Characterization of Reduced Graphene Oxide (rGO)-Loaded SnO₂ Nanocomposite and Applications in C₂H₂ Gas Detection

Abstract

Acetylene (C₂H₂) gas sensors were developed by synthesizing a reduced graphene oxide (rGO)-loaded SnO₂ hybrid nanocomposite via a facile two-step hydrothermal method. Morphological characterizations showed the formation of well-dispersed SnO₂ nanoparticles loaded on the rGO sheets with excellent transparency and obvious fold boundary. Structural analysis revealed good agreement with the standard crystalline phases of SnO₂ and rGO. Gas sensing characteristics of the synthesized materials were carried out in a temperature range of 100–300 °C with various concentrations of C₂H₂ gas. At 180 °C, the SnO₂–rGO hybrid showed preferable detection of C₂H₂ with high sensor response (12.4 toward 50 ppm), fast response-recovery time (54 s and 23 s), limit of detection (LOD) of 1.3 ppm and good linearity, with good selectivity and long-term stability. Furthermore, the possible gas sensing mechanism of the SnO₂–rGO nanocomposites for C₂H₂ gas were summarized and discussed in detail. Our work indicates that the addition of rGO would be effective in enhancing the sensing properties of metal oxide-based gas sensors for C₂H₂ and may make a contribution to the development of an excellent ppm-level gas sensor for on-line monitoring of dissolved C₂H₂ gas in transformer oil.

1. Introduction

Condition monitoring and fault diagnosis is crucial in ensuring the long-term safe and stable operation of power transformers. Latent faults in the power transformer in early stages can be found by monitoring the operation state of the insulation system, which mainly consists of oil and paper [1]. The occurrence of incipient thermal or discharge faults cause the C-H and C-C bond cleavage of oil-paper insulation system, leading to the formation of various dissolved gases, such as H₂, CH₄, C₂H₆, C₂H₄ and C₂H₂ [2]. These gases turn into insulated defects, which reduce the insulation strength of the oil-paper insulation system and develop into permanent faults. The composition and content of these dissolved gases must be detected in order to identify types and severity, as well as to accurately assess the insulation status of power transformer [3,4,5]. In addition, Acetylene (C₂H₂) is a colorless combustible hydrocarbon. It is the simplest alkyne, with a distinctive odor. It is widely used as a fuel and in many other industrial applications. For example, it is an important raw material for organic synthesis such as in 2-chloro-1,3-butadiene; in lighting; in welding and metal cutting; and in dry-cell batteries, etc. However, it is extremely unstable in its pure form and it is easy to incur fire or explosive accidents when it is liquefied, compressed, heated or mixed with air. Thus, the safety monitoring of C₂H₂ is vital during its production and handling. Use of gas sensors is one of the most widely and effectively used methods to detect the composition and content of C₂H₂ gas [6].

