Enhanced NO₂ Gas Sensing Properties Based on Rb-Doped ZnO/In₂O₃ Heterojunctions at Room Temperature: A Combined DFT and Experimental Study

Abstract

In this work, alkali metal Rb-loaded ZnO/In₂O₃ heterojunctions were synthesized using a combination of hydrothermal and impregnation methods. The morphology and structure of the synthesized samples were characterized by X-ray diffraction, field emission scanning electron microscopy, and transmission electron microscopy. The enhancement mechanism of the nitrogen dioxide gas sensing performance of the Rb-loaded ZnO/In₂O₃ heterojunctions was systematically investigated at room temperature using density-functional theory calculations and experimental validation. The experimental tests showed that the Rb-loaded ZnO/In₂O₃ sensor achieved an excellent response value of 24.2 for 1 ppm NO₂, with response and recovery times of 55 and 21 s, respectively. This result is 20 times higher than that of pure ZnO sensors and two times higher than that of ZnO/In₂O₃ sensors, indicating that the Rb-loaded ZnO/In₂O₃ sensor has a more pronounced enhancement in performance for NO₂. This study not only revealed the mechanism by which Rb loading affects the electronic structure and gas molecule adsorption behavior on the surface of ZnO/In₂O₃ heterojunctions but also provides theoretical guidance and technical support for the development of high-performance room-temperature NO₂ sensors.

1. Introduction

Nitrogen dioxide (NO₂) is a toxic and corrosive gas with an irritating odor. NO₂ not only has a significant negative impact on human health, but also has a significant damaging effect on the quality of the environment [1]. NO₂ is one of the major pollutants in industrial emissions and traffic exhausts and is strongly toxic and oxidizing [2]. The Health and Safety Rule Alert states that human beings should avoid spending more than 8 h in an environment with NO₂ at 1 ppm; otherwise, it can cause cardiovascular diseases [3]. Especially prolonged exposure to low concentrations of NO₂ (10 ppm) can lead to bronchitis and even death [4]. In addition, NO₂ can generate ozone (O₃) and acid rain through chemical reactions in the atmosphere, further harming the environment and ecosystem [5].

In recent years, semiconductor oxide-based gas sensors have received significant attention due to their simple structure, high sensitivity, and low cost. Among them, SnO₂ [6], ZnO [7], In₂O₃ [8], and WO₃ [9], as typical n-type semiconductor materials, are suitable materials for NO₂ sensor research due to their excellent electron transport properties and chemical stability. However, single materials often suffer from slow response speeds, poor selectivity, and high operating temperatures in practical applications. In order to overcome these drawbacks, researchers have begun to combine different materials to form heterojunctions, which can significantly improve the sensing performance. Among many heterojunction materials, ZnO/In₂O₃ [10] heterojunctions show excellent gas-sensing performance due to their good energy band matching and interfacial charge transfer properties. For example, Yuan et al. [11] prepared ZnO/In₂O₃ heterostructured nanosheets using a one-step hydrothermal method and discussed the effect of ZnO content on n-butanol gas-sensing performance. Liang et al. [12] achieved rapid detection of ppb-level NO gas in ZnO/In₂O₃ nanocomposites at room temperature (RT) using resonant tunneling modulation. Huang et al. [13] synthesized a one-dimensional ZnO/In₂O₃ nanofiber sensor based on a coaxial electrospinning method, which demonstrated superior selectivity and stability for ethanol gas.

Functionalization with noble [14,15], transition [16], and alkaline metals [17] has also emerged as an effective method to improve sensing performance. It has been reported that loaded metals can act as catalysts to promote the chemical reactions of target gas molecules and improve the response speed and sensitivity of sensors. Meanwhile, metal loading can change the electronic structure of semiconductor materials and regulate their surface electron density, thus affecting the conductivity change of the sensor. For example, Liu et al. [18] loaded Ag onto ZnO/In₂O₃ nanofibers, demonstrating excellent sensitivity, superior selectivity, and long-lasting stability for formaldehyde gas. Wang et al. [19] realized Au-decorated ZnO/In₂O₃ belt-tooth-shaped nanoheterostructures using a chemical vapor deposition process. The gas sensor exhibited excellent sensing performance for C₂H₂ gas at low operating temperatures. Guo et al. [20] designed ultra-small Pt nanoparticle-functionalized ZnO/In₂O₃ nanofibers using a new catalyst-loaded platform (Pt@ZIF-8). The experimental results showed that the Pt-ZnO/In₂O₃ nanofiber-based sensor has excellent acetone response, ultrafast response and recovery time, and a low detection limit.

