Highly Sensing and Selective Performance Based on Bi-Doped Porous ZnSnO3 Nanospheres for Detection of n-Butanol

In this study, pure zinc stannate (ZnSnO3) and bismuth (Bi)-doped ZnSnO3 composites (Bi-ZnSnO3) were synthesized via the in situ precipitation method, and their microstructures, morphologies, chemical components, sizes, and specific surface areas were characterized, followed by testing their gas sensing properties. The results revealed that Bi-ZnSnO3 showed superior gas sensing properties to n-butanol gas, with an optimal operating temperature of 300 °C, which was 50 °C lower than that of pure ZnSnO3. At this temperature, moreover, the sensitivity of Bi-ZnSnO3 to n-butanol gas at the concentration of 100 ppm reached as high as 1450.65, which was 35.57 times that (41.01) of ammonia gas, 2.93 times that (495.09) of acetone gas, 6.02 times that (241.05) of methanol gas, 2.54 times that (571.48) of formaldehyde gas, and 2.98 times that (486.58) of ethanol gas. Bi-ZnSnO3 had a highly repeatable performance. The total proportion of oxygen vacancies and chemi-adsorbed oxygen in Bi-ZnSnO3 (4 wt%) was 27.72% to 32.68% higher than that of pure ZnSnO3. Therefore, Bi-ZnSnO3 has considerable potential in detecting n-butanol gas by virtue of its excellent gas-sensing properties.


Introduction
n-Butanol, a typical volatile organic compound, has been widely applied to the production of various paint solvents, extractants, and plastic rubber products [1,2]. n-Butanol, which is flammable, gives off a unique pungent smell and tends to form explosive mixtures with air [3]. High-concentration n-butanol gas is hazardous to the human body, causing symptoms such as dizziness, headache, drowsiness, and dermatitis if exposed at >50 ppm for a long time [4,5]. Therefore, n-butanol gas must be rapidly and effectively monitored for the sake of human health and safety, but expensive and cumbersome monitoring instruments are usually required for traditional n-butanol detection methods, which are also marred by their complex operation, long response time(T(res)), etc. [6,7]. Hence, it is worthwhile to develop highly sensitive, highly selective gas sensors that can effectively monitor n-butanol. A variety of gas sensors have been developed, including semiconductor metal oxide gas sensors, primary battery-type oxygen sensors, and catalytic combustion-type sensors [8].
Among semiconductor metal oxides, the most extensively applied gas-sensitive materials at present, n-type semiconductors are the most studied, such as SnO 2 , TiO 2 , Fe 2 O 3 , In 2 O 3 , LaFeO 3 , WO 3 , and ZnO, and the gas-sensitive devices made of such materials integrate the merits of low cost, strong stability, easy manufacturing, and convenient use [9][10][11]. Nevertheless, single-metal-oxide gas sensors can no longer meet operational needs due to their high operating temperature and low sensitivity [12]. As a typical n-type metal oxide semiconductor, zinc stannate (ZnSnO 3 ), which can be easily prepared and conveniently used at a low cost, has been used to prepare sensor elements to detect all kinds of volatile organic gases [13]. The gas-sensing properties of ZnSnO 3 with different ZnSnO3 was prepared through an in situ precipitation method. First, ethanol aqueous solution (33.3 wt%) was prepared. Next, 4 mmol SnCl4•5H2O and 4 mmol Zn(NO3)2•7H2O were dispersed into 64 mL of an ethanol aqueous solution, followed by magnetic stirring for 30 min to form the uniform suspension marked as solution A. Afterwards, 48 mmol NaOH was added to a beaker containing 48 mL of deionized water and completely dissolved, forming solution B, which was then slowly added into solution A to be subjected to magnetic stirring for 30 min. Then, 120 mmol NaOH was dissolved in 32 mL of ethanol aqueous solution to form a clear solution, which was added dropwise into the abovementioned mixed solution. After stirring for 5 min, Bi(NO3)3•6H2O with different mass fractions was added at the doping concentration of 0 wt%, 2 wt%, 4 wt%, 5 wt%, or 7 wt%, followed by magnetic stirring for 30 min. It was transferred to a 500 mL round-bottom flask for refluxing at 85 °C for 3 h and left to stand for 12 h. The precipitates were filtered, rinsed, and dried. The obtained powder materials were heated in a 450 °C tube furnace at a heating rate of 5 °C/min for 3 h, yielding Bi-ZnSnO3 with different Bi contents (pure ZnSnO3 was acquired at the Bi doping concentration of 0 wt%). The synthesis process of Bi-ZnSnO3 is displayed in Figure 1. Under alkaline conditions, Zn 2+ , Sn 4+ , and Bi were doped to form the Bi-ZnSn(OH)6 precursor (Formula (1)). Then, Bi-ZnSn(OH)6 was calcinated at a high temperature, and crystals were removed to form Bi-ZnSnO3 (Formula (2)).