Metal oxides (MOs) that change their electrical conductance while exposed to ambient gas can be utilized as gas sensing materials [7]. Many MOs are applicable to detect the reducing and oxidizing gases, for instance, the conductance of CuO, NiO, In₂O₃, WO₃, TiO₂, Fe₂O₃, etc. show different response to gases. However, ZnO and SnO₂ are the main gas sensing materials for commercial applications and experimental research [8]. Pure SnO₂ is sensitive to all the dissolved gases, but it still has some shortcomings, such as high heating temperature, unsatisfied response and poor selectivity [9].Research on how to efficiently improve gas sensing properties is focused on surface modification [10], doping [11] and morphology or size control [12,13]. Recently, graphene has attracted much attention because of its two-dimensional nanostructure with extremely high specific area and electron mobility. The graphene sensor is expected to exhibit excellent gas sensing performance [14]. However, pure graphene is composed of periodic, closely-packed carbon atoms with a benzene ring structure, which is unfavorable for the adsorption of gas molecules. In this regard, the graphene derivative, which is decorated with oxygen functional groups to increase the adsorption site for gas molecules, was named reduced graphene oxide (rGO) [15]. In particular, many groups have reported the gas sensing properties of rGO-based materials, loaded with MOs nanoparticles, with respect to various gases. The NiO–rGO nanocomposite exhibited high response, good repeatability and acceptable selectivity toward CH₄ gas detection [16]. ZnO–rGO exhibited significantly higher responses than pure ZnO with respect to various gases, including O₂, SO₂, NO₂, CO, C₆H₆, and C₂H₅OH [17]. The MO–rGO (SnO₂, Fe₂O₃ and NiO) composite porous films exhibited much better sensing performances than those of the corresponding pure oxide films in detecting C₂H₅OH gas [18]. TiO₂–rGO shows good sensitivity and excellent selectivity to different concentrations of NH₃ from 5 ppm to 50 ppm at room temperature [19]. Latif and Dickert summarized the recent progress and gas sensing applications in graphene hybrid materials, including fabrication method of graphene-based sensors, nanostructures of hybrid materials, gas sensing mechanisms and future perspectives [20]. Nevertheless, there is few reports about the gas sensing properties of SnO₂ nanoparticle–reduced graphene oxide (SnO₂–rGO) nanocomposites for C₂H₂ gas detection at a ppm level.

Herein, we investigated the gas sensing properties (including optimum operating temperature, response to concentration, dynamic sensing transient, response and recovery time, selectivity and long-term stability) of SnO₂–rGO nanocomposites with respect to C₂H₂ gas detection. The SnO₂–rGO nanomaterials were synthesized via a simple two-step hydrothermal method and characterized by X-ray diffraction (XRD), transmission electron microscopy (TEM), energy dispersive spectroscopy (EDS), Raman spectra, X-ray photoelectron spectroscopy (XPS) and thermogravimetric analyses (TGA). Meanwhile, planar sensors were fabricated to examine the gas sensing performance with respect to C₂H₂ gas. The results indicate that the SnO₂–rGO can efficiently detect the C₂H₂ gas under a low operating temperature, showing better sensing performances than those of rGO alone. Finally, the possible gas sensing mechanisms of the SnO₂-rGO with respect to C₂H₂ detection have been discussed.

2. Experimental Method

2.1. Materials

The commercially available high-purity graphene oxide (GO) was purchased from XFNANO Materials Technology Co., Ltd. (Nanjing, China). Sodium stannate hydrate (Na₂SnO₃·3H₂O), citric acid hydrate (C₆H₈O₇·H₂O) and sodium hydroxide (NaOH) were purchased from Chuandong Chemical Reagent Co., Ltd. (Chongqing, China). All chemicals in our work were of analytical reagent grade and used as-received, without any further purification.

2.1.1. SnO₂ Nanoparticle Synthesis

SnO₂ nanoparticles were synthesized by the following hydrothermal method [21]. First, 2.5 mmol of Na₂SnO₃·3H₂O and 1.0 mmol of C₆H₈O₇·H₂O were dissolved in the mixture of 20 mL deionized water (DIW) and 20 mL anhydrous ethanol (EtOH), and the solution was vigorously stirred for 10 min. Subsequently, 15 mmol of NaOH was added dropwise into the above solution and stirred vigorously for 30 min. Then, the precursor solution was transferred to a 50 mL Teflon-lined autoclave which was heated at 180 °C for 14 h. Afterwards, the product was collected after centrifuging and washing for several times. Finally, the product was calcined at 400 °C for 2 h.

2.1.2. SnO₂–rGO Nanocomposite Preparation

SnO₂–rGO was prepared by compositing as-prepared SnO₂ nanoparticles and GO under the hydrothermal treatment and the detailed preparation process was presented as follows. The same amount (5 mg) of as-prepared SnO₂ nanoparticles and GO were added into 30 mL of DIW with stirring and ultrasonic dispersion for 30 min to achieve uniform dispersion. Then, the dispersion solution was transferred to 50 mL Teflon autoclave (180 °C, 12 h). The products were centrifuged and calcined at 200 °C for 2 h to obtain the SnO₂-rGO nanocomposite. For comparison, the rGO was prepared by the similar method without addition of SnO₂.

