Synthesis, Characterization and Enhanced Sensing Properties of a NiO/ZnO p–n Junctions Sensor for the SF6 Decomposition Byproducts SO2, SO2F2, and SOF2

The detection of partial discharge and analysis of the composition and content of sulfur hexafluoride SF6 gas components are important to evaluate the operating state and insulation level of gas-insulated switchgear (GIS) equipment. This paper reported a novel sensing material made of pure ZnO and NiO-decorated ZnO nanoflowers which were synthesized by a facile and environment friendly hydrothermal process for the detection of SF6 decomposition byproducts. X-ray diffraction (XRD), field emission scanning electron microscopy (FESEM), transmission electron microscopy (TEM), high resolution transmission electron microscopy (HRTEM), energy-dispersive X-ray spectroscopy (EDS) and X-ray photoelectron spectroscopy (XPS) were used to characterize the structural and morphological properties of the prepared gas-sensitive materials. Planar-type chemical gas sensors were fabricated and their gas sensing performances toward the SF6 decomposition byproducts SO2, SO2F2, and SOF2 were systemically investigated. Interestingly, the sensing behaviors of the fabricated ZnO nanoflowers-based sensor to SO2, SO2F2, and SOF2 gases can be obviously enhanced in terms of lower optimal operating temperature, higher gas response and shorter response-recovery time by introducing NiO. Finally, a possible gas sensing mechanism for the formation of the p–n junctions between NiO and ZnO is proposed to explain the enhanced gas response. All results demonstrate a promising approach to fabricate high-performance gas sensors to detect SF6 decomposition byproducts.


Introduction
Gas-insulated switchgear (GIS) is used extensively In electrical power systems, where it presents unique advantages, including small space occupation, low failure rates, outstanding breaking capacity, and so on [1,2]. In the event of a short circuit in big systems, the high-voltage circuit breaker (CB) in the switchgear is the last line of defense, so the GIS must remain healthy [3]. Sulfur hexafluoride (SF 6 ) as a widely used insulating medium in GIS due to its excellent insulating and arc-extinguishing properties [4,5]. Despite of the high reliability of GIS equipment, some internal defects could still lead to different degrees of partial discharge (PD) when the GIS equipment has been running for a long time [6]. When exposed to the energy of partial discharges, the SF 6 gas in GIS equipment could decompose and generate SF 4 , SF 3 , and SF 2 [7], as well as various low-fluorine sulfides. The decomposed SF 6

reacts
The chemical compositions of the prepared samples were conducted using energy-dispersive X-ray spectroscopy (EDS). X-ray photoelectron spectroscopy (XPS) was performed on an ESCLAB MKII using monochromatic Al Kα as the X-ray exciting source to investigate the chemical state of elements existing in the products.

The Fabrication of Sensors
Gas sensors were fabricated by a screen-printing technique with planar ceramic substrates [29] Figure 1 (the planar sensor) shows a structural drawing of the sensor, which mainly consists of three parts: ceramic substrate, Ag-Pd interdigital electrodes, and sensing materials. As prepared samples were ground into a fine powder and further mixed with distilled water and ethanol to form a homogeneous paste. This was subsequently screen printed onto the planar ceramic substrate to generate a uniform sensing film, and dried in air at 50 • C for 8 h. Finally, the fabricated gas sensor was aged at an aging test chamber for 12 days to improve the stability and repeatability before testing.

Gas Response Measurement
Gas sensing properties of the fabricated sensors to SOF 2 , SO 2 F 2 , and SO 2 were measured and automatically recorded by a CGS-1TP (Chemical Gas Sensor-1 Temperature Pressure) intelligent gas sensing analysis system, purchased from Beijing Elite Tech Co., Ltd. (Beijing, China) [30]. Figure 1 shows a schematic diagram of the CGS-1TP system. The planar sensor was laid on the test chamber and its electrodes fixed by adjusting the two probes on each side to collect electrical signals. It is convenient to gain sensor resistance, gas response, working temperature, environmental temperature, and relative humidity with this system. performed on an ESCLAB MKII using monochromatic Al Kα as the X-ray exciting source to investigate the chemical state of elements existing in the products.

The Fabrication of Sensors
Gas sensors were fabricated by a screen-printing technique with planar ceramic substrates [29] Figure 1 (the planar sensor) shows a structural drawing of the sensor, which mainly consists of three parts: ceramic substrate, Ag-Pd interdigital electrodes, and sensing materials. As prepared samples were ground into a fine powder and further mixed with distilled water and ethanol to form a homogeneous paste. This was subsequently screen printed onto the planar ceramic substrate to generate a uniform sensing film, and dried in air at 50 °C for 8 h. Finally, the fabricated gas sensor was aged at an aging test chamber for 12 days to improve the stability and repeatability before testing.