Preparation of Sensor Device
First, a small amount of the ZnSnO3 and Bi-ZnSnO3 samples were weighed and uniformly ground in a mortar. Second, a moderate amount of absolute ethanol was added into the ground samples to dissolve them into pastes, which were uniformly coated on the outer surface of the gas sensor. Next, the sensor was placed in an 80 °C vacuum-drying oven for 12 h, followed by aging in a 200 °C aging system for 12 h. Finally, the gas-sensing properties of the aging sensor were explored with air-sensing test equipment. The sensitivity of gas sensitive materials was defined as follows: S Ra/Rg (3) where Ra is the resistance (Ω) of the gas-sensing material in air and Rg is the resistance (Ω) of the gas-sensing material in the gas under test. Under alkaline conditions, Zn 2+ , Sn 4+ , and Bi were doped to form the Bi-ZnSn(OH) 6 precursor (Formula (1)). Then, Bi-ZnSn(OH) 6 was calcinated at a high temperature, and crystals were removed to form Bi-ZnSnO 3 (Formula (2)). 6 (1)

Preparation of Sensor Device
First, a small amount of the ZnSnO 3 and Bi-ZnSnO 3 samples were weighed and uniformly ground in a mortar. Second, a moderate amount of absolute ethanol was added into the ground samples to dissolve them into pastes, which were uniformly coated on the outer surface of the gas sensor. Next, the sensor was placed in an 80 • C vacuumdrying oven for 12 h, followed by aging in a 200 • C aging system for 12 h. Finally, the gas-sensing properties of the aging sensor were explored with air-sensing test equipment. The sensitivity of gas sensitive materials was defined as follows: where Ra is the resistance (Ω) of the gas-sensing material in air and Rg is the resistance (Ω) of the gas-sensing material in the gas under test.

XRD
The XRD patterns of ZnSnO 3 and Bi-ZnSnO 3 are exhibited in Figure 2. It was clearly observed that pure ZnSnO 3 had two weak and wide diffraction peaks at 2θ = 27.25 • and 58.59 • , indicating that amorphous ZnSnO 3 formed after the heat treatment [21]. This was The XRD patterns of ZnSnO3 and Bi-ZnSnO3 are exhibited in Figure 2. It was clearly observed that pure ZnSnO3 had two weak and wide diffraction peaks at 2θ = 27.25° and 58.59°, indicating that amorphous ZnSnO3 formed after the heat treatment [21]. This was because the calcination and dehydration of the precursor ZnSn(OH)6 destroyed the original H-O bond and changes the internal lattice arrangement, thereby forming amorphous ZnSnO3 [22]. In comparison with the patterns of pure ZnSnO3, new intense characteristic peaks could be observed in Bi-ZnSnO3 at 2θ = 27.47°, 33.16°, 35.01°, 42.42°, 46.47°, 52.38°, and 55.11°, which corresponded to the crystal faces of (021), (410), (240), (241), (431), (540), and (611), respectively. This result coincided with the standard data card (PDF#17-0320) of Bi, indicating that these newly generated diffraction peaks resulted from Bi doping. With the increase in the doping concentration of Bi, the intensity of diffraction peaks rose substantially, which further proved the successful doping of Bi.

SEM
The SEM images of pure ZnSnO3 and Bi-ZnSnO3 are presented in Figure 3. It can be seen from the SEM image ( Figure 3a) of ZnSnO3 that pure ZnSnO3 had a spherical structure with a coarse appearance and many holes on the surface. From the gaps of broken holes, the pure ZnSnO3 was hollowly structured, with a large pore volume and a diameter of about 900 nm, and each ZnSnO3 block consisted of many tiny, uniformly distributed nanoparticles. As shown in Figure 3b, Bi-doped ZnSnO3 had a similar appearance, namely, it remained a spherical hollow porous structure. The element distribution diagram was mainly used to study the composition and concentration of each component in the test materials. As further verified by Figure 3c-g, four elements, O, Bi, Zn, and Sn, mainly existed in Bi-ZnSnO3, among which Sn, Zn, and O were distributed quite uniformly, though the element distribution does not seem obvious in the figures due to too small doping concentrations of Bi. The mass ratios and the atomic percentages of the four elements are listed in Table 1. It can be intuitively seen that the atomic percentages of O, Bi, Zn, and Sn were 47.54%, 3.43%, 20.44%, and 28.58%, respectively. Evidently, the content of Bi was much lower than that of the other three elements.