2.1.3. Characterization

The microstructure and morphology of as-prepared samples were characterized. X-ray diffraction (XRD) patterns were recorded on a Bruker D8 Advance (Bruker, Karlsruhe, Germany) using Cu Kα radiation at 40 kV and 40 mA. Raman spectra were determined using a HORIBA Jobin Yvon LabRam (HORIBA Jobin Yvon, Paris, France) equipped with a 532 nm wavelength laser and full-range grating. Thermogravimetric analyses (TGA) were carried out with a NETZSCH STA 449F3 (Netzsch, Selb, Germany) unit at a heating rate of 5 °C/min under a nitrogen atmosphere. Scanning electron microscopy (SEM) images were observed by a FEI Nova Nano Field Emission SEM (FEI, Hillsboro, OR, USA) operated at 20 kV. Transmission electron microscopy (TEM) and energy dispersive spectroscopy (EDS) images were performed by an FEI Tecnai G² F20 S-TWIN (FEI, Hillsboro, OR, USA) operated at 200 kV. The Thermo ESCALAB 250Xi spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) was used to recorded the X-ray photoelectron spectroscopy (XPS) images.

2.2. Fabrication of Planar Gas Sensor

Aiming at testing the gas sensing of as-synthesized gas sensing nanomaterials, a planar resistivity-type sensor device was fabricated. The sensor is mainly made up of three layers: sensing materials, Ag–Pd electrodes and Al₂O₃ substrate. For the planar sensor depicted in Figure 1, Ag–Pd electrodes were interdigitally etched on Al₂O₃ with ≈300-nm-thick foil, 0.2-mm electrode width and spacing by screen-printing. The Ag–Pd pastes (9:1 by weight ratio) used in our experiment have been used for chemical sensors in many cases [22,23]. These substrates can work in atmospheres under 300 °C with stable electric and chemical properties, which are based on the interactions of Pd-doping in Ag pastes. The required quantity of nanocomposites was evenly dispersed in the mixture of DI water and ethanol ratio of 1:1, then the paste was dispersed coated onto the interdigital area of Ag–Pd electrodes to form uniform, dense and smoothly deposited thick films by screen-printing technique [24,25]. Finally, after drying overnight, aging occurred at 120 °C for 120 h to ensure surface resistance value remained stable. For this reason, the stability and repeatability of planar sensor was enhanced.

2.3. Gas Sensing Measurements

The gas sensing properties of the fabricated planar gas sensor for C₂H₂ were tested by our gas sensing analysis measurement system, which is made up of an automatic gas distribution equipment and the CGS-1TP intelligent gas sensing analysis system (Elite Technology Co., Ltd., Beijing, China), is shown in Figure 2. This measurement system is convenient for controlling and obtaining the experimental parameters (operating temperature, resistance characteristic, ambient condition, etc.).

The procedure of testing is described below. First, the fabricated planar gas sensor was fixed in the center of heating stage by two adjustable probes, which were used to collect the electrical signals of the sensing materials. Then, the operating temperature of sensor was set and air was delivered into a sealed chamber at a constant flow rate. When the resistance of the sensor remained stable, the C₂H₂ gas was injected into the chamber. After the response was achieved, the gas sensor was exposed to air again. The concentration of C₂H₂ gas was controlled by the mass flow controllers (MFCs) with the following equation:

$${\text{Gas}}_{\text{concentration}}(\text{ppm})=\frac{\text{Flow} {\text{rate}}_{{\text{C}}_2{\text{H}}_2}}{\text{Flow} {\text{rate}}_{\text{air}}+\text{Flow} {\text{rate}}_{{\text{C}}_2{\text{H}}_2}}$$

(1)

The sensor was measured at temperatures ranging between 100 °C and 300 °C for various concentrations of C₂H₂. The response of a sensor is calculated as S = R_a/R_g, where R_a and R_g are the resistances of the sensor in the air and C₂H₂ gas at certain concentrations, respectively. Response time and recovery time of the sensor were defined as the time required to reach 90% of the total resistance change [26]. The whole experiment process conducted at the same environment temperature and relative humidity (25 °C, 35%).