Although several studies have focused on the application of loaded metals and ZnO/In₂O₃ [21] heterojunctions for detecting NO₂ gas, the mechanism by which Rb loading enhances the NO₂-sensing performance of ZnO/In₂O₃ heterojunctions at room temperature has not yet been systematically investigated. Room-temperature detection can reduce energy consumption, making the sensor more economical and environmentally friendly in practical applications. In addition, NO₂ sensors under room-temperature conditions can realize a wider range of applications, including indoor air quality monitoring, portable environmental testing equipment, and mobile monitoring stations, which all place high demands on the power consumption and stability of the sensors. Therefore, it is of great practical importance to develop NO₂ sensors that can operate efficiently at room temperature.

The aim of this study was to investigate, for the first time, the synergistic effect of Rb-loaded ZnO/In₂O₃ heterojunctions in room-temperature NO₂ sensing using a combination of density functional theory (DFT) calculations and experiments. The Rb-loaded ZnO/In₂O₃ heterojunctions were prepared using hydrothermal and impregnation methods. The morphological structures and surface chemical states were observed using X-ray diffraction (XRD), transmission electron microscopy (TEM), and X-ray photoelectron spectroscopy (XPS). Through theoretical calculations and experimental validation, the mechanism by which loaded Rb influences the electronic structure of the ZnO/In₂O₃ heterojunction surface and the adsorption behavior of gas molecules was revealed. This is expected to provide new ideas for the development of high-performance, room-temperature NO₂ sensors, as well as theoretical guidance and technical support for future sensor design.

2. Materials and Methods

2.1. Preparation of ZnO and In₂O₃ Heterojunctions

Zinc nitrate (Zn(NO₃)₂·6H₂O), indium nitrate (In(NO₃)₃·XH₂O), and sodium hydroxide (NaOH) were purchased from Tianjin Kemiou Chemical Reagent Co., Ltd. (Tianjin, China). Zinc nitrate and indium nitrate with a stoichiometric ratio of 2:1 (0.6:0.3 g) were dissolved in 20 mL of deionized water and stirred at 80 °C for 15 min. Subsequently, they were transferred to a 50 mL pipette, and the solution was ultrasonically pulverized for 10 min while 10 mL of 2 mol/L sodium hydroxide solution was rapidly injected. Then, it was cross-washed with deionized water and ethanol three times and air-dried at 60 °C. Finally, the dried samples were placed in a tube furnace and heated to 850 °C at a heating rate of 5 °C/min and then calcined for 2 h. ZnO/In₂O₃ heterojunction nanocomplexes were obtained.

2.2. Preparation of Rb-Loaded ZnO/In₂O₃ Heterojunctions

Rubidium nitrate (RbNO₃) was purchased from Shanghai Maclin Biochemical Technology Co., Ltd. (Tianjin, China). First, 0.166 g of the ZnO/In₂O₃ heterojunction complex was added to 5 mL of deionized water and stirred for half an hour. Then, 0.147 g of rubidium nitrate was added to 5 mL of ethanol and stirred for half an hour. The above solutions were mixed and ultrasonicated for 60 min before being vacuum-dried at 70 °C for 24 h. Then, the powder was pyrolyzed in an argon atmosphere at 500 °C for 30 min to obtain 1 mol% Rb-ZnO/In₂O₃ [22]. Finally, 2% Rb-ZnO/In₂O₃ and 3% Rb-ZnO/In₂O₃ were prepared through adding 0.294 and 0.441 g of rubidium nitrate, respectively, with all other conditions held constant.

2.3. Material Characterization

The samples’ crystal structure and composition were determined using an X-ray diffractometer (XRD, D/Max 2400, Rigaku, Tokyo, Japan) in the 2θ region of 10°–80° with Cu Kα radiation at a rate of 3°/min. The morphology of the samples was obtained using a field emission scanning electron microscope (SEM; Apreo 2C, Thermo Scientific, Waltham, MA, USA). The internal nanostructure and elemental distribution of the samples were characterized using transmission electron microscopy (TEM; Talos F200S, Thermo Scientific, Waltham, MA, USA) at 200 kV. The composition and chemical states of the elements in the materials were analyzed using a Thermo Fisher K-ALPHA X-ray photoelectron spectrometer (XPS; Thermo Fisher, Waltham, MA, USA). The charge correction was performed with the C 1s peak (284.8 eV) as a reference.

2.4. Fabrication of Sensors and Measurement

A 2 mg sample was mixed with 25 µL of deionized water and thoroughly ground to form a slurry. This slurry was uniformly applied to an alumina substrate pre-coated with gold electrodes (5 × 10 mm). The sensor was dried at room temperature for 12 h to ensure complete evaporation of the water. A cross-sectional image of the sensor was observed using scanning electron microscopy (SEM), as shown in Figure S1. It can be seen that the sensing material formed a uniform sensing film on the substrate, with a thickness of approximately 65 µm.