SEM
The SEM images of pure ZnSnO 3 and Bi-ZnSnO 3 are presented in Figure 3. It can be seen from the SEM image ( Figure 3a) of ZnSnO 3 that pure ZnSnO 3 had a spherical structure with a coarse appearance and many holes on the surface. From the gaps of broken holes, the pure ZnSnO 3 was hollowly structured, with a large pore volume and a diameter of about 900 nm, and each ZnSnO 3 block consisted of many tiny, uniformly distributed nanoparticles. As shown in Figure 3b, Bi-doped ZnSnO 3 had a similar appearance, namely, it remained a spherical hollow porous structure. The element distribution diagram was mainly used to study the composition and concentration of each component in the test materials. As further verified by Figure 3c-g, four elements, O, Bi, Zn, and Sn, mainly existed in Bi-ZnSnO 3 , among which Sn, Zn, and O were distributed quite uniformly, though the element distribution does not seem obvious in the figures due to too small doping concentrations of Bi. The mass ratios and the atomic percentages of the four elements are listed in Table 1. It can be intuitively seen that the atomic percentages of O, Bi, Zn, and Sn were 47.54%, 3.43%, 20.44%, and 28.58%, respectively. Evidently, the content of Bi was much lower than that of the other three elements.

TEM
The microstructures of pure ZnSnO 3 and Bi-ZnSnO 3 were further explored via TEM. The TEM images of pure ZnSnO 3 are displayed in Figure 4a,b. The spherically structured ZnSnO 3 was 1.92 nm in size and was formed by the aggregation of many particles. In addition, the hollow structure of the sample clearly had many holes on it, in agreement with the SEM characterization results. As observed from the TEM image ( Figure 4c) of Bi-ZnSnO 3 , the sample did change much morphologically, keeping a spherical hollow porous structure with a size of 1.98 nm. The electronic diffraction pattern is illustrated in Figure 4c, presenting regularly arranged diffraction spots. Given this, Bi-ZnSnO 3 was judged to be a single-crystal structure within the range of selection.

TEM
The microstructures of pure ZnSnO3 and Bi-ZnSnO3 were further explored via TEM. The TEM images of pure ZnSnO3 are displayed in Figure 4a,b. The spherically structured ZnSnO3 was 1.92 nm in size and was formed by the aggregation of many particles. In addition, the hollow structure of the sample clearly had many holes on it, in agreement with the SEM characterization results. As observed from the TEM image ( Figure 4c) of Bi-ZnSnO3, the sample did change much morphologically, keeping a spherical hollow porous structure with a size of 1.98 nm. The electronic diffraction pattern is illustrated in Figure 4c, presenting regularly arranged diffraction spots. Given this, Bi-ZnSnO3 was judged to be a single-crystal structure within the range of selection.

BET
Specific surface area is one of the important parameters for measuring gas-sensitive materials. In this study, the specific surface area of prepared materials was determined through the N2 adsorption-desorption method. The N2 adsorption-desorption isothermal diagram of pure ZnSnO3 and Bi-ZnSnO3 is displayed in Figure 5a,b. According to the classification of International Union of Pure and Applied Chemistry (IUPAC), both materials had a type IV isothermal forming an H3 hysteresis loop, which is the characteristic isothermal of mesoporous materials [23,24]. The BET specific surface areas of pure ZnSnO3 and Bi-ZnSnO3 were 10.01 m 2 /g and 13.75 m 2 /g, respectively. The pore volume and the average pore size of pure ZnSnO3 were 0.02 cm 3 /g and 9.26 nm, and those of Bi-ZnSnO3 were 0.03 cm 3 /g and 25.51 nm. The pores might have been generated by the gaps generated in the formation of ZnSnO3 nanospheres. Furthermore, the mesoporous structure of microspheres and the enlargement of their specific surface area facilitated the gas adsorption and further enhanced the gas transport efficiency [25]. Therefore, we hypothesize that the changes in the specific surface area, the pore volume, and the pore size of Bi-ZnSnO3 exerted important influences on the gas sensitivity of this material.