3. Results and Discussion

3.1. Structural Characterization

The samples must be characterized before the measurement of gas sensing performance to identify the microstructure and morphology. The XRD patterns of rGO, SnO₂ and SnO₂–rGO are presented in Figure 3a. It is seen that a broad diffraction peak centered at 2θ = 24.5° with interlayer distance of 3.78 Å of rGO was observed, which is close to the typical (002) diffraction peak of graphite at 2θ = 26.6°. The peak of GO which might appear in the range of 2θ (10°–12°) was not observed, which further indicates that the GO had been almost completely reduced to rGO. Meanwhile, several strong peaks appeared at 2θ of 26.6°, 33.7°, 37.8° and 51.6°, in agreement with the (110), (101), (200) and (211) planes of rutile SnO₂ (JCPDS 77-0452). The main diffraction peaks of SnO₂–rGO were similar with those of SnO₂ and the characteristic peak of rGO is invisible as seen in Figure 3a, probably owing to the fact that the rGO is wrapped by SnO₂ nanocrystals through the hydrothermal treatment, or that the weak peak of rGO is swamped by the presence of the strong peak around 26.6° at (110) plane of the SnO₂ nanocrystals [27]. Figure 3b shows the Raman spectrum results of SnO₂–rGO, which can be used to prove the existence of rGO in the SnO₂–rGO nanocomposite. Two representative peaks were observed around at 1355 cm⁻¹ and 1586 cm⁻¹, corresponding with disorder-induced (D) band and tangential (G) band of rGO. Due to the ultrasonic exfoliation and hydrothermal treatment, the GO was reduced and more defects were introduced which enhance the performance of gas adsorption, leading to the D and G band of rGO having some minor peak shifts to 1360 cm⁻¹ and 1608 cm⁻¹ of GO. Moreover, the three typical peaks located at round 472 cm⁻¹, 641 cm⁻¹ and 781 cm⁻¹ can be matched with the E_g, A_g and B_2g vibration modes of rutile SnO₂ [28]. All these mentioned characteristic peaks appeared in the Raman spectrum of SnO₂–rGO. It was confirmed that the as-synthesized SnO₂–rGO nanocomposite was made up of rGO and SnO₂. The characteristics of combustion of SnO₂–rGO were investigated using thermogravimetric analysis (TGA) and the results are as shown in Figure 3c. The first 2.199% mass loss (approximately 100 °C) was due to water solvent molecules absorbed into the reduced GO bulk material. The following 5.096% decrease (400 °C) represents elimination of the remaining functional groups; further decomposition takes place up to 800 °C. Experimental settings were as follows: temperature scanning rate: 5 °C/min; temperature range 30–800 °C, and purging inert gas: N₂. SnO₂–rGO has good thermal stability in the range of temperatures (100–300 °C).

Figure 4a depicts the TEM and high-resolution TEM (HRTEM) micrographs of rGO. The quality of rGO is good because of its excellent transparency and obvious fold boundary. In Figure 4b, the SnO₂ nanoparticles with irregular grain size up to 10 nm in diameter were successfully synthesized. From the TEM image of SnO₂–rGO in Figure 4c,d, it can be seen that the SnO₂ nanoparticles were uniformly loaded on the rGO nanosheet surface. Meanwhile, the neighboring fringe spacing of the nanoparticle, dispersed on the surface of rGO, was 0.34 nm. This can be attributed to (110) plane of SnO₂ nanocrystals, indicating that the SnO₂–rGO nanocomposites were prepared. Figure 4e shows the selected area electron diffraction (SAED) pattern of SnO₂–rGO; all diffraction rings on this SAED pattern were exactly assigned to the reflections (110), (101), (211), (112) of the rutile SnO₂ phase and (002) of rGO, which agree well with the XRD pattern of SnO₂–rGO in Figure 3a. The energy dispersive spectroscopy (EDS) spectrum of SnO₂–rGO indicated that the Sn element is found in the constituent elements of SnO₂–rGO (C, O and Sn) and the content of the Sn element was calculated as 15.737 wt %, further verifying that SnO₂ was loaded on the surface of rGO.