Gas Response Measurement
Gas sensing properties of the fabricated sensors to SOF2, SO2F2, and SO2 were measured and automatically recorded by a CGS-1TP (Chemical Gas Sensor-1 Temperature Pressure) intelligent gas sensing analysis system, purchased from Beijing Elite Tech Co., Ltd. (Beijing, China) [30]. Figure 1 shows a schematic diagram of the CGS-1TP system. The planar sensor was laid on the test chamber and its electrodes fixed by adjusting the two probes on each side to collect electrical signals. It is convenient to gain sensor resistance, gas response, working temperature, environmental temperature, and relative humidity with this system. In this study, the gas response was defined as S = Ra/Rg, where Ra and Rg were the electrical resistance of the fabricated sensor in air and in the test gas, respectively. Response and recovery time was defined as the time required by the sensor to achieve 90% of the total resistance after injecting and removing the detected gas. All sensing measurements were tested under laboratory condition with room temperature (25 °C) and constant humidity (50% relative humidity), and repeated several times to ensure the reliability of the results.

Structural and Morphological Characterizations
The crystalline parameters of the as-prepared products were investigated by XRD and the results are presented in Figure 2. It is obvious that all of the peaks for both products can be well indexed and exactly match the standard pattern of ZnO (JCPDS. . No obvious diffraction peaks of NiO were observed in the pattern, suggesting that the introducing of NiO did not change the original ZnO structure. It may be proposed that the amount of NiO is very small and it is highly In this study, the gas response was defined as S = Ra/Rg, where Ra and Rg were the electrical resistance of the fabricated sensor in air and in the test gas, respectively. Response and recovery time was defined as the time required by the sensor to achieve 90% of the total resistance after injecting and removing the detected gas. All sensing measurements were tested under laboratory condition with room temperature (25 • C) and constant humidity (50% relative humidity), and repeated several times to ensure the reliability of the results.

Structural and Morphological Characterizations
The crystalline parameters of the as-prepared products were investigated by XRD and the results are presented in Figure 2. It is obvious that all of the peaks for both products can be well indexed and  . No obvious diffraction peaks of NiO were observed in the pattern, suggesting that the introducing of NiO did not change the original ZnO structure. It may be proposed that the amount of NiO is very small and it is highly dispersed in the ZnO samples. Moreover, the high peak intensities clearly indicate the high degree of crystallization of both of the samples, corresponding to a lower number of lattice defects [31]. No other doping diffraction peaks were observed from Figure 2, which suggest a high purity of our as-prepared products. dispersed in the ZnO samples. Moreover, the high peak intensities clearly indicate the high degree of crystallization of both of the samples, corresponding to a lower number of lattice defects [31]. No other doping diffraction peaks were observed from Figure 2, which suggest a high purity of our as-prepared products. 30   The overall surface morphologies of the pure and NiO-decorated ZnO nanostructures were firstly studied by FESEM. From Figure 3a,b it is clear that both products consist of a large amount of beautiful microflowers with smooth and approximately uniform size, and no other morphologies are observed. Comparing Figure 3a,b, the micro-morphological structures of the pure and NiO-decorated ZnO nanostructures are almost similar, which demonstrate that the introduction of NiO only has a slight impact on the morphology of pure ZnO products. Further morphology characterization was performed by TEM and shown in Figure 3c. The single microflower was composed of nanorods that connected to each other through the center to form 3D flower-like structures. This is in agreement with recently reported synthesized flower-like nanostructure [32]. The HRTEM image is shown in Figure 3d and the interplanar spacing of 0.28 nm as illustrated is consistent well with the (100) planes of ZnO (JCPDS. . And the interplanar distances of 0.24 nm agree well with the lattice spacing of the (111) planes of the cubic NiO [33], which indicate that NiO has already exists in the composites. The overall surface morphologies of the pure and NiO-decorated ZnO nanostructures were firstly studied by FESEM. From Figure 3a,b it is clear that both products consist of a large amount of beautiful microflowers with smooth and approximately uniform size, and no other morphologies are observed. Comparing Figure 3a,b, the micro-morphological structures of the pure and NiO-decorated ZnO nanostructures are almost similar, which demonstrate that