BET
Specific surface area is one of the important parameters for measuring gas-sensitive materials. In this study, the specific surface area of prepared materials was determined through the N 2 adsorption-desorption method. The N 2 adsorption-desorption isothermal diagram of pure ZnSnO 3 and Bi-ZnSnO 3 is displayed in Figure 5a,b. According to the classification of International Union of Pure and Applied Chemistry (IUPAC), both materials had a type IV isothermal forming an H3 hysteresis loop, which is the characteristic isothermal of mesoporous materials [23,24]. The BET specific surface areas of pure ZnSnO 3 and Bi-ZnSnO 3 were 10.01 m 2 /g and 13.75 m 2 /g, respectively. The pore volume and the average pore size of pure ZnSnO 3 were 0.02 cm 3 /g and 9.26 nm, and those of Bi-ZnSnO 3 were 0.03 cm 3 /g and 25.51 nm. The pores might have been generated by the gaps generated in the formation of ZnSnO 3 nanospheres. Furthermore, the mesoporous structure of microspheres and the enlargement of their specific surface area facilitated the gas adsorption and further enhanced the gas transport efficiency [25]. Therefore, we hypothesize that the changes in the specific surface area, the pore volume, and the pore size of Bi-ZnSnO 3 exerted important influences on the gas sensitivity of this material.

XPS
The surface chemical composition and element distribution were analyzed via XPS. The XPS survey spectrum of Bi-ZnSnO3 is displayed in Figure 6a. Characteristic peaks of C, Zn, Sn, O, and Bi existed in the composites, indicating the presence of these five elements in the samples. The high-resolution pattern of Bi4f is exhibited in Figure 6b, where the peaks at the binding energy of 159.08 eV and 164.08 eV corresponded to Bi4f7/2 and

XPS
The surface chemical composition and element distribution were analyzed via XPS. The XPS survey spectrum of Bi-ZnSnO 3 is displayed in Figure 6a. Characteristic peaks of C, Zn, Sn, O, and Bi existed in the composites, indicating the presence of these five elements in the samples. The high-resolution pattern of Bi4f is exhibited in Figure 6b, where the peaks at the binding energy of 159.08 eV and 164.08 eV corresponded to Bi4f 7/2 and Bi4f 5/2 of Bi, respectively. The high-resolution patterns of Zn2p in pure ZnSnO 3 and Bi-ZnSnO 3 are shown in Figure 6c. The binding energies of Zn2p 3/2 and Zn2p 1/2 in Bi-ZnSnO 3 were 1021.68 and 1044.78 eV, respectively, which showed a shift of 0.3 eV in comparison with the binding energy (1021.98 eV and 1045.08 eV) in pure ZnSnO 3 . This may be because Bi doping changed the electron density on the ZnSnO 3 surface and reduced the binding energy. The high-resolution patterns of Sn3d in pure ZnSnO 3 and Bi-ZnSnO 3 are displayed in Figure 6d. The binding energies of Sn3d 5/2 and Sn3d 3/2 in Bi-ZnSnO 3 were located at 486.38 and 494.78 eV, respectively, showing a shift of 0.2 eV compared with the binding energy (486.58 eV and 494.98 eV) in pure ZnSnO 3 , another change that we attributed to Bi doping. In addition, the energy gap between two corresponding peaks of Sn3d was 8.4 eV, indicating that Sn existed in the form of +4 valence [26]. The high-resolution patterns of O1s in pure ZnSnO 3 and Bi-ZnSnO 3 are displayed in Figure 6e,f, where the O1s peak could be divided into three characteristic peaks: lattice oxygen, oxygen vacancies, and chemi-adsorbed oxygen. From Figure 6e, the three characteristic peaks of pure ZnSnO 3 were successively located at 530.18, 531.48, and 532.48 eV, respectively, with area ratios of 72.28%, 21.68%, and 6.04%, respectively. As shown in Figure 6f, the three characteristic peaks of Bi-ZnSnO 3 were located at 530.38, 531.68, and 532.58 eV, with area ratios of 67.34%, 26.10%, and 6.56%, respectively. It is widely accepted that the higher the proportions of oxygen vacancies and chemi-adsorbed oxygen are, the better the gas sensing properties of the sensor will be [27,28]. In this study, the total proportion of oxygen vacancies and chemi-adsorbed oxygen in Bi-ZnSnO 3 was 32.66%, which was higher than 27.72% in pure ZnSnO 3 , further proving that Bi replaced the lattice atoms in pure ZnSnO 3 so as to form oxygen defects.