To further investigate the composition and chemical state of the element, X-ray photoelectron spectroscopy (XPS) spectra of the SnO₂–rGO nanocomposite were carried out and are shown in Figure 5. As depicted in Figure 5a, the peaks of C 1s, O 1s, Sn 4d, Sn 3d and Sn 3p can be observed in survey spectrum. The two peaks with binding energies of 487.5 eV and 496.0 eV corresponded to Sn 3d_5/2 and Sn 3d_3/2 respectively, which indicated the existence of Sn²⁺. Figure 5c shows the C 1s spectrum and binding energy of C-C, C-O; C=O and O-C=O are assigned at 284.7 eV, 285.6 eV, 286.8 eV and 288.9 eV, respectively [29]. Figure 5d displays the O 1s spectrum, which has been separated into three peaks at 530.4 eV, 531.5 eV and 533.7 eV. The peak centered at 530.4 eV was attributed to oxidized metal ions (O_L) in the lattice of Sn–O. The binding energy at 531.5 eV is ascribed to oxygen vacancies (O_V) in the surface of SnO₂–rGO. The high binding energy peak located at 533.7 eV originated from the oxygen atoms (O_S) chemisorbed on the surface of synthesized materials [30]. Based on above results, it is concluded that the as-synthesized nanocomposites are composed of SnO₂ and rGO.

3.2. C₂H₂ Gas Sensing Properties

C₂H₂ gas is one of the main fault characteristic gases dissolved in power transformer oil, which can indicate the transformer discharge faults effectively. Thus, we carried out the gas sensing properties of the fabricated gas sensors for C₂H₂ gas. Figure 6a shows the resistance curve (resistance of the sensor in the air versus temperature) of rGO, SnO₂ and SnO₂–rGO gas sensors. Then the operating temperatures of rGO, SnO₂ and SnO₂–rGO gas sensors were investigated at a temperature range of 100 °C to 300 °C, with an interval of 20 °C. Figure 6b shows the relationship between response and temperature for those gas sensors to 50 ppm C₂H₂ gas. It is obvious that the responses of rGO–SnO₂ first ascended and then descended with the temperature increasing, while those of rGO and SnO₂ increased slowly in the temperature range of measurement. Although the reduced graphene oxide gas sensor shows a good response to gases at room temperature [31], the SnO₂ gas sensor usually works at high temperatures (about 360 °C) [32]. This behavior might be attributed to the potential barrier formed by the chemisorbed oxygen ions (O_n⁻_(ads)) on the SnO₂ surface, which prevents the C₂H₂ molecules from reacting at low temperature. The SnO₂ gas sensor at low temperature possibly underwent a small chemical activation (the adsorption–desorption kinetics of test gas on the sensing materials), and showed a small response. At high temperatures, due to the enhanced chemical activation, the adsorbed gas molecules may escape before their reaction, leading to a response drop. Due the presence of SnO₂ in the SnO₂–rGO nanocomposite, the optimum operating temperature is between that of rGO and SnO₂. The operating temperature when the response of SnO₂–rGO reached the maximum value about 12.4 was 180 °C, which is called the optimum operating temperature and proved that the introduction of rGO was beneficial to reducing the optimum operating temperatures for the SnO₂-based gas sensor.

Figure 6c shows the response-recovery curve of SnO₂-rGO to different concentrations of C₂H₂ gas from 0.5 ppm to 50 ppm at its optimum operating temperature (180 °C). When SnO₂–rGO gas sensor was exposed to the C₂H₂ gas, its response clearly increased. In contrast, the response of SnO₂–rGO gas sensor was decreased backed to the initial state after air was delivered into the chamber. To further study the response-recovery performance, the dynamic sensing transient of SnO₂–rGO sensor to 1 ppm C₂H₂ gas was shown in Figure 6d. By definition, the response and recovery times of SnO₂–rGO were estimated to be around 54 s and 23 s, respectively.

