −6.323 kHz, −11.298 kHz and −13.548 kHz for 50 ppm H₂S gas, based on NaNO₂, Zn(NO₃)₂, Pb(NO₃)₂ and Cu(NO₃)₂ treatments, respectively. The detailed comparison of the responses at other concentrations is shown in Table S1.

The real-time curve of the sensor response to different concentrations of H₂S gas after Cu(NO₃)₂ ligand exchange treatment is shown in Figure 6a. The response of the sensor with different ligand exchange treatments compared with the untreated Bi₂S₃ nanoribbon sensor for different H₂S gas concentrations is shown in Figure 6b. For different concentrations of H₂S, the sensor exhibited an approximately linear response under 10 ppm. And with the increase of H₂S concentration, the response of the sensors gradually tends to saturation, which may be due to the limited number of gas adsorption sites in the sensitive films leading to H₂S adsorption saturation. Among them, as demonstrated in Figure 6b, the Cu(NO₃)₂-treated sensor exhibits the maximum response. This phenomenon may be attributed to the Cu ions exhibiting a preference for the adsorption of H₂S gas, which has been confirmed in previous reports [31]. Therefore, the subsequent study was conducted on the basis of the Cu(NO₃)₂ solution treatment.

The response of the Cu(NO₃)₂-treated sensor to 10 ppm H₂S gas was tested under 20 μW/cm² light (395 nm, 465 nm, 520 nm, 590 nm and 625 nm) as well as in a dark environment, respectively. Figure 7a presents the real-time response curves of the Cu(NO₃)₂-treated SAW sensor to 10 ppm H₂S gas at different wavelength light activation. Figure 7b shows the corresponding line plots of response comparison at different wavelength light gains. The results show that the response of the sensor to H₂S gas varies with light wavelength, and the light-activated responses are all enhanced compared to the dark environment. Specifically, the response value of the sensor in the dark under 10 ppm H₂S gas was −6.1 kHz, while the frequency shift of the sensor operating at 590 nm light-activated was improved to −11.5 kHz. This represents a substantial enhancement of 188.5% in performance compared to the dark condition.

The response/recovery time of a gas sensor is also an indispensable indicator for evaluating its performance. The response time of a gas sensor is defined as the time required for the response to reach 90% of the maximum value when the sensor is exposed to the target gas environment; the recovery time is defined as the time required for the response to recover from the maximum value to 90% of the initial state [32]. The response/recovery times of the Bi₂S₃ nanoribbons SAW sensor for 10 ppm H₂S gas at 590 nm light-activated are 27 s/21 s, as shown in Figure 7c. In a similar manner, the response/recovery times of the SAW sensor with different wavelengths light-activated for 10 ppm H₂S gas were analyzed (see Figure 7d). It was found that the response/recovery speeds of the sensor with different wavelengths light-activated were improved compared with those in the dark environment. In particular, the response/recovery becomes faster and faster as the wavelength of the activated light increases. This enhancement can be attributed chiefly to the variation in the number of photogenerated carriers excited by different wavelengths of light at 20 μW/cm². The increase in response speed is due to light activation changing the adsorption kinetics on the surface of the sensitive material. This process leads to a decrease in the gas adsorption reaction potential and an acceleration of target gas adsorption and desorption.

The response of the SAW gas sensor with 590 nm light-activated to H₂S gas (0.5–50 ppm) is shown in Figure 8a. When exposed to H₂S gas, the center frequency of the sensor decreases to a negative frequency shift. When the sensor is exposed to air again, the center frequency of the sensor gradually returns to the initial frequency value. It is clear that the frequency shifts of the sensor gradually increase as the target H₂S concentration increases. The sensor’s response to 50 ppm H₂S can reach a maximum of −18.2 kHz. The sensor displays a response of −2.8 kHz even when the H₂S concentration is reduced to 0.5 ppm. This demonstrates the excellent performance of the sensor, which can reach sub-ppb levels.

A comparison of the responses of the sensor with light-activated and dark shows that the response is enhanced by 134% (50 ppm H₂S), as shown in Figure 8b. Furthermore, a 191% improvement in response was observed at a concentration of 10 ppm, with a 163% improvement even at a low concentration of 2 ppm. The enhancement of the light-activated response is particularly pronounced for low concentrations of H₂S below 10 ppm, which facilitates the detection of low concentrations of H₂S.