The Working Temperature of Bi-ZnSnO3
In general, the properties of metal-oxide gas sensors are closely related to their operating temperature [29]. Too high of an operating temperature not only increases the wear rate of gas-sensitive elements but also raises their application and production costs, so a  In general, the properties of metal-oxide gas sensors are closely related to their operating temperature [29]. Too high of an operating temperature not only increases the wear rate of gas-sensitive elements but also raises their application and production costs, so a significant research objective in the field of metal-oxide gas sensors is to reduce their operating temperature [30]. Figure 7 shows the temperature response curves of pure ZnSnO 3 and Bi-ZnSnO 3 of different mass ratios to 100 ppm n-butanol gas. Under normal circumstances, the sensitivity of gas sensors would first grow and then decline with the rise in ambient temperature. This is because both the adsorption rate and the desorption rate are accelerated by a rise in the ambient temperature, but the former is faster than the latter, thus substantially enhancing the material sensitivity. As the ambient temperature further rises after reaching a certain level, the acceleration of the desorption rate exceeds that of the adsorption rate, thus lowering the sensitivity of the gas-sensitive material [31,32]. First, the temperature response of Bi-ZnSnO 3 at the doping concentrations of 0 wt%, 2 wt%, 4 wt%, 5 wt%, and 7 wt% was tested. It can be clearly observed from the temperature response diagram that the five composites operated optimally at 350 • C, 375 • C, 300 • C, 300 • C, and 300 • C, corresponding to the respective sensitivities of 173.25, 339.56, 1450.65, 60.71, and 40.81. Thus, the Bi-ZnSnO 3 reached the highest sensitivity at the doping concentration of 4 wt%, being 8.37 times that of pure ZnSnO 3 . Moreover, the optimal operating temperature of the composite was 300 • C, 50 • C lower than that of pure ZnSnO 3 . The above results show that the optimal proportion of Bi is doped in ZnSnO 3 can effectively improve the gas-sensing properties of the matrix material ZnSnO 3 in addition to lowering the optimal operating temperature.

The Sensitivity Performance of Bi-ZnSnO3 for N-Butanol
As shown in Figure 8a, the sensitivities of ZnSnO3 and Bi-ZnSnO3 with different contents of Bi increased to various extents as the concentration of n-butanol gas increased from 5 to 500 ppm, with Bi-ZnSnO3 (4 wt%) presenting the fastest growth rate in sensitivity, which indicated that the optimal doping concentration of Bi was 4 wt%. In addition, the growth rate of Bi-ZnSnO3 (4 wt%) first increased and then declined with the increase in the concentration of n-butanol gas. This was because a rise in the concentration of nbutanol brings about an increasing number of molecules that occupy a limited number of adsorption sites on the surface of the sensing material, which eventually become saturated and, thus, slow the growth in sensitivity. The concentration-response curves of pure ZnSnO3 and Bi-ZnSnO3 (4 wt%) to n-butanol gas are shown in Figure 8b,