The sensor’s performance was evaluated using a dynamic gas flow device, as shown in Figure 1. All gases used in this study were supplied in certified cylinders by Dalian GuangMing Special Gas Products Co., Ltd. (Dalian, China). The composition of the simulated air was 70% N₂ and 30% O₂. All other gases were at a concentration of 100 ppm. The simulated air and target gases were mixed into different concentrations in a mixing tank under the precise control of a mass flow controller (MFC; Criterion D500, Horina, Kyoto, Japan) and introduced into the test chamber at a steady flow rate of 300 sccm. The sensors were placed in a test chamber for electrical measurement, with resistance values recorded by a source meter (34470A, Keysight, Santa Rosa, CA, USA) and a computer-aided measurement instrument.

The gas response value was defined as S = R_a/R_g (ammonia, acetone, methanol, formaldehyde, and ethanol) or R_g/R_a (NO₂), where R_a and R_g are the resistance values of the sensor exposure to air and target gas, respectively. The error on the response value was calculated by applying error propagation, considering the load resistance tolerance and instrument uncertainty on the output voltage acquisition [23]. The sensor response and recovery times were determined through measuring the time taken for the sensor to reach 90% of the saturation resistance value during adsorption and desorption [24].

Sensitivity: The sensor was exposed to increasing concentrations ranging from 1 to 20 ppm of NO₂ to sound out different application scenarios.

Repeatability: A 2 mol% Rb ZnO/In₂O₃ film was exposed to five cycles of 10 ppm of NO₂.

Selectivity: The concentrations of the selected gases (NO₂, ethanol, formaldehyde, methanol, acetone, and ammonia) were determined on the basis of the World Health Organization’s annual publication (Air Quality Guidelines) [25], as well as average test levels reported in the literature.

The limit of detection (LOD) of the sensor can be calculated using the following Equation (1) [6].

LOD = 3σ/s

(1)

where σ is the standard deviation of the response value obtained from the blank measurements, where 40 data points were tested in the absence of the target gas; s is the slope of the calibration curve.

2.5. Periodic DFT Calculation Details

In this work, density functional theory (DFT) calculations were conducted using the Vienna Ab initio Simulation Package (VASP 5.2) [26,27,28]. Generalized gradient approximation (GGA) was applied for the exchange-correlation energy and interatomic interactions with Perdew–Burke–Ernzerh adsorption configurations. The cut-off energy was set at 400 eV. Monkhorst–Pack k-point meshes (2 × 6 × 1) were employed for structural relaxation within the Brillouin zone. The convergence criteria for energy and force were 1 × 10⁻⁴ eV and 3 × 10⁻² eV/Å, respectively. In this work, the original computational model was adopted with a ZnO fibrillar zincite structure and an In₂O₃ rhombic-centered hexahedral structure, as shown in Figure 2.

3. Results and Discussion

3.1. Characterization of the Samples

Figure 3 illustrates the XRD spectra of 1–3 mol% Rb-ZnO/In₂O₃ and ZnO/In₂O₃ composites. The XRD patterns of all four samples exhibited characteristic peaks of In₂O₃ and ZnO, as shown in Figure 3a. Among them, those diffraction peaks with 2θ values of 30.49°, 50.88°, and 60.49° belong to the crystalline facets of (222), (440), and (622) of In₂O₃ (JCPDS No. 65-3170), respectively. Meanwhile, those diffraction peaks with 2θ values of 31.60°, 33.30°, and 36.08° belong to the crystalline facets of (100), (002), and (101) of ZnO (JCPDS No. 79-0208), respectively. No other characteristic peaks were detected, indicating the high purity of the obtained composites. It is noteworthy that the loaded Rb metal element, due to its low content, resulted in the absence of characteristic diffraction peaks. This situation has been reported in several literature works [20]. High-resolution XRD maps of the four samples are shown in Figure 3b, in which the characteristic diffraction peaks of In₂O₃ (222) shifted as the loading of Rb increased. This is due to the fact that the ionic radius of Rb⁺ is larger than that of In³⁺ and Zn²⁺ [29,30,31,32].

Figure 4a shows an SEM image of ZnO/In₂O₃. The morphology consists of nanoparticles with diameters of approximately 400 and 100 nm, where the larger nanoparticles are ZnO and the smaller ones are In₂O₃. The SEM image is in general agreement with the images previously reported in the literature, which indicates that the sensitive materials prepared in this study are correct [12]. Figure 4b shows an SEM image of the ZnO/In₂O₃ composite with 2 mol% Rb loading. It can be seen that the loading of Rb did not destroy the overall structure of the ZnO/In₂O₃ heterojunction. The small particles of the In₂O₃ and Rb particles were dispersed around ZnO. Figure S2 shows SEM images of 1 and 3 mol% Rb-ZnO/In₂O₃, and it can be seen that several samples have similar morphology sizes, with only minor differences.