The response curve of SnO₂–rGO gas sensor to various concentrations (from 0.5 ppm to 500 ppm) of C₂H₂ gas at optimum operating temperatures of 180 °C was plotted in Figure 7a. The response of sensor increased steeply from 0.5 ppm to 10 ppm and then increased slowly from 10 ppm to 500 ppm. Sensitivity is one of the most important parameters, which measures the variation of the response as a function of the variation of the concentration (c). The figure of the sensitivity as shown in Figure 7a can be calculated as the derivative of the response curve of SnO₂–rGO. Moreover, Figure 7b shows that response versus concentration maintained good linearity when the gas concentration was below 10 ppm, with R² = 0.994, and the equation for the response y and concentration x can be depicted as:

y = 0.447x + 1.421

(2)

In this case, we defined that the limit of detection is the lowest concentration when the response reached 2. Thus, it can be calculated that when sensing response (y) goes to 2 the concentration (x) goes to 1.3 ppm.

Generally, a chemical sensor may be sensitive not only to its intended chemical for measurement, but also to a number of chemicals that can equally be present in the environment, so selectivity is another important property of gas sensors. The figure of the selectivity can be calculated using Equation (13) presented in [33]. The response of SnO₂–rGO gas sensors to 50 ppm different gases (hydrogen, carbon monoxide, carbon dioxide, methane and acetylene) at 180 °C are described in Figure 7c. The highest response is observed for C₂H₂ sensing performance, which is almost 2.7, 4.5, 9.6 and 1.5 times higher than for H₂, CO, CO₂ and CH₄, respectively. All those results reveal that the SnO₂–rGO sensor can be used for C₂H₂ detection. Better response to C₂H₂ might be attributed to higher reactivity and smaller bond energy of H-C≡C-H (490 kJ/mol) compared with other test gases. Although the bond energy of CH₄ (431 kJ/mol) and H₂ (436 kJ/mol) is close to C₂H₂, interaction strength between the sensing layer and target gas is an effective factor on the sensor response [34]. In the case of the SnO₂–rGO gas sensor, H₂ and CO may decompose easily at about 200 °C, where the exothermic reactions become impulsive, hence showing higher response [35]. The fabricated sensor did not show a remarkable response toward CO₂ due to insufficient interaction strength between sensing layer and exposed gas [36]. Moreover, based on the molecular frontier orbital theory, we calculated the highest occupied molecular (HOMO) and lowest unoccupied molecular (LUMO) values of hydrocarbon gases and conducted response testing of SnO₂. The results showed that the ability of losing electron weakens in the following order: C₂H₄, C₂H₂, C₂H₆ and CH₄, which is in strong agreement with our experimental data [32]. Generally, sensor selectivity is a complicated problems associated with the intricate film nanostructure and complicated sensing mechanisms. Sensor selectivity may need further investigation with respect to critical applications [37]. Furthermore, the stability of SnO₂–rGO sensor upon exposure to 10 ppm, 50 ppm and 200 ppm C₂H₂ gas was investigated. The measurements lasted for 30 days and responses were tested every 5 days. As shown in Figure 7d, the responses to 10 ppm, 50 ppm and 100 ppm C₂H₂ gas were kept at an almost constant level. These observed results demonstrate that the proposed sensor has good long-term stability. Lastly, the comparison of the C₂H₂ gas sensing properties of the SnO₂/rGO gas sensor with previously reported results in C₂H₂ gas sensors using other nanomaterials are shown in

Table 1. Comparison of previously reported and our work on C₂H₂ gas sensors.