the introduction of NiO only has a slight impact on the morphology of pure ZnO products. Further morphology characterization was performed by TEM and shown in Figure 3c. The single microflower was composed of nanorods that connected to each other through the center to form 3D flower-like structures. This is in agreement with recently reported synthesized flower-like nanostructure [32]. The HRTEM image is shown in Figure 3d and the interplanar spacing of 0.28 nm as illustrated is consistent well with the (100) planes of ZnO (JCPDS. . And the interplanar distances of 0.24 nm agree well with the lattice spacing of the (111) planes of the cubic NiO [33], which indicate that NiO has already exists in the composites. dispersed in the ZnO samples. Moreover, the high peak intensities clearly indicate the high degree of crystallization of both of the samples, corresponding to a lower number of lattice defects [31]. No other doping diffraction peaks were observed from Figure 2, which suggest a high purity of our as-prepared products.  The overall surface morphologies of the pure and NiO-decorated ZnO nanostructures were firstly studied by FESEM. From Figure 3a,b it is clear that both products consist of a large amount of beautiful microflowers with smooth and approximately uniform size, and no other morphologies are observed. Comparing Figure 3a,b, the micro-morphological structures of the pure and NiO-decorated ZnO nanostructures are almost similar, which demonstrate that the introduction of NiO only has a slight impact on the morphology of pure ZnO products. Further morphology characterization was performed by TEM and shown in Figure 3c. The single microflower was composed of nanorods that connected to each other through the center to form 3D flower-like structures. This is in agreement with recently reported synthesized flower-like nanostructure [32]. The HRTEM image is shown in Figure 3d and the interplanar spacing of 0.28 nm as illustrated is consistent well with the (100) planes of ZnO (JCPDS. . And the interplanar distances of 0.24 nm agree well with the lattice spacing of the (111) planes of the cubic NiO [33], which indicate that NiO has already exists in the composites.  In order to investigate the element components of the as synthesized samples, energy dispersive X-ray spectroscopy measurements were conducted. Figure 4 shows the EDS spectra of 3 at-% NiO-decorated ZnO products. It is clear that the elemental compositions of Ni, O and Zn, and the atomic percentage of Ni in the samples is calculated to be 2.72 at-%. Meanwhile no other EDS peaks have been found, which confirms our products are of high purity. The chemical compositions and chemical states of the samples were studied by X-ray photoelectron spectroscopy analysis, and the corresponding results are shown in Figure 5. As observed in Figure 5a, the XPS wide survey spectrum of the synthesized hierarchical samples confirms the existence of Zn, Ni, O, and C. It is proposed that C includes not only C compounds adsorbed by the samples from air, but also oily dirt from the apparatus. Figure 5b shows the high resolution scanning XPS spectra of Ni 2p. It is observed from the Ni 2p3/2 main peak and its satellite peak are located at 855.98 eV and 861.58 eV, respectively, while the Ni 2p1/2 main peak and its satellite peak appear at 873.38 eV and 879.68 eV, respectively. These results are in agreement with the previously reported values and further confirm the presence of NiO [34,35]. From Figure 5c, it is clear that the prominent peaks of Zn 2p3/2 and Zn 2p1/2 are at the binding energies of 1021.58 eV and 1044.68 eV. The observed spin-orbit splitting of Zn 2p between Zn 2p3/2 and Zn 2p1/2 is about 23 eV, which is consistent with the corresponding value of pure ZnO, indicating a normal state of Zn 2+ in the synthesized sample [36]. In the spectra of O 1s as shown in Figure 5d, the peaks are located at 530.88 eV and 531.38 eV, and 532.33 eV respectively, which could be attributed to O 2− ions [37,38]. Therefore, it can be deduced that the Zn and Ni are present as ZnO and NiO in this sample, respectively. In order to investigate the element components of the as synthesized samples, energy dispersive X-ray spectroscopy measurements were conducted. Figure 4 shows the EDS spectra of 3 at-% NiO-decorated ZnO products. It is clear that the elemental compositions of Ni, O and Zn, and the atomic percentage of Ni in the samples is calculated to be 2.72 at-%. Meanwhile no other EDS peaks have been found, which confirms our products are of high purity. In order to investigate the element components of the as synthesized samples, energy dispersive X-ray spectroscopy measurements were conducted. Figure 4 shows the EDS spectra of 3 at-% NiO-decorated ZnO products. It is clear that the elemental compositions of Ni, O and Zn, and the atomic percentage of Ni in the samples is calculated to be 2.72 at-%. Meanwhile no other EDS peaks have been found, which confirms our