Figure 5a shows a TEM image of 2 mol% Rb-ZnO/In₂O₃. The central part is a ZnO block approximately 800 nm in size, surrounded by small In₂O₃ particles approximately 50 nm in size and dispersed Rb particles. It can be seen that the 2 mol% Rb-ZnO/In₂O₃ composite maintains excellent distribution and size homogeneity. Figure 5b presents a HR-TEM image of 2 mol% Rb-ZnO/In₂O₃. It clearly shows the contact interface between two different crystal structures with lattice spacings of 0.24 and 0.26 nm, corresponding to the In₂O₃ (411) and ZnO (002) faces, respectively. Meanwhile, the lattice spacing of 0.19 nm belongs to the (220) crystal surface of Rb, as shown in Figure 5c. It can be seen from the inset that the Rb atoms are ordered and uniform in size and have no other coordinating atoms. This means that the crystal consists of Rb atoms [33]. Figure 5d shows the EDS elemental mapping analysis of 2 mol% Rb-ZnO/In₂O₃. This analysis shows that Zn (green) is distributed in the middle to form a block, and In (blue), O (red), and Rb (yellow) are distributed around ZnO. Where no other elements were detected, the corresponding elemental content estimation is shown in Figure 5e. This confirms that the prepared Rb-loaded ZnO/In₂O₃ composites consisted of In, O, Zn, and Rb. All of the above characterization results confirm the successful loading of Rb into the ZnO/In₂O₃ heterojunction composites. The dimensions of the ZnO/In₂O₃ heterojunction, as well as the heterostructure of ZnO/In₂O₃, were not changed before or after loading.

The X-ray photoelectron spectra (XPS) of ZnO/In₂O₃ and 2 mol% Rb-ZnO/In₂O₃ are shown in Figure S3. Figure S3a shows the full XPS spectra of ZnO/In₂O₃ and 2 mol% Rb-ZnO/In₂O₃, indicating the presence of zinc, indium, and oxygen. Figure S3b shows the Zn 2p spectra with two significant peaks at 1021.70 and 1044.68 eV, corresponding to Zn 2p_3/2 and Zn 2p_1/2, respectively, with a splitting value of about 23 eV. This confirms that the Zn atoms in all samples are in a fully oxidized state [34]. In Figure S3c, the In 3d spectra show peaks at 444.22 (In 3d_5/2) and 451.72 eV (In 3d_3/2) with a splitting value of about 7.5 eV, indicating the presence of In3⁺ [35]. X-ray photoelectron spectroscopy (XPS, Figure S3d) of the Rb 3d measurements shows that the binding energy at 103.6 eV belongs to the Rb 3d peak. Additionally, the peak for Rb 3d⁺ occurs at 102.3 eV, indicating that only a small fraction of the elemental Rb is oxidized [36]. The atomic percentages of ZnO/In₂O₃ and 2 mol% Rb-ZnO/In₂O₃ are shown in Table S1.

