The Sensing Selectivity of Gas Sensors Based on Different Sn-Doped Indium Oxide Films

Abstract

The gas-sensitive performance and selectivity of gas sensors via different Sn-doped indium oxide (In₂O₃) films have been investigated. The response characteristics were significantly enhanced to methanol (CH₄O), ethanol (C₂H₆O), and acetone (C₃H₆O) with the increase in Sn content, while the response time and the recovery time became shorter. The sensor exhibited the strongest response to ethanol, followed by acetone and then methanol with all the ratios of In₂O₃ (90%, 85%, and 80%) and SnO₂ (10%, 15%, and 20%). The mechanism of Sn doping on the gas sensing selectivity was calculated using the density functional theory (DFT) method, which perfectly explained the experimental results. The sensors demonstrated high selectivity towards ethanol, even in the presence of interfering gases. In addition, the sensors showed effective detection of the target gas with 10 ppb and demonstrated good repeatability. This work systematically analyzed the priority selectivity of In₂O₃-based gas sensors, providing a new path for gas detection in multi-interference and complex environments.

1. Introduction

Methanol, ethanol, and acetone are the most commonly used organic solvents in human production and daily life, which have demonstrated a wide range of potential applications in various areas such as the food industry, biotechnology, medical sterilizations, chemical engineering, and health protection [1]. However, methanol is toxic and volatile, and can be oxidized to formaldehyde and formic acid through inhalation or skin contact, posing a serious threat to the human nervous and hematological systems [2]. Ethanol is volatile and flammable, and individuals exposed to high ethanol concentrations for prolonged periods of time may encounter a range of health issues such as respiratory disturbances, headaches, drowsiness, eye irritation, and liver damage [3]. Acetone is also volatile and flammable, and penetrates the human body via multiple routes, causing damage to lung, liver, and kidney tissues [4].

Currently, electrochemical sensors, contact combustion sensors, infrared sensors, and semiconductor gas sensors are the main gas sensors in the field of gas detection [5,6,7,8]. In semiconductor sensors, the metal oxide semiconductor (MOS) has emerged as the most widely used gas-sensitive material due to its exceptional performance characteristics, such as high response, rapid response and recovery speeds, excellent long-term stability, low economic cost, and ease of system integration [9]. However, this type of sensor still faces several challenges, especially in terms of poor gas selectivity and limited anti-interference ability. To address these issues, researchers have been exploring many methods to improve sensor performance, such as doping [10,11,12], surface modification with noble metals [13,14,15], constructing heterojunction structures [16,17,18], and designing nanoscale architectures [19,20,21].

As an n-type semiconductor material, indium oxide (In₂O₃) exhibits great potential for gas sensing applications due to its wide bandgap (ranging from 3.5 to 4.3 eV) [22], high electrical conductivity, and ability to detect multiple gases [23]. Our previous work prepared many kinds of gas sensors based on In₂O₃ materials with different nanostructures, such as nanowires and nano-pyramids, by magnetron sputtering [24,25]. Other researchers have also created lots of structures in this area, such as Co₃O₄/In₂O₃ heterojunctions [26] and 3D flower-like In₂O₃ nanosheets [27].

In this study, a large number of gas sensors with different Sn-doped In₂O₃ films were prepared via magnetron sputtering. Mixture gas tests were conducted; the three typical gases of methanol, ethanol, and acetone were selected as test targets. The prepared Sn-doped In₂O₃ sensors exhibited extremely high selectivity to ethanol, followed by acetone and methanol. The density functional theory (DFT) method was applied to reveal the mechanism of Sn doping on the gas sensing performance of crystal structures. Compared with the existing studies, this paper systematically explores for the first time the regulatory effect of different Sn doping ratios on the gas-sensitive selectivity of In₂O₃ thin films and combines DFT calculations to reveal the electronic mechanism of Sn-O bond adsorption of methanol, ethanol, and acetone. This work provides a new strategy for optimizing the selectivity of MOS sensors through precise doping.