products are of high purity. The chemical compositions and chemical states of the samples were studied by X-ray photoelectron spectroscopy analysis, and the corresponding results are shown in Figure 5. As observed in Figure 5a, the XPS wide survey spectrum of the synthesized hierarchical samples confirms the existence of Zn, Ni, O, and C. It is proposed that C includes not only C compounds adsorbed by the samples from air, but also oily dirt from the apparatus. Figure 5b shows the high resolution scanning XPS spectra of Ni 2p. It is observed from the Ni 2p3/2 main peak and its satellite peak are located at 855.98 eV and 861.58 eV, respectively, while the Ni 2p1/2 main peak and its satellite peak appear at 873.38 eV and 879.68 eV, respectively. These results are in agreement with the previously reported values and further confirm the presence of NiO [34,35]. From Figure 5c, it is clear that the prominent peaks of Zn 2p3/2 and Zn 2p1/2 are at the binding energies of 1021.58 eV and 1044.68 eV. The observed spin-orbit splitting of Zn 2p between Zn 2p3/2 and Zn 2p1/2 is about 23 eV, which is consistent with the corresponding value of pure ZnO, indicating a normal state of Zn 2+ in the synthesized sample [36]. In the spectra of O 1s as shown in Figure 5d, the peaks are located at 530.88 eV and 531.38 eV, and 532.33 eV respectively, which could be attributed to O 2− ions [37,38]. Therefore, it can be deduced that the Zn and Ni are present as ZnO and NiO in this sample, respectively. The chemical compositions and chemical states of the samples were studied by X-ray photoelectron spectroscopy analysis, and the corresponding results are shown in Figure 5. As observed in Figure 5a, the XPS wide survey spectrum of the synthesized hierarchical samples confirms the existence of Zn, Ni, O, and C. It is proposed that C includes not only C compounds adsorbed by the samples from air, but also oily dirt from the apparatus. Figure 5b shows the high resolution scanning XPS spectra of Ni 2p. It is observed from the Ni 2p 3/2 main peak and its satellite peak are located at 855.98 eV and 861.58 eV, respectively, while the Ni 2p 1/2 main peak and its satellite peak appear at 873.38 eV and 879.68 eV, respectively. These results are in agreement with the previously reported values and further confirm the presence of NiO [34,35]. From Figure 5c, it is clear that the prominent peaks of Zn 2p 3/2 and Zn 2p 1/2 are at the binding energies of 1021.58 eV and 1044.68 eV. The observed spin-orbit splitting of Zn 2p between Zn 2p 3/2 and Zn 2p 1/2 is about 23 eV, which is consistent with the corresponding value of pure ZnO, indicating a normal state of Zn 2+ in the synthesized sample [36]. In the spectra of O 1s as shown in Figure 5d, the peaks are located at 530.88 eV and 531.38 eV, and 532.33 eV respectively, which could be attributed to O 2− ions [37,38]. Therefore, it can be deduced that the Zn and Ni are present as ZnO and NiO in this sample, respectively. Therefore, it can be deduced that the Zn and Ni are present as ZnO and NiO in this sample, respectively.  Figure 6 presents the electric resistance properties in pure air of the as prepared pure ZnO and NiO-decorated ZnO sensors at various operating temperatures from 100 °C to 400 °C. As seen in Figure 6, the resistance values of the two sensors decrease when the operating temperature increases, which is an intrinsic characteristic of a semiconductor gas sensor. Compared with pure ZnO sensor, the NiO-decorated ZnO sensor exhibits a higher resistance value at the same working temperature, which might be beneficial to the following gas sensing performances. In general, the sensing performance of fabricated gas sensors is strongly influenced by the operating temperature. In order to demonstrate that the addition of NiO nanoparticles is an effective way to enhance the gas sensing properties of ZnO-based gas sensor, the effects of various different operating temperatures ranging from 140 °C to 360 °C on fabricated sensors to 100 ppm of SO2, SOF2,   Figure 6 presents the electric resistance properties in pure air of the as prepared pure ZnO and NiO-decorated ZnO sensors at various operating temperatures from 100 • C to 400 • C. As seen in Figure 6, the resistance values of the two sensors decrease when the operating temperature increases, which is an intrinsic characteristic of a semiconductor gas sensor. Compared with pure ZnO sensor, the NiO-decorated ZnO sensor exhibits a higher resistance value at the same working temperature, which might be beneficial to the following gas sensing performances.   