3.2. Gas-Sensing Performance

This work is dedicated to the development of high-performance, room-temperature sensors, so the operating temperature of the gas-sensing tests was uniformly chosen to be room temperature (24 °C). The response values of the ZnO at 140 °C and the ZnO/In₂O₃ heterojunction and 1–3 mol% Rb-loaded ZnO/In₂O₃ sensors at 1, 5, 10, and 20 ppm NO₂ at room temperature were tested, as shown in Figure 6a. It can be seen that the response values of ZnO at an operating temperature of 140 °C were very low (four concentration response values: 1.55, 1.68, 1.82, and 1.93, respectively). Meanwhile, both the ZnO/In₂O₃ sensor and the Rb-ZnO/In₂O₃ sensor showed an enhanced effect on the detection of NO₂. The response values of the pure ZnO/In₂O₃ sensor at four concentrations were 11.7, 37.1, 48.3, and 63, respectively; the response values of 1 mol% Rb-ZnO/In₂O₃ at four concentrations were 16.3, 39.6, 52.4, and 72, respectively; the response values of 2 mol% Rb-ZnO/In₂O₃ at four concentrations were 24.2, 42.9, 58.7, and 83.1; and the response values of 3 mol% Rb-ZnO/In₂O₃ at four concentrations were 15.56, 36.96, 51.37, and 70. The ZnO/In₂O₃ sensor not only achieved room-temperature detection of NO₂ but also demonstrated a significantly improved response value compared to the ZnO sensor. The 2 mol% Rb-ZnO/In₂O₃ sensor achieved a higher response value for NO₂ compared to the ZnO/In₂O₃ sensor, and at a low concentration of 1 ppm, the response value was approximately twice as high. The above experiments show that the Rb-loaded ZnO/In₂O₃ sensor is significantly superior to the ZnO and ZnO/In₂O₃ sensors. Meanwhile, the highest sensitivity to NO₂ was achieved at 2 mol% Rb loading. Therefore, the subsequent gas sensitivity experiments were mainly investigated using the 2 mol% Rb-ZnO/In₂O₃ sensor. Then, the sensitivities of the ZnO/In₂O₃ and 2 mol% Rb-ZnO/In₂O₃ sensor were tested for 100 ppm of ammonia, acetone, methanol, formaldehyde, and ethanol and 1 ppm of NO₂ gas, respectively, as shown in Figure 6b. The two sensors showed excellent gas sensitivity performance for the detection of 1 ppm of NO₂ (the response value of the 2 mol% Rb-ZnO/In₂O₃ sensor was 24.2 and of pure the ZnO/In₂O₃ sensor was 11.7), which is clearly different from the other 100 ppm gases (the response values of both sensors were less than 2). To investigate the selectivity of the Rb-ZnO/In₂O₃ sensors, the selectivity coefficients were defined as S_A/S_B, where S_A and S_B are the values of the sensor response to NO₂ gas and other gases, respectively [15]. The selectivity coefficients of the ZnO/In₂O₃ sensor for ethanol, formaldehyde, methanol, acetone, and ammonia were calculated to be 10.35, 6.22, 6.29, 8.48, and 8.36, respectively. Meanwhile, the selectivity coefficients of the 2 mol% Rb-ZnO/In₂O₃ sensor for ethanol, formaldehyde, methanol, acetone, and ammonia were 20.02, 11.28, 11.49, 18.08, and 15.37, respectively. These results indicate that the introduction of Rb improved the sensitivity and selectivity of the ZnO/In₂O₃ sensor.

Subsequently, the response of the sensor was tested in the concentration range of 1–20 ppm, as was the transient response and recovery at 1 ppm. Figure 7a,b depict the dynamic response changes of the ZnO sensor when exposed to 1–20 ppm of NO₂ at 140 °C and the response recovery curve at 1 ppm of NO₂. The sensor exhibited poor NO₂ response values of only 1.55 at 1 ppm and response and recovery times of 1187 and 1236 s, respectively. Therefore, the sensor’s performance for NO₂ detection at a high temperature of 140 °C is limited. Figure 7c–f, to the contrary, shows the dynamic response changes of the pure ZnO/In₂O₃ and 2 mol% Rb-ZnO/In₂O₃ sensors when exposed to 1–20 ppm of NO₂ at RT, as well as the response recovery curves at 1 ppm of NO₂. The sensor response values for pure ZnO/In₂O₃ at 1, 5, 10, and 20 ppm of NO₂ can be seen in Figure 7c. The resistance of the sensor was essentially stable at a baseline value when synthetic air was introduced, and the resistance showed an increasing trend when NO₂ gas was introduced. Figure 7d showed the transient response curve of the pure ZnO/In₂O₃ sensor in a low-concentration 1 ppm of NO₂ atmosphere, with response and recovery times of 168 and 182 s, respectively. Upon comparison, the ZnO/In₂O₃ sensor showed a substantial increase in the response value at high temperatures (from 1.55 to 11.7) and a reduction in the response and recovery times compared to ZnO.

Figure 8a shows the linear fit curve of the 2 mol% Rb-ZnO/In₂O₃ sensor at an NO₂ concentration of 1–20 ppm. The R² value of the 2 mol% Rb-ZnO/In₂O₃ sensor was close to 1.0, indicating good linearity of the sensor in detecting NO₂. This is important for practical applications requiring gas detection over a wide range. The calculated LOD for the 2 mol% Rb-ZnO/In₂O₃ sensor was 17 ppb, suggesting that this gas sensor can be used for NO₂ detection at lower concentration levels. Figure 8b shows the repeatability test of the 2 mol% Rb-ZnO/In₂O₃ sensor at 10 ppm of NO₂. After five repeated passes of 10 ppm of NO₂ gas, the sensor showed a stable response and could return to the original baseline resistance. This means that the 2 mol% Rb-ZnO/In₂O₃ sensor has good repeatability. In addition, the long-term stability of the 2 mol% Rb-ZnO/In₂O₃ sensor for 10 ppm of NO₂ is shown in Figure 8c. The response fluctuated slightly during the 28-day test period, demonstrating excellent long-term stability.