The relationship between H₂S concentration (0.5~50 ppm) and frequency shifts of the sensor is shown in Figure 8c. The frequency response of the sensor in the low concentration range of H₂S from 0.5 to 8 ppm shows good linearity with R² of 0.981. The good linearity of the sensor facilitates the commercial application and development of the sensor. As the H₂S concentration increased (greater than 10 ppm), the frequency change in the sensor leveled off, which was limited by the number of adsorption sites of sensitive area leading to saturation of H₂S adsorption. The theoretical detection limit (LOD) of the sensor can be calculated based on the signal-to-noise ratio with the following formula [33]:

$$LOD\left(ppm\right)=3{\displaystyle \frac{SD}{S}}$$

(1)

where the standard deviation ($SD$) of the 300 data points of the baseline is 0.0344 and $S$ is the slope of the linear fit line of 1.08306 in Figure 8c. According to Equation (1), the $LOD$ is calculated to be about 95.2 ppb. A comparative evaluation of H₂S sensing capabilities between the SAW gas sensor developed in this study and previously reported SAW-based devices is presented in

Table 1. Comparison of sensitivity and response time of SAW H₂S gas sensor.

Sensitive Materials	Operating Frequency (MHz)	Concentration (ppm)	|Δf| (kHz)	Response Time (s)	Ref.
Fe2O3/SiO2	200	1	0.8	<200	[34]
CuO-Al2O3	200	1	15	>1000	[35]
CuO	200	2	9	>3000	[36]
CuO/TiO2	200	10	34	531	[37]
CuO@V2C	163	20	39	54	[38]
CuO	200	4	30	>200	[39]
Bi2S3	241	10	11.5	27	This work

. The results demonstrate that the Bi₂S₃ nanoribbon-functionalized sensor exhibits superior performance characteristics, particularly with respect to enhanced frequency response and accelerated response/recovery time.

The repeatability of the sensor was evaluated by repeated exposure to 10 ppm H₂S, as shown in Figure 8d. The four frequency shifts of the sensor are almost constant with less than 2% variation in the response value, suggesting a good repeatability of the sensor. Figure 8e reveals the sensor selectivity by comparing the response of various gases commonly, including H₂S, NO₂, NH₃, H₂ methanol, ethanol and acetone gases. The response of the above interfering gases is less than 10% of the frequency response of the 10 ppm H₂S gas, indicating that the SAW sensor based on Bi₂S₃ nanoribbons has a good selectivity for H₂S gas.

Subsequently, the response of the Bi₂S₃-based SAW sensor to 10 ppm H₂S was tested under varying relative humidity conditions (30~80%). As illustrated in Figure S3, the frequency response (Δf) of the SAW sensor increases with the rise in RH values (30~80%). This phenomenon can be attributed to the significant adsorption of H₂S and H₂O molecules on the sensitive area under high relative humidity, which leads to an increase in the mass loading of the SAW sensor and amplifies the frequency shifts [35,38].

3.4. Sensing Mechanism of SAW H₂S Sensors

The gas-sensing mechanism can be elucidated in terms of the chemical reaction that occurs between the sensitive material and H₂S. In air, oxygen molecules are adsorbed on the surface of the sensitive material and extract electrons from the conduction band of Bi₂S₃ films, formatting of adsorbed oxygen ions $O_{2\left(ads\right)}^{-}$, $O_{\left(ads\right)}^{-}$ and $O_{\left(ads\right)}^{2-}$ on the surface of the Bi₂S₃ nanoribbon [40]. The specific reaction process is shown in Figure 9a and Equations (2)–(5). Electron abstraction from the surface of the Bi₂S₃ nanoribbons leads to the formation of an electron depletion layer in the n-type Bi₂S₃ nanoribbons.