Figure 6 presents the electric resistance properties in pure air of the as prepared pure ZnO and NiO-decorated ZnO sensors at various operating temperatures from 100 °C to 400 °C. As seen in Figure 6, the resistance values of the two sensors decrease when the operating temperature increases, which is an intrinsic characteristic of a semiconductor gas sensor. Compared with pure ZnO sensor, the NiO-decorated ZnO sensor exhibits a higher resistance value at the same working temperature, which might be beneficial to the following gas sensing performances. In general, the sensing performance of fabricated gas sensors is strongly influenced by the operating temperature. In order to demonstrate that the addition of NiO nanoparticles is an effective way to enhance the gas sensing properties of ZnO-based gas sensor, the effects of various different operating temperatures ranging from 140 °C to 360 °C on fabricated sensors to 100 ppm of SO2, SOF2, and SO2F2 were respectively investigated and illustrated in Figure 7. It is found that the response of In general, the sensing performance of fabricated gas sensors is strongly influenced by the operating temperature. In order to demonstrate that the addition of NiO nanoparticles is an effective way to enhance the gas sensing properties of ZnO-based gas sensor, the effects of various different operating temperatures ranging from 140 • C to 360 • C on fabricated sensors to 100 ppm of SO 2 , SOF 2 , and SO 2 F 2 were respectively investigated and illustrated in Figure 7. It is found that the response of    Figure 8a, the response continuously increases with gas concentrations ranging from 5 to 800 ppm. Moreover, it is clear that the NiO-decorated ZnO sensor demonstrates enhanced sensing properties for SO2 detection compared with the pure ZnO sensor. Figure 8b exhibits a linear calibration curve of the sensors with SO2 concentration in the range of 5~100 ppm. In this area, the linear fitting functions of pure and NiO-decorated ZnO gas sensors are y = 0.288x + 3.397 and y = 0.761x + 13.049, respectively. Their linear correlation coefficients R 2 are as high as 0.996 and 0.987, which indicates that the response of SO2 concentration variation at low concentration is close to linear.

Gas-Sensing Properties
The gas responses of the pure and NiO-decorated ZnO sensors to SOF2 with different concentrations at their respective optimum working temperatures were measured and are shown in Figure 9. The responses of the fabricated sensors increase rapidly in linearity with increasing SOF2 concentration below 200 ppm, but increase much more slowly above 200 ppm. Moreover, we can also find that the response of the NiO-decorated ZnO sensor is nearly two times higher, on average, than that of the pure ZnO sensor. As Figure 9b exhibits, the linear relationship of the pure and NiO-decorated ZnO sensors between the sensing response and gas concentration in the range of 5~100 ppm are fitted by the equation y = 0.121x + 0.797 and y = 0.203x + 1.398, with high liner correlation coefficients R 2 of 0.993 and 0.994, respectively.   Figure 8a, the response continuously increases with gas concentrations ranging from 5 to 800 ppm. Moreover, it is clear that the NiO-decorated ZnO sensor demonstrates enhanced sensing properties for SO 2 detection compared with the pure ZnO sensor. Figure 8b exhibits a linear calibration curve of the sensors with SO 2 concentration in the range of 5~100 ppm. In this area, the linear fitting functions of pure and NiO-decorated ZnO gas sensors are y = 0.288x + 3.397 and y = 0.761x + 13.049, respectively. Their linear correlation coefficients R 2 are as high as 0.996 and 0.987, which indicates that the response of SO 2 concentration variation at low concentration is close to linear.
The gas responses of the pure and NiO-decorated ZnO sensors to SOF 2 with different concentrations at their respective optimum working temperatures were measured and are shown in Figure 9. The responses of the fabricated sensors increase rapidly in linearity with increasing SOF 2 concentration below 200 ppm, but increase much more slowly above 200 ppm. Moreover, we can also find that the response of the NiO-decorated ZnO sensor is nearly two times higher, on average, than that of the pure ZnO sensor. As Figure 9b exhibits, the linear relationship of the pure and NiO-decorated ZnO sensors between the sensing response and gas concentration in the range of 5~100 ppm are fitted by the equation y = 0.121x + 0.797 and y = 0.203x + 1.398, with high liner correlation coefficients R 2 of 0.993 and 0.994, respectively.  Figure 10 shows the responses of the pure and NiO-decorated ZnO sensors as a function of SO2F2 with different concentrations at their respective optimum operating temperatures. As Figure  10a shows that the gas responses of the sensors increase rapidly with increasing SO2F2 concentration, but increase much more slowly with the further increase of the SO2F2 concentration and nearly reach their saturation at about 600 ppm. Furthermore, the response of the NiO-decorated ZnO sensor is higher than that of the pure one. From Figure 10b the linear fitting functions of pure and NiO-ZnO gas sensors to 5~100 ppm of SO2F2 are y = 0.175x + 0.270 and y = 0.358x + 1.752, respectively. And their linear correlation coefficients R 2 were calculated to be 0.990 and 0.995, respectively.