2. Experimental Sections

2.1. Preparation of Gas Sensors

Magnetron sputtering was employed to fabricate gas-sensitive materials onto the surfaces of ceramic tubes. Three targets with different In₂O₃/SnO₂ mass ratios (In₂O₃:SnO₂ = 90 wt%:10 wt%, designated as S1; 85 wt%:15 wt%, designated as S2; and 80 wt%:20 wt%, designated as S3) were selected. During the sputtering process, a sputtering power of 300 W (in RF mode), a deposition temperature of 300 °C, a sputtering pressure of 2 Pa, an argon flow rate of 25 sccm, and a sputtering time of 4 min were used to ensure the uniformity and densification of the films. Then, the samples were annealed at 480 °C for 30 s through a rapid annealing furnace. And then, the annealed samples were welded to a heating wire and a six-pin base using solder wire. The sensor was composed of a ceramic tube and a heating wire. The electrodes on the surface of the ceramic tube were platinum electrodes, and the distance between the electrodes was 1 mm. The sensors were heated, and the heating temperature was incrementally raised from 240 °C to 390 °C (the corresponding heating voltage was 3.8 V to 5.2 V), with each 25 °C increase occurring every 10 min. At last, the sensors were heated at 390 °C for 30 min until the resistance of the sensors stabilized. The specific experimental process is shown in Figure 1.

2.2. Sensing Selectivity Test Method

The gas sensitive test was conducted using a WS-30A (Winsen, Zhengzhou), with a measurement voltage of 5 V and a heating voltage of 4.28 V (corresponding to a test temperature of 280 °C). The gas sensitive test was carried out in an 18 L sealed space. The required liquid volume corresponding to the target gas concentration was calculated using Equation (1). When the sensor was used to detect reducing gases, the response could be defined as Equation (2).

$$\mathrm{volume}={\displaystyle \frac{{10}^9\times C_{\mathrm{gas}}\times M\times V}{22.4\times {\rho }_{\mathrm{liquid}}\times \omega }}$$

(1)

$$S=R_{\mathrm{air}}/R_{\mathrm{gas}}$$

(2)

In Equation (1), the volume represents the extracted liquid volume, C_gas is the gas concentration corresponding to the extracted liquid, M is the relative molecular mass of the gas, V is the volume of the sealed space, ρ_liquid is the density of the extracted liquid, and ω is the mass fraction of the extracted liquid. In Equation (2), S is the response value of the gas sensor, R_air is the resistance of the sensor in air, and R_gas is the resistance of the sensor in the target gas.

For the priority selectivity test of mixed gases, two micro-syringes were used to extract the liquid volumes corresponding to the required concentrations of the target gas and interfering gas, respectively. These were then mixed in the evaporation station in sequence. The liquid was evaporated to gas, and a small fan in the sealed space was used to evenly distribute the mixed gas in the sealed space for testing.

2.3. Computational Methods

The Cambridge Sequential Total Energy Package (CASTEP) module in Materials Studio (MS) was utilized for the calculations. This module was based on the pseudopotential plane-wave method of DFT, and Generalized Gradient Approximation (GGA) in the form of Perdew–Burke–Ernzerhof (PBE) was selected as the exchange–correlation functional. Additionally, the DFT-D2 method was introduced for dispersion correction to enhance the calculation accuracy [28,29]. The In₂O₃ material exhibits both a cubic ferromanganese structure and a hexagonal corundum-type structure [30]. In this study, the more stable cubic ferromanganese structure was chosen for our calculations, which possesses a lattice parameter of 10.15 Å and a forbidden band width of 2.45 eV [31]. A three-dimensional schematic of the structure is depicted in Figure 2a.