The Sensitivity Performance of Bi-ZnSnO 3 for N-Butanol
As shown in Figure 8a, the sensitivities of ZnSnO 3 and Bi-ZnSnO 3 with different contents of Bi increased to various extents as the concentration of n-butanol gas increased from 5 to 500 ppm, with Bi-ZnSnO 3 (4 wt%) presenting the fastest growth rate in sensitivity, which indicated that the optimal doping concentration of Bi was 4 wt%. In addition, the growth rate of Bi-ZnSnO 3 (4 wt%) first increased and then declined with the increase in the concentration of n-butanol gas. This was because a rise in the concentration of n-butanol brings about an increasing number of molecules that occupy a limited number of adsorption sites on the surface of the sensing material, which eventually become saturated and, thus, slow the growth in sensitivity. The concentration-response curves of pure ZnSnO 3 and Bi-ZnSnO 3 (4 wt%) to n-butanol gas are shown in Figure 8b, . It was clear that the sensitivity of the two gas-sensitive materials to n-butanol gas increased progressively with the n-butanol gas concentration. That is because more and more available n-butanol molecules participate in the redox reaction, contributing to the gradual increase in the sensitivity of the materials. At a given n-butanol gas concentration, the sensitivity of Bi-ZnSnO 3 (4 wt%) was much higher than that of pure ZnSnO 3 due to the modification of Bi. Additionally, the repeatability of the gas sensor was excellent, as the sensitivity increased rapidly after adsorbing n-butanol gas and quickly returned to the initial state when it was exposed to the air. The linear relationship between the sensitivity of Bi-ZnSnO3 and the concentration of n-butanol gas (5-500 ppm) is shown in Figure 9, from which it can be seen that, with the increase in gas concentration, the sensitivity of Bi-ZnSnO3 (4 wt%) composites increased rapidly in the n-butanol environment with a concentration of less than 100 ppm. When the n-butanol concentration exceeded 100 ppm, the sensitivity of the composites increased rapidly. This indicates that the detection limit of the material gradually approaches a peak, so it may be necessary to modify the material in other ways to better detect higher concentrations of n-butanol gas. The two-stage linear correlation coefficients of the Bi-ZnSnO3 (4 wt%) composite are R1 2 = 0.9921and R2 2 = 0.9838, respectively. The results show that the composite can accurately detect n-butanol gas in the range of 5 ppm-500 ppm in practical application. T(res)/T(rec) plays an important role in the evaluation of the properties of sensors [33,34]. T(res) is defined as the time it takes for a gas sensor to reach 90% of its maximum sensitivity to the measured gas, while T(rec) stands for the time it takes for the resistance of a gas sensor to reach 90% of its resistance variation when the sensor is away from the measured gas. We drew the response curves of the two materials to 100 ppm n-butanol gas at their respective optimal operating temperatures ( Figure 8d). The T(res) and T(rec) of pure ZnSnO 3 were 10 s and 17.5 s, while those of Bi-ZnSnO 3 were 8 s and 17 s. By comparison, the T(res) and T(rec) of Bi-ZnSnO 3 (4 wt%) were significantly shorter than those of ZnSnO 3 , meaning that the response and recovery performance of the composites was significantly higher, which in turn improved the application value of the materials.
The linear relationship between the sensitivity of Bi-ZnSnO 3 and the concentration of n-butanol gas (5-500 ppm) is shown in Figure 9, from which it can be seen that, with the increase in gas concentration, the sensitivity of Bi-ZnSnO 3 (4 wt%) composites increased rapidly in the n-butanol environment with a concentration of less than 100 ppm. When the n-butanol concentration exceeded 100 ppm, the sensitivity of the composites increased rapidly. This indicates that the detection limit of the material gradually approaches a peak, so it may be necessary to modify the material in other ways to better detect higher concentrations of n-butanol gas. The two-stage linear correlation coefficients of the Bi-ZnSnO 3 (4 wt%) composite are R 1 2 = 0.9921and R 2 2 = 0.9838, respectively. The results show that the composite can accurately detect n-butanol gas in the range of 5 ppm-500 ppm in practical application.

Sensitivity of ZnSnO3 and Bi-ZnSnO3 to Different VOCs
Its good selectivity can be attributed to the different adsorption capacities and reducibilities of the sensor surface for various measured gases. To further evaluate the selectivity of the two sensor materials, the sensitivities of pure ZnSnO3 and Bi-ZnSnO3 (4 wt%) gas-sensitive elements to six different volatile organic compounds at 100 ppm were measured at their optimal operating temperatures. The results are shown in Figure 10a. The sensitivity of Bi-ZnSnO3 (4 wt%) to n-butanol gas reached 1450.65, which was 35.37 times that (41.01) of ammonia gas, 2.93 times that (495.09) of acetone gas, 6.02 times that (241.05) of methanol gas, 2.54 times that (571.48) of formaldehyde gas, and 2.98 times that (486.58) of ethanol gas. By comparison, the sensitivity of Bi-ZnSnO3 (4 wt%) to n-butanol was much higher than that of other gases, demonstrating a great improvement over pure ZnSnO3. Additionally, Bi-ZnSnO3 presented excellent selectivity for n-butanol gas, probably because Bi doping increased the specific surface area of ZnSnO3 and gave rise to many active sites on the surfaces, thus improving its gas-sensing properties.