Table 1. Comparison between the NO₂ sensing performance of the 2 mol% Rb-ZnO/In₂O₃ gas sensor and other literature results.

Sensing Materials	Working Temperature	Concentration (ppm)	Response (Rg/Ra or Ra/Rg)	Response–Recovery Time	Reference
Au-porous ZnO nanowires	RT	1	2.3	Not present	[37]
Pd-ZnO nanowires	100 °C	1	13.5	141/177 s	[38]
ZnSe/ZnO	200 °C	8	10.42	98/141 s	[39]
Pt-ZnO/PrGO	RT	5	1.76	528/702 s	[40]
In2O3/ZnO nanofibers	RT	1	6.0	36/68 s	[41]
ZnO/In2O3	RT	5	2.21	78/610 s	[42]
ZnO/In2O3	RT	10	29.1	61/39 s	[12]
2 mol% Rb-ZnO/In2O3	RT	1	24.2	55/21 s	This work

shows a comparative analysis of the NO₂ sensing performance of this sensor with other sensors documented in the literature. As illustrated in the figure, the Rb-ZnO/In₂O₃ sensor exhibited superior sensitivity compared to both the noble metal-modified ZnO and ZnO-based heterojunctions. Furthermore, in comparison to other sensors, the response value and response–recovery time of the 2 mol% Rb-ZnO/In₂O₃ sensor in this study were again enhanced. This may be attributed to the increase in active sites resulting from the loading of Rb and the rise in surface chemisorbed oxygen, which augments the redox rate. In conclusion, these findings offer novel insights into the potential for enhancing the NO₂ sensing performance of ZnO-based sensors.

3.3. Gas-Sensing Mechanism

Previous reports have generally used the surface depletion model to explain the sensing mechanism, which depends on the conductivity change of the sensing material during the adsorption and desorption of the target gas as shown in Figure 9 [43]. Oxygen is first adsorbed on the surface of ZnO/In₂O₃, then converted into chemisorbed oxygen species according to Equations (2) and (3) [1]. During this process, a large number of free electrons are trapped in their conduction band. As a result, a very thin electron depletion layer tends to form at the grain boundaries, as shown in Figure 9a. Subsequently, when Rb-ZnO/In₂O₃ is in a nitrogen dioxide atmosphere, due to the high electron affinity of nitrogen dioxide molecules, it can simultaneously trap free electrons from the conduction band of Rb-ZnO/In₂O₃ and react with chemisorbed oxygen species on the surface, as shown in Equations (4) and (5) [44]. In this process, the carrier concentration in the sensing material further decreases, the layer thickness increases further, and the conductivity decreases significantly, as shown in Figure 9b. Finally, the sensor returns to its original state after the nitrogen dioxide supply is turned off, as shown in Equation (6) [12,45].

$${\mathrm{O}}_2\left(\mathrm{gas}\right) \to {\mathrm{O}}_2\left(\mathrm{ads}\right)$$

(2)

$${\mathrm{O}}_2\left(\mathrm{ads}\right)+{\mathrm{e}}^{-}\to {\mathrm{O}}_2^{-}\left(\mathrm{ads}\right)$$

(3)

$${\mathrm{N}\mathrm{O}}_2\left(\mathrm{g}\mathrm{a}\mathrm{s}\right)+{\mathrm{e}}^{-}\to {\mathrm{N}\mathrm{O}}_2^{-}\left(\mathrm{a}\mathrm{d}\mathrm{s}\right)$$

(4)

$${\mathrm{N}\mathrm{O}}_2\left(\mathrm{g}\mathrm{a}\mathrm{s}\right)+{\mathrm{O}}_2^{-}+{2\mathrm{e}}^{-}\to {\mathrm{N}\mathrm{O}}_2^{-}\left(\mathrm{a}\mathrm{d}\mathrm{s}\right)+{2\mathrm{O}}^{-}$$

(5)

$${\mathrm{N}\mathrm{O}}_2^{-}\left(\mathrm{a}\mathrm{d}\mathrm{s}\right)+{2\mathrm{O}}^{-}\to {\mathrm{N}\mathrm{O}}_2\left(\mathrm{g}\mathrm{a}\mathrm{s}\right)+{\mathrm{O}}_2+3{\mathrm{e}}^{-}$$

(6)

It has been reported that Rb-loaded ZnO/In₂O₃ can improve the sensing performance of nitrogen dioxide gas at room temperature for two main reasons: first, the loaded Rb metal can act as a catalyst to promote chemical reactions of the target gas molecules and provide more active sites to enhance the adsorption and decomposition of these gas molecules [46,47]. Second, the loaded Rb metal can increase the electron transfer between ZnO/In₂O₃ and the gas and regulate the surface electron density of the ZnO/In₂O₃ material, thus improving the sensitivity and selectivity of the sensor [48]. Even though the above findings have been widely reported, the mechanism by which Rb loading influences the adsorption behavior of gas molecules on the surface of ZnO/In₂O₃ heterojunctions has not yet been thoroughly investigated. Therefore, we performed density functional theory-based calculations to reveal the effect of Rb-ZnO/In₂O₃ on NO₂.