Considering the characteristics of the crystal growth direction of In₂O₃, the (110) low-index plane was selected for calculations because of its heightened activity during the crystal growth process, as illustrated in Figure 2b. In the calculation setup, the plane-wave truncation energy was set to 380 eV, and the K-point grid in the Brillouin zone was set to 3 × 3 × 3. To ensure the accuracy of the calculation results, the spin state of the electrons and dipole correction were fully considered [32]. Furthermore, a vacuum layer of 20 Å was introduced along the Z-direction to mitigate inter crystalline interactions arising from periodic boundary conditions.

In the doping simulations, substitutional doping was employed. The In₂O₃ structure contains two different types of In atoms at the 8b and 24d positions. A single Sn atom is more likely to replace an In atom at the 8b position, and two Sn atoms tend to replace In atoms at both the 8b and 24d positions, respectively [33]. We adopted the doping order of b-d-b-d in this study. As a result, three Sn atoms were successfully doped into the structure, as shown in Figure 2d. For all the calculations, the tolerances for energy, maximum force, and maximum displacement were set as 10⁻⁵ eV/atom, 0.03 eV/Å, and 0.001 Å, respectively.

3. Results and Discussion

3.1. Morphologies and Elemental Composition for Different Films

As shown in Figure 3, the SEM images show the morphology of different Sn-doped In₂O₃ films with the thickness of 77.7 nm, 61.8 nm, and 46.2 nm, and the insets are the AFM results. The root mean square (RMS) error for S1, S2, and S3 was 0.878 nm, 0.684 nm, and 0.520 nm, respectively, which means the doping of Sn can decrease the surface roughness of films. This phenomenon indicates that the doping of the Sn has a significant inhibitory effect on the grain growth of In₂O₃ [34], and the reduction in grain size transforms the sensor into grain-controlling mode [35].

Table 1. Elemental contents of samples synthesized via sputtering targets with varied composition ratios.

	O (At%)	In (At%)	Sn (At%)
S1	28.41	67.35	4.48
S2	27.18	61.91	10.92
S3	27.33	61.96	10.70

provides detailed content data for each constituent element in three different samples. As the mass ratio of In₂O₃:SnO₂ in targets decreases, the contents of In and O elements in the prepared films exhibit a notable decreasing trend, while the Sn element content increases accordingly. It is noteworthy that the difference in Sn content between S2 and S3 is negligible, suggesting that the doping amount of Sn tends to stabilize below a certain critical ratio.

3.2. Gas Sensing Properties of Sensors

The working temperature of the sensors was set at 280 °C to ensure both high response and a long lifetime. As the Sn content increases, the response to the three target gases is significantly enhanced, and the response and recovery times are significantly reduced. This confirms the positive modulation effect of Sn doping on the gas sensing performance of In₂O₃. In particular, the sensors exhibit the strongest response to ethanol under the same concentration conditions, followed by acetone and then methanol. The response value of S1 to 100 ppm methanol was 2.21, and the response and recovery times were 6/19 s. The response value of S1 to 100 ppm ethanol was 2.76, and the response and recovery times were 5/8 s. The response value of S1 to 100 ppm acetone was 2.71, and the response and recovery times were 6/17 s. The response value of S2 to 100 ppm ethanol was 2.78, and the response and recovery times were 14/31 s. The response value of S2 to 100 ppm ethanol was 3.36, and the response and recovery times were 9/20 s. The response value of S2 to 100 ppm acetone was 3.34, and the response and recovery times were 9/25 s. The response value of S3 to 100 ppm methanol was 2.27, and the response and recovery times were 6/19 s. The response value of S3 to 100 ppm ethanol was 2.72, and the response and recovery times were 6/34 s. The response value of S3 to 100 ppm acetone was 3.03, and the response and recovery times were 3/16 s.