Sensitivity of ZnSnO 3 and Bi-ZnSnO 3 to Different VOCs
Its good selectivity can be attributed to the different adsorption capacities and reducibilities of the sensor surface for various measured gases. To further evaluate the selectivity of the two sensor materials, the sensitivities of pure ZnSnO 3 and Bi-ZnSnO 3 (4 wt%) gas-sensitive elements to six different volatile organic compounds at 100 ppm were measured at their optimal operating temperatures. The results are shown in Figure 10a. The sensitivity of Bi-ZnSnO 3 (4 wt%) to n-butanol gas reached 1450.65, which was 35.37 times that (41.01) of ammonia gas, 2.93 times that (495.09) of acetone gas, 6.02 times that (241.05) of methanol gas, 2.54 times that (571.48) of formaldehyde gas, and 2.98 times that (486.58) of ethanol gas. By comparison, the sensitivity of Bi-ZnSnO 3 (4 wt%) to n-butanol was much higher than that of other gases, demonstrating a great improvement over pure ZnSnO 3 . Additionally, Bi-ZnSnO 3 presented excellent selectivity for n-butanol gas, probably because Bi doping increased the specific surface area of ZnSnO 3 and gave rise to many active sites on the surfaces, thus improving its gas-sensing properties.
The response values of Bi-ZnSnO 3 (4 wt%) to a gas concentration of 100 ppm at 300 • C were tested repeatedly within 28 days (four times, once a week) (Figure 10b). The sensitivity obtained from the first test was 1450.65, while that of the fourth test was 1431.60, which indicated that the sensitivity of the composites decreased slightly as the test progressed. Even so, the sensitivity remained above 95% of the initial response value, showing that this gas-sensitive material has good stability for repeated use. that (41.01) of ammonia gas, 2.93 times that (495.09) of acetone gas, 6.02 times that (241.05) of methanol gas, 2.54 times that (571.48) of formaldehyde gas, and 2.98 times that (486.58) of ethanol gas. By comparison, the sensitivity of Bi-ZnSnO3 (4 wt%) to n-butanol was much higher than that of other gases, demonstrating a great improvement over pure ZnSnO3. Additionally, Bi-ZnSnO3 presented excellent selectivity for n-butanol gas, probably because Bi doping increased the specific surface area of ZnSnO3 and gave rise to many active sites on the surfaces, thus improving its gas-sensing properties. The response values of Bi-ZnSnO3 (4 wt%) to a gas concentration of 100 ppm at 300 °C were tested repeatedly within 28 days (four times, once a week) (Figure 10b). The sensitivity obtained from the first test was 1450.65, while that of the fourth test was 1431.60, The performance parameters of the Bi-ZnSnO 3 gas sensor to n-butanol are compared with those of other gas sensors in Table 2. Despite a higher operating temperature, the Bi-ZnSnO 3 gas sensor showed a much better sensitivity to n-butanol gas at 100 ppm than other gas sensors based on different gas-sensitive materials. The optimal operating temperature of the ZnSnO 3 sensor was reduced to 300 • C due to the doping of Bi. All these findings indicate that Bi-ZnSnO 3 , serving as coating materials for gas sensors, is significantly valuable in research and in real-world applications.