In the previously reported literature, the construction of heterojunctions with ZnO (002) and In₂O₃ (104) facets leads to a stable configuration [12]. Therefore, the (002) facet of the ZnO model and the (104) facet of the In₂O₃ model were investigated. In order to better match ZnO and In₂O₃ heterojunctions, single cell ZnO and In₂O₃ were expanded into 3 × 2 × 1 ZnO and 2 × 1 × 1 In₂O₃ supercells, respectively.

The degree of mismatch between the lattice parameters of the ZnO/In₂O₃ heterojunctions was calculated to be 1.09%, thus meeting the requirements for constructing heterostructure models (i.e., the mismatch is less than 5%). When the degree of mismatch was less than 5%, the constructed heterojunction was not susceptible to lattice mismatch [28,49]. The constructed ZnO/In₂O₃ was optimized, and the front and top views of the optimized structure are shown in Figure 10a. After structural optimization, a transition layer was formed at the interface, new In-O-Zn bonds appeared, and the structure did not collapse with broken bonds.

In order to construct the Rb-loaded ZnO/In₂O₃ model, three adsorption sites were selected for Rb-loaded adsorption. The stability of the Rb-loaded ZnO/In₂O₃ structure can be judged by calculating the binging energy (E_b) as follows [50]:

E_b = E_Rb-ZnO/In2O3 − E_ZnO/In2O3 − E_Rb

(7)

where E_Rb-ZnO/In2O3 represents the total energy of the Rb-ZnO/In₂O₃ model, E_ZnO/In2O3 represents the energy of the ZnO/In₂O₃ model, and E_Rb represents the energy of the Rb monomer. The more negative the E_b, the stronger the structure. According to

Table 2. Binding energy (E_b) parameters of Rb at different adsorption sites of the ZnO/In₂O₃ model.

Structure Configurations	Adsorption Sites	Eb (eV)
Rb-ZnO/In2O3	Zntop	−0.094
Rb-ZnO/In2O3	Otop	−0.325
Rb-ZnO/In2O3	M	−0.386

, the binding energies were −0.094 eV for Zn_top, −0.325 eV for O_top, and −0.386 eV for M. In comparison, it was found that the binding energy of the overall system was the smallest when Rb was adsorbed on the M site of ZnO/In₂O₃. Therefore, the Rb-ZnO/In₂O₃ structure at the M adsorption site was the most stable, as shown in Figure 10b.

In order to better discuss the adsorption energy in the gas-sensitive mechanism analysis, the calculation of adsorption energy (E_ads) is defined as follows [51]:

E_ads = E_Total − E_Rb-ZnO/In2O3 − E_Gas

(8)

where E_Total, E_Rb-ZnO/In2O3, and E_Gas are the total energy of the system, the energy of the Rb-ZnO/In₂O₃ model, and the energy of the gas molecules, respectively, and the magnitude of the adsorption energy represents the strength of the adsorption between the gas molecules and the gas-sensitive material.

In order to discuss the charge transfer quantity better in the gas-sensitive mechanism analysis, the charge transfer quantity ($∆\rho$) can be calculated and defined as follows [52]:

$$∆\rho ={\rho }_{Total}-{\rho }_{Rb-ZnO/In2O3}-{\rho }_{Gas}$$

(9)

where ${\rho }_{Total}$, ${\rho }_{Rb-ZnO/In2O3}$, and ${\rho }_{Gas}$ are the total charge of the surface and adsorbed molecules, the charge of the Rb-ZnO/In₂O₃ model, and the charge of the molecules, respectively. The magnitude of the charge transfer represents the concentration of carriers between the gas molecules and the gas-sensitive material.

Before calculating the adsorption energy and charge transfer, structural optimization is required. Figure 11 shows the optimal adsorption models of the ZnO/In₂O₃ structure and Rb-loaded ZnO/In₂O₃ structures for NO₂. Figure S4 shows the optimal adsorption model of the ZnO/In₂O₃ and Rb-loaded ZnO/In₂O₃ structures for other gases. The calculated adsorption energies and charge transfers of the two structures for different gases are shown in

Table 3. Adsorption energy (E_ads), charge transfer (Q), and bonding distance (d) of ZnO/In₂O₃ and Rb-ZnO/In₂O₃ on different gas molecules.