Figure 4 displays the dynamic response–recovery curves of three sensors toward methanol, ethanol, and acetone across a concentration range from 10 ppb to 500 ppm. For 500 ppm ethanol, the response value of the three sensors were measured as 3.61, 4.51, and 4.62, respectively. Notably, the response of S1 to these three target gases was significantly lower compared to those of S2 and S3. Given the similarity in Sn content between S2 and S3, their response performances toward the aforementioned gases exhibit minimal variability. Furthermore, all three sensors demonstrate effective detection of low concentrations of methanol, ethanol, and acetone. The limit of detection (LOD) of the three prepared sensors for methanol, ethanol, and acetone are 10 ppb. Among them, the response values of S2 to 10 ppb of methanol, ethanol, and acetone are 1.18, 1.16, and 1.11 respectively.

Figure 5 presents the response characteristics of the sensors in the case of pairwise gas mixing (the corresponding 2D image is shown in Supplementary Figure S1). In these tests, a fixed single gas concentration of 150 ppm was used, while the other two gases were added at concentrations ranging from 50 ppm to 250 ppm.

Figure 5a reveals an enhancement trend in the response of the three sensors to methanol when acetone or ethanol are added, indicating a synergistic adsorption effect among the mixed gases. As the concentration of the added gas increases, so does the interference with the detection results. The interference effect of ethanol on methanol is stronger than that of acetone on methanol. When the gas of acetone was mixed to 150 ppm, the changes in response value in methanol were 17.8%, 27.8%, and 30.7% for S1, S2, and S3, respectively. In contrast, when the same concentration of ethanol was mixed, the changes were 42.3%, 37.6%, and 41.1%, respectively. So, the S1 sensor exhibits the best selectivity for methanol when acetone is mixed in, and the S2 sensor shows the best selectivity for methanol when ethanol is mixed in.

Figure 5b demonstrates the response characteristics with mixing methanol or acetone into ethanol. No matter how the concentration of methanol changes, the anti-interference performance on ethanol’s detection is best. The response value changes of the three sensors to ethanol ranged from 1% to 1.4% at 150 ppm methanol. In contrast, when mixed with 150 ppm acetone, the magnitude of response value changes to ethanol increased significantly (3.9%, 8.6%, and 8.5%, respectively), suggesting that the interference effect of acetone on ethanol is stronger than that of methanol on ethanol. Among the three samples, all sensors show minimal interference effects in ethanol detection within a methanol-containing environment. And in an acetone-containing environment, S1 demonstrates better selectivity for ethanol.

Figure 5c shows the response characteristics adding methanol or ethanol into acetone. When 150 ppm methanol was added, the response value changes of three sensors were 7.7%, 3.8%, and 1.4%, respectively. However, mixing ethanol at the same concentration (150 ppm) led to significant changes in the three sensors (30.3%, 22.9%, and 17.4%, respectively). In the detection of acetone, methanol exerts less interference effect than ethanol. Regardless of the gas mixed in, S3 exhibits superior selectivity toward acetone. This indicates that the increasing Sn content enhances the selectivity of In₂O₃ for acetone detection. Although the Sn content of samples S2 and S3 is similar, with S2 having 0.22 at% more Sn than S3, the results of the concentration test showed no significant difference in their responses to methanol and ethanol. However, there was a marked difference in their responses to acetone, with S3 exhibiting a stronger response than S2. In the selectivity test, S3 demonstrated superior selectivity for acetone compared to S2. These results indicate that there is a threshold for the doping content of Sn in In₂O₃. When the doping content of Sn reaches this threshold, the selectivity for acetone is the best.

As shown in Figure 6a−c, five consecutive cycles of the repeatability test were conducted for the three semiconductor gas sensors at 100 ppm ethanol. The test results indicate that the response value of each sensor maintains a high degree of consistency over the repeated measurements, with the maximum deviation not exceeding 1.3%. This fully verifies the stability and repeatability of Sn-doped In₂O₃ sensors in gas detection.