According to the sensing results mentioned above, it can be concluded that the gas responses of the NiO-decorated ZnO sensor to SO2, SO2F2 and SOF2 is higher than that of the pure one. Moreover, the responses of the prepared sensors to these three test gases at low concentration have a good linear relationship. These good linear relationships indicate that the fabricated gas sensors are more likely to be used in practice [39,40].  Figure 10 shows the responses of the pure and NiO-decorated ZnO sensors as a function of SO2F2 with different concentrations at their respective optimum operating temperatures. As Figure  10a shows that the gas responses of the sensors increase rapidly with increasing SO2F2 concentration, but increase much more slowly with the further increase of the SO2F2 concentration and nearly reach their saturation at about 600 ppm. Furthermore, the response of the NiO-decorated ZnO sensor is higher than that of the pure one. From Figure 10b the linear fitting functions of pure and NiO-ZnO gas sensors to 5~100 ppm of SO2F2 are y = 0.175x + 0.270 and y = 0.358x + 1.752, respectively. And their linear correlation coefficients R 2 were calculated to be 0.990 and 0.995, respectively.
According to the sensing results mentioned above, it can be concluded that the gas responses of the NiO-decorated ZnO sensor to SO2, SO2F2 and SOF2 is higher than that of the pure one. Moreover, the responses of the prepared sensors to these three test gases at low concentration have a good linear relationship. These good linear relationships indicate that the fabricated gas sensors are more likely to be used in practice [39,40].  Figure 10 shows the responses of the pure and NiO-decorated ZnO sensors as a function of SO 2 F 2 with different concentrations at their respective optimum operating temperatures. As Figure 10a shows that the gas responses of the sensors increase rapidly with increasing SO 2 F 2 concentration, but increase much more slowly with the further increase of the SO 2 F 2 concentration and nearly reach their saturation at about 600 ppm. Furthermore, the response of the NiO-decorated ZnO sensor is higher than that of the pure one. From Figure 10b the linear fitting functions of pure and NiO-ZnO gas sensors to 5~100 ppm of SO 2 F 2 are y = 0.175x + 0.270 and y = 0.358x + 1.752, respectively. And their linear correlation coefficients R 2 were calculated to be 0.990 and 0.995, respectively.
According to the sensing results mentioned above, it can be concluded that the gas responses of the NiO-decorated ZnO sensor to SO 2 , SO 2 F 2 and SOF 2 is higher than that of the pure one. Moreover, the responses of the prepared sensors to these three test gases at low concentration have a good linear relationship. These good linear relationships indicate that the fabricated gas sensors are more likely to be used in practice [39,40].  Figure 11 presents the response and recovery curves of the NiO-decorated ZnO sensor to 100 ppm SO2, SO2F2 and SOF2 working at their respective optimum operating temperature. As seen in Figure 10, the common feature of these three test gases is that the response of the sensor increases rapidly when gas is injected into the gas chamber for sensing and dramatically decreases to its initial value when the gas is removed. According to the definition above, the response times to 100 ppm SO2, SO2F2 and SOF2 were calculated to be about 12, 16 and 18 s, while the recovery times were calculated to be about 16, 20 and 22 s, respectively.  Figure 11. Response-recovery curves of the NiO-decorated ZnO gas sensor to 100 ppm SO2, SO2F2 and SOF2.
Finally, the long-time stability and repeatability of the NiO-decorated ZnO sensor to 100 ppm SO2, SO2F2 and SOF2 at their respective optimum working temperatures were measured and shown in Figure 12. As the figure shows, all of the gas responses change slightly and remain at a nearly constant value during the long experimental cycles, which indicates the excellent long-term stability and repeatability of the fabricated sensors for detecting these three kinds of SF6 decomposition byproducts.  Figure 11 presents the response and recovery curves of the NiO-decorated ZnO sensor to 100 ppm SO 2 , SO 2 F 2 and SOF 2 working at their respective optimum operating temperature. As seen in Figure 10, the common feature of these three test gases is that the response of the sensor increases rapidly when gas is injected into the gas chamber for sensing and dramatically decreases to its initial value when the gas is removed. According to the definition above, the response times to 100 ppm SO 2 , SO 2 F 2 and SOF 2 were calculated to be about 12, 16 and 18 s, while the recovery times were calculated to be about 16, 20 and 22 s, respectively.  Figure 11 presents the response and recovery curves of the NiO-decorated ZnO sensor to 100 ppm SO2, SO2F2 and SOF2 working at their respective optimum operating temperature. As seen in Figure 10, the common feature of these three test gases is that the response of the sensor increases rapidly when gas is injected into the gas chamber for sensing and dramatically decreases to its initial value when the gas is removed. According to the definition above, the response times to 100 ppm SO2, SO2F2 and SOF2 were calculated to be about 12, 16 and 18 s, while the recovery times were calculated to be about 16, 20 and 22 s, respectively.  Figure 11. Response-recovery curves of the NiO-decorated ZnO gas sensor to 100 ppm SO2, SO2F2 and SOF2.