Ref.	Material	Temperature (°C)	Measuring Range (ppm)	Response (S = Ra/Rg)	τResponse (s)	τRecovery (s)
[38]	Pure SnO2	300	200–10,000	6.3 (10,000 ppm)	34	>10 min
[39]	Sm2O3/SnO2	180	1–5000	63.8 (1000 ppm)	3	17
[40]	Pt/ZnO	300	50–10,000	43 (1000 ppm)	6	65
[41]	ZnO/rGO 1	250	30–1000	18.2 (100 ppm)	100	28
Our	SnO2–rGO	180	0.5–500	12.4 (50 ppm)	54	23

¹ rGO: reduced graphene oxide.

.

3.3. Gas Sensing Mechanism

According to our experimental results, SnO₂–rGO nanocomposites have excellent performance with respect to detection of C₂H₂, which shows potential for application. The possible sensing mechanism is discussed as follows. Commonly, the sensing mechanism of MO gas sensors is explained by the adsorption–desorption model, the resistance changes of the MOs through the interaction between adsorbed oxygen ions (O²⁻, O⁻ or O₂⁻) and target gas molecules. As we know, rGO nanosheets prepared by chemical or thermal treatment will show a p-type semiconductor characteristic [42]. Thus, the heterojunctions (n-n or n-p) can be formed when rGO is composited with MOs. This will affect the gas sensing performance by impacting on the resistances change of the nanocomposites. In order to explain the gas sensing mechanism in SnO₂–rGO nanocomposites as thoroughly as possible, several possible mechanisms will be considered, as follows.

Firstly, the nanostructure of as-synthesized SnO₂ grains are composed of nanoparticles, as shown in Figure 8b. The adsorption–desorption model and boundary barrier model are described in Figure 8a. As the SnO₂ works as a n-type semiconductor characteristic, when the SnO₂ grain is exposed to the air, the O₂ molecules in air will adsorbed on the surface of the SnO₂ grain to form the adsorbed oxygen ions (O²⁻, O⁻ or O₂⁻). As the electrons on the surface of SnO₂ are captured by O₂ molecules, a wide electron depletion layer (DL) is formed. When the SnO₂ which contains adsorbed oxygen ions comes into contact with reducing gases, the trapped electrons will release back to the surface of SnO₂ and the width of the DL will be reduced. For example, the possible interaction of C₂H₂ has been proposed in the reference [41] as follows:

O₂ + 2e⁻ ↔ 2O⁻(ads)

(3)

C₂H₂ + O⁻(ads) → CO₂ + H₂O + e⁻

(4)

Meanwhile, the electron transfer between different SnO₂ nanoparticles is an important factor for gas sensing behavior. If the electron in the nanoparticle is to transfer to another, the boundary barrier which is formed by space charge must be overcome. The contact resistance in the interface will affect the resistance of the gas sensor. When the SnO₂ is in the air, the electron on the surface of SnO₂ will be trapped and the boundary is raised. In contrast, in the reducing gases the electron is released back to the surface and the barrier is lowered. The increase or decrease of the width of the DL and boundary barrier according to the interaction between adsorbed oxygen ions make an important contribution to the sensing performance of SnO₂.

Secondly, for SnO₂–rGO nanocomposites possessing n-type character, electrons acted as electron carriers. The structure of SnO₂–rGO is shown in Figure 8b and the mechanism is investigated in Figure 8c. When it is exposed to air, three types of electron DL may be formed, as shown in Figure 8c. DL1 is the first electron depletion layer, which is formed on the surface of SnO₂ nanoparticles owing to the adsorbed oxygen ions [43]; DL2 is the second electron depletion layer, which is formed by the electrons and transfers from the surface of SnO₂ to the rGO during the formation of p-n heterojunctions; and DL3 is the third electron depletion layer. It appears in the area where SnO₂ is embed in rGO nanosheets and forms the p-n heterojunctions [18]. All these electron depletion layers will prevent the migration of electron, leading to the high-resistance state of the nanocomposites. When the reducing gas is introduced, the width of these three types DLs will undergo different changes. The trapped electrons in the DL1 will be released back to the SnO₂, decreasing the width of the DL1; generally, that of DL2 is supposed to be decreased because the reducing gas molecules may be adsorbed on the surface of rGO and donate the electrons to it. However, the work functions of SnO₂ and rGO are 4.55 eV and 4.75 eV, respectively. Because the work function of SnO₂ is smaller than that of rGO, the electrons in SnO₂ will transfer to rGO to equate the Fermi level and form a barrier (0.2 eV), as shown in Figure 8c. While the reducing gas is introduced, the value of (E_C-E_F) in SnO₂ will decrease the barrier for electron transfer from SnO₂ to rGO. As a result, the number of electrons in SnO₂ will be sufficient, leading to a smaller resistance of SnO₂ and a broadened width of DL2. Through the relative resistance change of SnO₂ is much larger than that of rGO, the width of DL3 in p-n heterojunctions is reduced. The specific nanostructure of SnO₂-rGO nanocomposites which could cause the different types of electron depletion layers has made great contribution to the change of resistance.