Humidity, as one of the pivotal factors influencing the performance of gas sensors, was manipulated within a confined environment to assess the sensors’ gas sensing response characteristics under four distinct humidity conditions (45% RH, 60% RH, 75% RH, and 90% RH), as depicted in Figure 6d–f. The experimental results indicate that as humidity increased from 45% RH to 60% RH, the response of S1, S2, and S3 towards methanol, ethanol, and acetone decreased significantly. Specifically, the response value of the three sensors to 200 ppm methanol dropped to 79.93%, 74.97%, and 71.64% of their original value, respectively. For testing with 200 ppm ethanol, the response reductions were to 77.96%, 69.37%, and 69.34%, respectively. And the declines changed to 78.38%, 67.51%, and 69.83% under 200 ppm acetone, respectively. Notably, S1 exhibited superior humidity resistance compared to S2 and S3.

As humidity further increased, the decreasing trend in the sensors’ response to the target gases gradually slowed down. The decline in response with increasing humidity is attributed to the dissociation of water molecules at oxygen vacancies (Vo) on the oxide surface [36], leading to the occupation of active sites on the sensor’s surface. Both In₂O₃ and SnO₂ belong to n-type oxides, which are inherently rich in Vo on their surfaces. Additionally, the doping of Sn further enhances the Vo concentration within In₂O₃, shown as Equation (3). The humidity stability advantage of S1 could be attributed to its lower Sn doping amount, resulting in the Vo concentration being lower than that of S2/S3. Since Vo is the main site for water molecule adsorption, S1 demonstrates stronger humidity resistance compared to S2 and S3. Additionally, as humidity continued to rise, the adsorption of water molecules by the sensors gradually approached saturation, resulting in a flattened trend in the changes in the sensors’ response value.

$$I{n}_2{O}_3+Sn\to I{n}_{2-x}S{n}_x{O}_{3-2x}+x{V}_2+x{V}_o$$

(3)

3.3. Gas Adsorption on Sn-Doped In₂O₃ (110) Surface

As shown in Figure 7, numerous adsorption sites with varying physical environments exist on the Sn-doped In₂O₃ (110) surface, which is a characteristic that provides Sn-doped In₂O₃ with significant adsorption capabilities. In view of the variability in chemical bonding between molecules of different gases, their adsorption behaviors on the Sn-doped In₂O₃ (110) surface exhibit diversity. To initially investigate the adsorption characteristics of methanol, ethanol, and acetone at different positions within Sn-doped In₂O₃, a simplified calculation model was employed first.

The calculation results reveal that, for methanol and ethanol, the optimal adsorption position in In₂O₃ is located at the In-O bond. After doping with Sn atoms, the best adsorption position shifts to the Sn-O bond. Methanol and ethanol primarily interact with Sn-doped In₂O₃ through their H-O bonds. Similarly, the adsorption behavior of acetone resembles that of methanol and ethanol. In In₂O₃, the optimal adsorption position for acetone is also at the In-O bond, which changes to the Sn-O bond following Sn doping. Acetone interacts primarily with Sn-doped In₂O₃ through its C=O bond during adsorption. This indicates that the Sn-O and In-O bonds in Sn-doped In₂O₃ are the primary reactive sites for alcohols and ketone gases. The optimal adsorption configurations are illustrated in Figure 8.

$$E_{\mathrm{ads}}=E_{Sn-dopedI{n}_2{O}_3+\mathrm{gas}}-E_{Sn-dopedI{n}_2{O}_3}-E_{\mathrm{gas}}$$

(4)

In Equation (4), E_{Sn-Doped In2O3+gas} represents the total energy of Sn-doped In₂O₃ and the target gas after adsorption, while E_{Sn-Doped In2O3} and E_gas denote the energies of Sn-doped In₂O₃ and the target gas, respectively, in their unadsorbed states.