Finally, the long-time stability and repeatability of the NiO-decorated ZnO sensor to 100 ppm SO2, SO2F2 and SOF2 at their respective optimum working temperatures were measured and shown in Figure 12. As the figure shows, all of the gas responses change slightly and remain at a nearly constant value during the long experimental cycles, which indicates the excellent long-term stability and repeatability of the fabricated sensors for detecting these three kinds of SF6 decomposition byproducts. Finally, the long-time stability and repeatability of the NiO-decorated ZnO sensor to 100 ppm SO 2 , SO 2 F 2 and SOF 2 at their respective optimum working temperatures were measured and shown in Figure 12. As the figure shows, all of the gas responses change slightly and remain at a nearly constant value during the long experimental cycles, which indicates the excellent long-term stability and repeatability of the fabricated sensors for detecting these three kinds of SF 6 decomposition byproducts.

Sensing Mechanisms
The sensing mechanism for ZnO, a typical n-type semiconductor gas sensing material, has been interpreted in some former papers [41,42]. It is believed that the sensing properties of ZnO chemical gas sensors are mainly determined by the change of electric resistance, which is fundamentally attributed to the chemical adsorption and desorption process of target gas molecules on the surface of gas sensing materials [43]. In an air atmosphere, oxygen could be absorbed on the ZnO surface acting as a trap capturing electrons from its conduction band to form a depletion region on the surface. This depletion region would lead to an increase of the sensor resistance. When exposed to a target gas, the test gas molecules react with the adsorbed oxygen and the trapped electrons are released back into the conduction band, and thus as decreased resistance is measured [44]. According to the definition of gas response (Ra/Rg) the response of the sensor increases.
It was observed in the present study that the sensing properties of ZnO were greatly enhanced due to the introduction of NiO. Two main reasons satisfactorily explain this interesting performance. It has been acknowledged that the gas sensing properties of semiconductor nanomaterials with different morphologies are different [45]. In this study, the synthesized flower-like nanomaterials offer a larger surface accessibility, which will be favorable to adsorption and desorption of target gas molecules, and this unique morphology is helpful to improve the gas sensing response. Moreover, the formation of p-n junctions between NiO and ZnO in the grain boundaries results in an enhancement of gas sensing. A hetero p-n junction system has been proposed to explain the enhanced gas response of Co3O4-WO3 [46], NiO-SnO2 [47], LaFeO3-SnO2 [48] heterocontacts.
As we all know, ZnO shows n-type conductivity by electrons and NiO shows p-type conductivity by holes. By introducing NiO into ZnO, the electrons in ZnO and holes in NiO transport in opposite directions until the system reaches equalization of the Fermi levels, leading to the formation of a p-n junction, as shown in Figure 13a [49]. This is the thermal equilibrium of the formation of a p-n junction [50]. When the composite was exposed to air, oxygen molecules can absorb on the surface of the material and form chemically adsorbed oxygen. With different operating temperatures chemisorbed oxygen exists in various forms [51], as shown in the reaction Equations (1)-(4).

Sensing Mechanisms
The sensing mechanism for ZnO, a typical n-type semiconductor gas sensing material, has been interpreted in some former papers [41,42]. It is believed that the sensing properties of ZnO chemical gas sensors are mainly determined by the change of electric resistance, which is fundamentally attributed to the chemical adsorption and desorption process of target gas molecules on the surface of gas sensing materials [43]. In an air atmosphere, oxygen could be absorbed on the ZnO surface acting as a trap capturing electrons from its conduction band to form a depletion region on the surface. This depletion region would lead to an increase of the sensor resistance. When exposed to a target gas, the test gas molecules react with the adsorbed oxygen and the trapped electrons are released back into the conduction band, and thus as decreased resistance is measured [44]. According to the definition of gas response (Ra/Rg) the response of the sensor increases.