Thirdly, the formation of p-n heterojunctions between SnO₂ and rGO can help to enhance the gas sensing properties, and includes many other factors. As the rGO is obtained by chemical or thermal treatment of graphene oxide, there may be many oxygen functional groups on the surface of rGO, such as hydroxyl (-OH) and carboxyl (-COOH) which will provide more adsorption sites for the formation of adsorbed oxygen ions and target gas molecules [44,45]. Besides, the introduction of rGO can effectively prevent the aggregation of SnO₂ nanoparticles, because there are many overlaps among the SnO₂ nanoparticles that must be generated, reducing the exposed active surface [18]. Due to these factors, the resistance variation shows a larger change when compared to individual SnO₂ or rGO.

According to the proposed possible mechanisms, we conclude that the SnO₂-rGO gas sensing mechanism combines with the effect of the adsorbed oxygen ions and grain boundary barrier of SnO₂, the structural characteristics of rGO and the formation of local p-n heterojunctions between SnO₂ and rGO. All these sources can be considered as the resistance modulation of the SnO₂–rGO nanocomposites. We suggest that the combination of two sensing mechanisms, being related to rGO-SnO₂ p-n heterojunctions and structural characteristics of rGO, is responsible for the enhancement of the gas sensing properties of rGO-loaded SnO₂. Thus, we can safely conclude that this hybrid sensing mechanism, which combines the effects of high specific surface area and local p-n heterojunctions, is primarily responsible for the enhancement of the sensitivity. In comparison, the adsorbed oxygen ions (O²⁻, O⁻ or O₂⁻) and grain boundary barrier of SnO₂ will make contributions to the enhancement of the gas sensing properties for most of metal oxide semiconductor-based gas sensor. However, the rGO-SnO₂ p-n heterojunctions (change of electron depletion layer), oxygen functional groups and high specific surface area (greater number of adsorption sites) are special characteristics for rGO–SnO₂ hybrid materials due to the due to the presence of rGO. Further investigations to study the exact sensing mechanism are needed.

4. Conclusions

This paper demonstrated a high performance C₂H₂ gas sensor based on a SnO₂–rGO nanocomposite, which was synthesized via a facile two-step hydrothermal method. The as-prepared SnO₂-rGO nanocomposites were characterized by XRD, Raman, TEM, EDS and XPS. The planar gas sensor was fabricated to investigate the gas sensing properties of C₂H₂ gas. The gas sensing properties of the as-fabricated planar gas sensor to C₂H₂ were tested using exposure to various concentrations of C₂H₂ gas at the optimum operating temperature (180 °C). The C₂H₂ gas sensor exhibited a good dynamic response (0.5–50 ppm), low limit of detection (1.3 ppm), high response (S = 12.4–50 ppm), fast response–recovery time (54 s, 23 s), good selectivity, long-term stability and maintained good linearity from 0.5 ppm to 10 ppm. Furthermore, the possible gas sensing mechanisms of the SnO₂–rGO nanocomposites were summarized and discussed. Our work may contribute to the development of a ppm-level gas sensor for on-line monitoring of dissolved C₂H₂ gas in transformer oil.