As shown in Figure 9a, the adsorption energies of the three gases on the In-O bond are all negative, confirming that the adsorption process between the gases and the sensitive material is exothermic and spontaneous. Furthermore, these adsorption energies on the In-O bond are all lower than −0.4 eV [37], indicating the occurrence of chemisorption between the three gases and In₂O₃. The lower the adsorption energy, the stronger the binding between gas molecules and the sensitive material, and the more pronounced the electron transfer, leading to a higher response value. Among the three gases, ethanol exhibits the smallest adsorption energy, suggesting that In₂O₃ has a better response to ethanol compared to methanol and acetone. Additionally, the adsorption energies of the three gases on the Sn-O bond are lower than those on the In-O bond, further indicating that doping In₂O₃ with Sn significantly enhances its response to methanol, ethanol, and acetone. Notably, the most remarkable change in adsorption energy was observed for acetone, highlighting the stronger effect of Sn doping on ketone gases.

Figure 9b illustrates the Charge Density Difference (CDD) of ethanol and Sn-doped In₂O₃. In this plot, the yellow region represents the electron depletion region, which is concentrated near the H-O bond at the contact interface between ethanol and Sn-doped In₂O₃ (110). Conversely, the blue region indicates the electron enrichment region, which is mainly distributed on the Sn-doped In₂O₃ (110) surface. This phenomenon suggests that, as a reducing gas, ethanol loses electrons and acts as an electron donor during adsorption with Sn-doped In₂O₃, an n-type semiconductor sensitive material. Consequently, the resistance of Sn-doped In₂O₃, acting as an electron acceptor, is decreased. Chemisorption between ethanol and Sn-doped In₂O₃ is achieved through significant charge transfer.

Figure 9c displays the Density of States (DOS) of Sn-doped In₂O₃ before and after gas adsorption. After adsorption, the DOS shifts to lower energy, and the DOS at the fermi energy level increases, indicating an enhancement in electrical conductivity. This occurs because ethanol, acting as a reducing agent, donates electrons to Sn-doped In₂O₃ during adsorption. Figure 9d shows the Partial Density of States (PDOS) of Sn-doped In₂O₃ after ethanol adsorption. In the energy range from −9 eV to −3 eV, there are obvious multi-orbital overlaps among the 2p orbitals of the O atoms and the 5p orbitals of the Sn atoms of Sn-doped In₂O₃ and the 1s orbitals of the H atoms and the 2p orbitals of the O atoms of ethanol. This indicates the occurrence of hybridization and strong interactions between these orbitals. In particular, the electron cloud overlap of the O atom of Sn-doped In₂O₃ has the largest area and contributes the most significantly to the adsorption process, suggesting that the main active site for adsorption in Sn-doped In₂O₃ is located at the O atom.

3.4. Gas Sensing Mechanism

For n-type MOS sensors, the majority carriers are electrons, as depicted in Figure 10b. When the sensitive material on the surface of the gas sensor is exposed to air, oxygen molecules capture electrons from the material, thereby being converted into adsorbed oxygen. This process results in the formation of an Electron Depletion Layer (EDL) on the surface of the sensitive material [38]. The formation of an EDL significantly reduces the electron concentration in the sensitive material, which subsequently decreases its conductivity and increases its resistance. When the sensor is exposed to an atmosphere containing reducing gases, such as ethanol, these gas molecules undergo a redox reaction with the adsorbed oxygen, releasing the previously captured electrons back into the conduction band of the material. This reaction leads to an increase in electron concentration, an enhancement in conductivity, and a corresponding decrease in resistance [39]. A MOS sensor relies on the change in resistance to effectively detect target gases. The reactions between adsorbed oxygen and gases such as methanol, ethanol, and acetone on the surface of In₂O₃ are described by reaction Equations (5)–(12) [40,41,42]:

$$O_{2(\mathrm{gas})}\to O_{2(\mathrm{ads})}$$

(5)

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

(6)

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

(7)