It was observed in the present study that the sensing properties of ZnO were greatly enhanced due to the introduction of NiO. Two main reasons satisfactorily explain this interesting performance. It has been acknowledged that the gas sensing properties of semiconductor nanomaterials with different morphologies are different [45]. In this study, the synthesized flower-like nanomaterials offer a larger surface accessibility, which will be favorable to adsorption and desorption of target gas molecules, and this unique morphology is helpful to improve the gas sensing response. Moreover, the formation of p-n junctions between NiO and ZnO in the grain boundaries results in an enhancement of gas sensing. A hetero p-n junction system has been proposed to explain the enhanced gas response of Co 3 O 4 -WO 3 [46], NiO-SnO 2 [47], LaFeO 3 -SnO 2 [48] heterocontacts.
As we all know, ZnO shows n-type conductivity by electrons and NiO shows p-type conductivity by holes. By introducing NiO into ZnO, the electrons in ZnO and holes in NiO transport in opposite directions until the system reaches equalization of the Fermi levels, leading to the formation of a p-n junction, as shown in Figure 13a [49]. This is the thermal equilibrium of the formation of a p-n junction [50]. When the composite was exposed to air, oxygen molecules can absorb on the surface of the material and form chemically adsorbed oxygen. With different operating temperatures chemisorbed oxygen exists in various forms [51], as shown in the reaction Equations (1)- (4).
where R is one of SF6 decomposition component gases, and O − (ads) denotes the adsorbed oxygen on the material's surface. Owing to the electron-hole recombination, the hole concentration of p-type NiO decreases, and as a result the concentration gradient of the p-n junction is reduced. Thus, the effect of carriers transport becomes weak and the depletion layer becomes thin [50,54], as shown in Figure 13b. The resistance of the sensor in the test gas is further decreased [55]. In short, compared with a pure ZnO sensor, the formation of p-n junctions greatly increase the resistance of the ZnO sensor in air and further decreases the resistance in the test gas. Thus, based on the definition of gas response (Ra/Rg), the responses of the gases are greatly enhanced due to the variation of resistance [49].

Conclusions
In summary, pure and 3 at-% NiO-decorated ZnO sensing materials were successfully synthesized via a one-step hydrothermal method and characterized by XRD, FESEM, TEM, HRTEM, EDS and XPS, respectively. Their gas sensing properties toward the SF6 decomposition byproducts SO2, SO2F2, and SOF2 were systemically researched with a CGS-1TP gas sensing analysis system. Comparative results demonstrate that the as prepared NiO-decorated ZnO sensing material exhibits enhanced gas sensing properties, including lower optimal working temperature, higher gas response and shorter response-recovery time. Such good performance is attributable to the formation of p-n junctions between NiO and ZnO. This work provides an effective way for designing high-performance gas sensors to detect SF6 decomposition byproducts like SO2, SO2F2, and SOF2. Moreover, the resistance of the NiO-decorated ZnO in air (Ra) will be even higher than without the heterojunctions (pure ZnO) due to a new depletion layer between NiO and ZnO [33,52]. Nevertheless, when the composite was exposed to test gases (SO 2 , SOF 2 and SO 2 F 2 ), the gas molecules react with the absorbed oxygen species of gas sensing material and release the electrons back to the material. The three SF 6 decomposed components all play the role of an electron donating gas [53]. The reaction equation is as follows: R + O − (ads) «= =» RO (ads) + e − where R is one of SF 6 decomposition component gases, and O − (ads) denotes the adsorbed oxygen on the material's surface. Owing to the electron-hole recombination, the hole concentration of p-type NiO decreases, and as a result the concentration gradient of the p-n junction is reduced. Thus, the effect of carriers transport becomes weak and the depletion layer becomes thin [50,54], as shown in Figure 13b. The resistance of the sensor in the test gas is further decreased [55]. In short, compared with a pure ZnO sensor, the formation of p-n junctions greatly increase the resistance of the ZnO sensor in air and further decreases the resistance in the test gas. Thus, based on the definition of gas response (Ra/Rg), the responses of the gases are greatly enhanced due to the variation of resistance [49].

Conclusions
In summary, pure and 3 at-% NiO-decorated ZnO sensing materials were successfully synthesized via a one-step hydrothermal method and characterized by XRD, FESEM, TEM, HRTEM, EDS and XPS, respectively. Their gas sensing properties toward the SF 6 decomposition byproducts SO 2 , SO 2 F 2 , and SOF 2 were systemically researched with a CGS-1TP gas sensing analysis system. Comparative results demonstrate that the as prepared NiO-decorated ZnO sensing material exhibits enhanced gas sensing properties, including lower optimal working temperature, higher gas response and shorter response-recovery time. Such good performance is attributable to the formation of p-n junctions between NiO and ZnO. This work provides an effective way for designing high-performance gas sensors to detect SF 6 decomposition byproducts like SO 2 , SO 2 F 2 , and SOF 2 .