Structure Configurations	Gas	Eads (eV)	Q (|e|)	d (Å)
ZnO/In2O3	NO2	−1.176	0.031	1.462
ZnO/In2O3	C2H5OH	−0.186	0.001	3.122
ZnO/In2O3	CH3COCH3	−0.319	0.026	2.523
Rb-ZnO/In2O3	NO2	−0.642	0.480	2.339
Rb-ZnO/In2O3	C2H5OH	−0.115	0.008	2.823
Rb-ZnO/In2O3	CH3COCH3	−0.278	0.005	2.837

.

It can be seen from Table 3 that after NO₂ gas adsorption on the ZnO/In₂O₃ structure, an N-O bond was formed between the gas molecule and the surface, with a bond length of 1.462 Å due to the adsorption effect. Compared to the adsorption of the other two gases, NO₂ had the shortest bond length and the largest adsorption energy, indicating that ZnO/In₂O₃ has the best adsorption capacity for NO₂ among the three gases. Meanwhile, the Rb-loaded ZnO/In₂O₃ structure also had the largest adsorption energy for NO₂ among the three gases. From Figure 11b, it can be seen that Rb-N and Rb-O bonds were formed between the Rb-ZnO/In₂O₃ structure and NO₂, and NO₂ gas molecules adsorbed with Rb had a bond length of 2.339 Å, which is the shortest bond length compared to the other gases as well. This implies that the Rb-ZnO/In₂O₃ structure also had the best adsorption capacity for NO₂. It is noteworthy that the adsorption energy of Rb-ZnO/In₂O₃ for NO₂ was smaller than that of ZnO/In₂O₃ for NO₂. This means that NO₂ gas molecules are more easily desorbed on the surface of Rb-ZnO/In₂O₃, resulting in a faster recovery time of the Rb-ZnO/In₂O₃ sensor. This is consistent with the gas-sensitive test results observed previously (the recovery time of the Rb-loaded ZnO/In₂O₃ sensor was indeed shortened from 182 to 21 s). Interestingly, the charge transfer of NO₂ adsorbed on Rb-ZnO/In₂O₃ was significantly larger than that of NO₂ adsorbed on ZnO/In₂O₃. Figure 12 shows the charge difference plots of the adsorbed NO₂ molecules for both structures, and the surface electrons of both structures were transferred toward NO₂. However, it is obvious that the charge density of Rb-ZnO/In₂O₃ was much larger, and its charge transfer was 15 times that of ZnO/In₂O₃. The Rb-loaded metal increased the amount of electron transfer between the ZnO/In₂O₃ surface and NO₂ gas, resulting in greater electron capture by NO₂ gas from the ZnO/In₂O₃ surface. This means that the carrier concentration in the sensing material was further reduced, the depletion layer became thicker, and the conductivity decreased significantly, thus increasing the sensitivity and selectivity of the sensor. The above calculations fully illustrate the depth of the sensitization mechanism of the Rb-loaded ZnO/In₂O₃ sensor to NO₂ gas.

4. Conclusions

In conclusion, pure and alkali metal Rb-loaded ZnO/In₂O₃ was successfully prepared using a combination of hydrothermal and impregnation methods. The experimental tests showed that the 2 mol% Rb-loaded ZnO/In₂O₃ sensor achieved an excellent response value of 24.2 to 1 ppm of NO₂, with response and recovery times of 55 and 21 s, respectively. The response value, response time, and recovery time of the Rb-loaded ZnO/In₂O₃ sensor were significantly better than those of the pure ZnO and ZnO/In₂O₃ sensors. This indicates that Rb-loaded ZnO/In₂O₃ sensors have a more obvious performance enhancement for NO₂.

The adsorption energies of NO₂ with Rb-ZnO/In₂O₃ and ZnO/In₂O₃ and the charge transfer were calculated using DFT, and we found that NO₂ molecules interacted more strongly with the former than the latter, while the degree of charge transfer was larger. These results suggest that the introduction of Rb plays an important role in increasing the adsorption capacity as well as the carrier concentration of the ZnO/In₂O₃ radical. In this work, the mechanism by which Rb loading enhances the NO₂-sensing performance of ZnO/In₂O₃ heterojunctions was thoroughly explored through theoretical calculations and experimental validation, providing theoretical guidance and technical support for the development of high-performance room-temperature NO₂ sensors. Future studies can further explore other rare metal loadings and their effects on the sensing performance of different gases, while the combination of advanced DFT computational simulation techniques will help to better understand the surface and interfacial properties of the materials, thus guiding the development and application of new sensing materials.