The Sensing Mechanism
The adsorption and desorption of oxygen molecules on the surfaces of n-type semiconductors lead to resistance changes, so the gas-sensing properties of n-type semiconductors were compared [41]. When ZnSnO 3 is exposed to air, oxygen molecules in the air will be adsorbed on its surface to form chemi-adsorbed oxygen, generating ionized oxygen O − 2 , O − , and O 2− . As the temperature increases, oxygen molecules capture free electrons from the conduction band of ZnSnO 3 , resulting in a decrease in electron concentration and the formation of a depletion layer, which increases the resistance of the gas sensor [42]. Hence, the sensitive mechanism of n-butanol detection can be illustrated by the Wolkenstein model [43], as follows: The reaction in Formulas (4) and (5) mainly occurred at a temperature less than 100 • C (T < 100 • C); the reaction in Formula (6) mainly occurred when 100 • C < T < 300 • C; and the reaction in Formula (7) mainly occurred when T > 300 • C. A redox reaction between nbutanol gas and ionized oxygen species is initiated by the contact of gas-sensitive elements and n-butanol gas, which releases the captured electrons back to the conduction band of ZnSnO 3 , reducing the thickness of the electron depletion layer, increasing the carrier concentration, and lowering the resistance of the gas-sensitive elements. According to the test results, the optimal operating temperatures of the prepared materials are above 300 • C, and the chemical reactions in this process are mainly as follows [44][45][46]: Combined with the XPS results, these results show that the total proportion of oxygen vacancies and chemi-adsorbed oxygen in Bi-ZnSnO 3 (4 wt%) was from 27.72% to 32.68% higher than that of pure ZnSnO 3 . The increase in oxygen vacancies indicated that more electrons could be captured from the conduction band of Bi-ZnSnO 3 while chemi-adsorbed oxygen directly participated in the reaction of n-butanol gas to release more electrons back to the conduction band of Bi-ZnSnO 3 , leading to a stronger change in the electron depletion layer. This suggests that the introduction of Bi is beneficial to the surface adsorption of oxygen. In addition, according to the BET characterization, the specific surface area of Bi-ZnSnO 3 was also improved, which led to more active sites on the gas sensor surface. Finally, the gas-sensing properties of Bi-ZnSnO 3 were promoted by a more complete electron depletion layer due to an increasing number of oxygen molecules adsorbed on the surface [47]. The gas-sensing mechanism of Bi-ZnSnO 3 is shown in Figure 11. Combined with the XPS results, these results show that the total proportion of oxygen vacancies and chemi-adsorbed oxygen in Bi-ZnSnO3 (4 wt%) was from 27.72% to 32.68% higher than that of pure ZnSnO3. The increase in oxygen vacancies indicated that more electrons could be captured from the conduction band of Bi-ZnSnO3 while chemiadsorbed oxygen directly participated in the reaction of n-butanol gas to release more electrons back to the conduction band of Bi-ZnSnO3, leading to a stronger change in the electron depletion layer. This suggests that the introduction of Bi is beneficial to the surface adsorption of oxygen. In addition, according to the BET characterization, the specific surface area of Bi-ZnSnO3 was also improved, which led to more active sites on the gas sensor surface. Finally, the gas-sensing properties of Bi-ZnSnO3 were promoted by a more complete electron depletion layer due to an increasing number of oxygen molecules adsorbed on the surface [47]. The gas-sensing mechanism of Bi-ZnSnO3 is shown in Figure  11.

Conclusions
ZnSnO3 and Bi-ZnSnO3 were first synthesized via the in situ precipitation method and then were characterized as nanospheres, followed by a study on the gas-sensing properties to n-butanol gas. The test results revealed that Bi-ZnSnO3, compared with pure ZnSnO3, was a superior sensor of n-butanol gas. In particular, Bi-ZnSnO3 (4 wt%) possessed the highest sensitivity of 1450.65, approximately 8.37 times that of pure ZnSnO3, at an optimal operating temperature of 300 °C, which was 50 °C lower than that of pure ZnSnO3. Moreover, Bi-ZnSnO3 had better selectivity and repeatability. The total proportion of oxygen vacancies and chemi-adsorbed oxygen in Bi-ZnSnO3 (4 wt%) was from 27.72% to 32.68% higher than that of pure ZnSnO3. The increase in oxygen vacancies indicated that more electrons could be captured from the conduction band of Bi-ZnSnO3, while chemi-adsorbed oxygen directly participated in the reaction of n-butanol gas to release more electrons back to the conduction band of Bi-ZnSnO3, leading to a stronger change in the electron depletion layer. Altogether, our results suggest that Bi-ZnSnO3 have

Conclusions
ZnSnO 3 and Bi-ZnSnO 3 were first synthesized via the in situ precipitation method and then were characterized as nanospheres, followed by a study on the gas-sensing properties to n-butanol gas. The test results revealed that Bi-ZnSnO 3 , compared with pure ZnSnO 3 , was a superior sensor of n-butanol gas. In particular, Bi-ZnSnO 3 (4 wt%) possessed the highest sensitivity of 1450.65, approximately 8.37 times that of pure ZnSnO 3 , at an optimal operating temperature of 300 • C, which was 50 • C lower than that of pure ZnSnO 3 . Moreover, Bi-ZnSnO 3 had better selectivity and repeatability. The total proportion of oxygen vacancies and chemi-adsorbed oxygen in Bi-ZnSnO 3 (4 wt%) was from 27.72% to 32.68% higher than that of pure ZnSnO 3 . The increase in oxygen vacancies indicated that more electrons could be captured from the conduction band of Bi-ZnSnO 3 , while chemi-adsorbed oxygen directly participated in the reaction of n-butanol gas to release more electrons back to the conduction band of Bi-ZnSnO 3 , leading to a stronger change in the electron depletion layer. Altogether, our results suggest that Bi-ZnSnO 3 have great potential in the detection of n-butanol gas owing to its excellent gas-sensing properties, in contrast to traditional n-butanol sensors.