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

(8)

$$C{H}_{3}OH+3O^{-}\to C{O}_2+2H_{2}O+3e^{-}$$

(9)

$$C_2{H}_{5}OH+O^{-}\to C{H}_{3}CHO+H_{2}O+e^{-}$$

(10)

$$C{H}_{3}CHO+5O^{-}\to 2C{O}_2+2H_{2}O+5e^{-}$$

(11)

$$C{H}_{3}COC{H}_3+8O^{-}\to 3C{O}_2+3H_{2}O+8e^{-}$$

(12)

Both In₂O₃ and SnO₂ are n-type semiconductors, as illustrated in Figure 10a. The work function, electron affinity, and band gap of In₂O₃ are 4.3 eV, 3.7 eV, and 3.6 eV, respectively. The corresponding parameters for SnO₂ are 4.9 eV, 3.7 eV, and 3.6 eV, respectively [43]. When In₂O₃ is in contact with SnO₂, a contact interface is formed due to the misalignment of their fermi energy levels (where the fermi energy level of In₂O₃ is higher than that of SnO₂), and electrons continuously flow from In₂O₃ to SnO₂ until an equilibrium state is established, forming an n-n heterojunction structure [44]. During this process, the SnO₂ side accumulates electrons to form an electron-accumulation layer, while the In₂O₃ side, due to electron loss, forms a wider EDL, resulting in an increase in In₂O₃’s resistance. When the sensor with an n-n heterojunction structure (composed of In₂O₃ and SnO₂) is exposed to a reducing gas environment, the reducing gas reacts with adsorbed oxygen, causing the thickness of the EDL to decrease and the resistance of the sensor to decrease dramatically. This significant resistance change greatly enhances the response [45]. In addition, the doping of Sn not only effectively inhibits the growth of In₂O₃ grains, resulting in a larger specific surface area for the material, but also increases the content of Vo in In₂O₃ [46]. The increase in Vo provides more adsorption sites for gas molecules, which significantly improves the gas sensing performance of In₂O₃.

Among the three target gas molecules, ethanol has the smallest adsorption energy at the In-O bond, and the main component of the three sensors is still In₂O₃. Therefore, its response effect to ethanol is the best. The adsorption energy difference of acetone at the Sn-O bond and the In-O bond is the greatest, which indicates that the doping of Sn element is more conducive to the adsorption of acetone. Given that the Sn content of S2 and S3 is significantly higher than that of S1, S1 is more inclined to promote the adsorption of alcohol gases, thereby demonstrating higher anti-interference performance. Meanwhile, S2 and S3, due to their higher Sn content, are more conducive to the adsorption of acetone.

4. Conclusions

In this study, three types of Sn-doped In₂O₃-based sensors with varying indium-tin mass ratios were prepared using magnetron sputtering followed by a rapid annealing process. Systematic gas-sensitive performance tests were conducted for these sensors with methanol, ethanol, and acetone. Additionally, DFT calculations were performed to gain insights into the sensing mechanisms. The experimental results revealed that the Sn-doped In₂O₃ sensors exhibited high selectivity to ethanol. Even when high or low concentrations of methanol or acetone were mixed with ethanol, the strongest selectivity of the sensors was towards ethanol. By adjusting the indium/tin mass ratios in the Sn-doped In₂O₃ composition, we observed a significant enhancement in the gas-sensitive performance of Sn-doped In₂O₃. In particular, the sensors in this study demonstrated excellent performance for ultra-low-concentration gas detection. This enhancement can be attributed to the formation of a heterostructure at the interface of In₂O₃ and SnO₂, which not only broadens the EDL on the surface of In₂O₃ but also introduces additional Vo. Considering that water molecules are prone to adsorb and dissociate at Vo on the oxide surface, we found that S1, with a lower Sn content, showed superior stability under varying humidity conditions compared to S2 and S3. All in all, by adjusting the Sn content in In₂O₃ films, sensors with high selectivity for different gases and stable gas detection capability under varying humidity conditions could be achieved. These findings provide strong technical support for the accurate monitoring of mixed gases in complex space environments.