$$O_{2\left(gas\right)}\to O_{2\left(ads\right)}$$

(2)

$$O_{2\left(ads\right)}+e^{-}\to O_{2\left(ads\right)}^{-}$$

(3)

$$O_{2\left(ads\right)}^{-}+e^{-}\to {2O}_{\left(ads\right)}^{-}$$

(4)

$$O_{\left(ads\right)}^{-}+e^{-}\to O_{\left(ads\right)}^{2-}$$

(5)

$${2H}_{2}S+{3O}_{2\left(ads\right)}^{-}\to {2SO}_2+{2H}_{2}O+{3e}^{-}$$

(6)

$$H_{2}S+{3O}_{\left(ads\right)}^{-}\to {SO}_2+H_{2}O+{3e}^{-}$$

(7)

$$H_{2}S+{3O}_{\left(ads\right)}^{2-}\to {SO}_2+H_{2}O+{6e}^{-}$$

(8)

Upon exposure to H₂S gas, the H₂S molecules react with the adsorbed oxygen on the surface of the Bi₂S₃ nanoribbons, resulting in the release of electrons back to the Bi₂S₃ surface (Figure 9b). The process can be explained in detail by Equations (6)–(8), and the schematic is shown in Figure 9b. This process leads to changes in the mass and conductance of the sensitive film on the surface of the SAW sensor.

In the case of SAW gas sensors, the gas molecules react with the sensitive material on the surface of the sensor, thereby affecting the propagation of the SAW. The frequency shift of the SAW sensor is attributed to a combination of mass loading effect, acoustic and electrical effects, and viscoelastic effect, as follows [9,41]:

$${\displaystyle \frac{\Delta f}{f_0}}\cong {\displaystyle \frac{\Delta v}{v_0}}=-C_m{f}_0\Delta \left({\rho }_s\right)+4C_e{f}_0\Delta \left(h{G}^{\prime }\right)-{\displaystyle \frac{K^2}{2}}\Delta \left({\displaystyle \frac{1}{1+{\left(\frac{v_0{c}_s}{{\sigma }_s}\right)}^2}}\right)$$

(9)

where $\Delta f$ signifies the frequency shift, $f_0$ denotes the center frequency, $\Delta v$ and $v_0$ represent the change in wave velocity and the initial wave velocity, $C_m$ and $C_e$ are the coefficients of mass sensitivity and elasticity of the material. The change in mass per unit area is denoted by ${\rho }_s$, the thickness of the film is represented by $h$, and the shear modulus is indicated by $G^{\prime }$. The electromechanical coupling coefficient is denoted by $K^2$, the capacitance per unit length of the substrate is denoted by $c_s$, and the conductivity of the film is denoted by ${\sigma }_s$.

According to Equation (9), an increase in the conductivity and an increase in the mass of the sensitive film leads to a decrease in the center frequency of the SAW (negative frequency shift). In order to distinguish between the respective effects of mass loading effect and acoustic electrical effects on frequency shift, a 50 nm Au film was deposited on the sensitive area of the SAW delay line device. The highly conductive metal films eliminate the frequency shift of the SAW device due to acoustoelectric effects. Additionally, the mass loading effect remains unaffected by the Au film. As illustrated in Figure S4, the responses of the sensors with and without Au film to 10 ppm H₂S were −7.81 kHz and −11.57 kHz, respectively. This finding indicates that the response exhibited by Bi₂S₃-based SAW sensors to H₂S is the result of a combination of mass loading and acoustic electric effects, with the mass loading effect contributing 67.5% of the total response.

4. Conclusions

In summary, we have prepared a fast and highly sensitive light-activated SAW H₂S sensor based on 1D Bi₂S₃ nanoribbons. The optimal performance of the Cu(NO₃)₂-treated sensor was determined by investigating ligand exchange with different inorganic salts. The Cu(NO₃)₂-treated sensor exhibited a response value of −6.08 kHz for 10 ppm H₂S, which is 203% of the untreated sensor. Furthermore, the performance of the SAW sensor was evaluated under light activation with different wavelengths. When exposed to 10 ppm H₂S under yellow light activation (590 nm), the sensor demonstrated a frequency shift of −11.57 kHz, which corresponded to 191% of the response in the dark. Furthermore, when the H₂S concentration was reduced to 500 ppb, the sensor exhibited a response of −2.8 kHz with a theoretical detection limit of 31.8 ppb. Therefore, light-activated SAW H₂S sensors based Bi₂S₃ nanoribbons present a promising avenue for enhancing the overall performance of gas sensors and the application of nanomaterials in the domain of gas sensing technology.