The present study generated a TiO₂ nanotube array by an anodic oxidation method using a two-electrode system. A platinum metal piece was used as cathode, whereas a titanium piece was used as anode. The experimental processes are as follows: first, 0.5 mm thick Ti foil (area of 0.8 cm × 2.0 cm and purity of 99.94%) was burnished with sandpaper, soaked in 30% HCl solution, and heated to 80 °C for 20 min to remove the surface oxidation layer. Then, the Ti pieces were cleaned by washing with deionized water. The clean Ti pieces acted as the anode, whereas the platinum pieces acted as the cathode in the two-electrode electrochemical electrolysis pool. Between the two electrodes, a constant 20 V of anodic oxidation was applied continuously for 2 h. The electrolyte concentration was 0.1 M HF solution, and the electrolyte pH value was tested using a Model 3000 pH meter. Magnetic stirring was applied to ensure the uniformity of the Ti electrode surface electric current and temperature in the oxidation process and reduce the influence of the double electric layer between the electrolyte and electrode interface. After the reaction, the TiO₂ nanotube array was cleaned by washing with deionized water, and dried in air heated from 2 °C/min to 500 °C in a muffle furnace for 1 h, and then removed after the temperature dropped to room temperature.

A gas-sensitive element of the TiO₂ nanotube array is different from a traditional gas-sensitive element. A TiO₂ nanotube array is generated directly on a metal titanium piece, not coated on traditional Si or A1₂O₃ substrates. In the present study, a conductive silver glue was applied directly to create an electrical contact on the TiO₂ nanotube array surface. The electrodes and TiO₂ nanotube array were pasted firmly together to connect the leads. A schematic diagram of the sensor structure is shown in Figure 1.

Figure 2 presents the detection test device for the TiO₂ nanotube array sensor response measurement of the SF₆ decomposition products. Shown in the figure are the following: 1. quartz glass tube; 2. thermal resistance probe; 3. carbon nanotube sensors; 4. ceramic heating slices; 5. vacuum form; 6. vacuum pump; 7. vent ducts; 8. terminals; 9. AC regulator; 10. temperature display apparatus; 11. impedance analyzer; 12. gas flow meter; and 13. inlet ducts.

In the current experiment, the standard gas containing the SF₆ decomposition products that needs to be measured was injected through the inlet. The gas flow meter controlled and detected the measured gas flow rate. The ceramic heater and heat resistance of the probe controlled and measured the sensor surface temperature. The TiO₂ nanotube array sensor was placed in an airtight quartz glass tube. The sensor detects the characteristics of the sensor resistance, and records the resistance of the whole process through the impedance analyzer. The relative variation of the resistance of the TiO₂ nanotube array sensor (i.e., sensitivity) was calculated using the following formula: R % = ( R − R 0 ) / R 0 × 100 %where R is the sensor resistance after the injection of detected gas and R₀ is resistance in N₂.

The response time of the sensor is the same as 90% of the amount of time that its resistance changes to the maximum amount. The TiO₂ nanotube array has adsorption effects with the oxygen in air and water vapor; hence, to eliminate those factors, this experiment used the dynamic method [13]. The specific steps are as follows: before the sensitivity response test, high-purity N₂ was first injected at a flow rate of 0.1 L/min, and at the same time, connected to the heating power supply. The voltage regulator was adjusted to control the surface temperature of the sensor (required to maintain a certain temperature) until the TiO₂ nanotube sensor array resistance was stable. The value obtained for R₀ was recorded.

Second, one of the SF₆ gas decomposition products, namely SO₂, was passed, and the gas flow velocity in the device was maintained (the same as the previous N₂ gas flow velocity). At this time, the sensor resistance exhibited pronounced changes and achieved stability (waves near one resistance) immediately. In the process, the resistance value R was recorded. Finally, when the sensor resistance was stable, high-purity N₂ was injected again at 0.1 L/min velocity, until the resistance of the sensor gradually achieved numerical stability.

The sample was observed under a scanning electron microscope (SEM). In the present experiment, a JEOL JSM-7000 field emission SEM (Japan) was used. As observed from the SEM images, the anodic oxidation method and the above experimental scheme can grow a TiO₂ nanotube array with a high order and directional growth, whose pipe diameter is about 80 nm and length of about 300 nm (shown in Figure 3).

Figure 4 shows an X-ray diffraction spectrum diagram of the TiO₂ nanotube array. From the figure, the crystal face peak of strong anatase (A in the figure) exists at 2θ = 25.3°, and the 101 crystal face peak of weak rutile exists at 2θ = 27.4° (R in the figure). These findings indicate that the TiO₂ nanotube array is mainly anatase, and a small amount of rutile phase is observed.

The performance of metal oxide semiconductor gas-sensitive materials is greatly influenced by the working temperature. The present study tested the SO₂ gas sensor response curve of the TiO₂ nanotube array sensor at different working temperatures.

The prepared sensor was placed inside the mentioned test device (Figure 2). Through the temperature control device, the surface of the sensor was heated, and its surface temperature was controlled. In the current study, the gas-sensitive characteristics of the TiO₂ nanotube array sensor were tested with 50 ppm SO₂ at surface temperatures ranging from 20 °C to 400 °C.

Figure 5 shows the curve of the sensitivity of the TiO₂ nanotube array sensor at different working temperatures (i.e., surface temperature). The chart indicates that when work temperature is lower, the sensitivity of the sensor increases with the rise in its working temperature. When the temperature reaches 200 °C, the sensor reaches it maximum sensitivity at −76%. When the working temperature continues to rise, the sensitivity tends to be saturated and remains basically unchanged. Therefore, the best working temperature for the TiO₂ nanotube array sensor is about 200 °C.

Figure 6 shows the curve of the response time of the TiO₂ nanotube array sensor at different working temperatures. In the figure, the response time of the sensor decreases with the rise in the working temperature, and has a certain linear relationship with the temperature. Through the linear fit, the linear correlation coefficient R² is 0.98.

The molecular motion and diffusion of the gas speed up because of the increase in temperature, and the gas absorption and the dissociation rate of the sensor's surface increase so the sensor response time decreases with the increasing temperature.

According to procedures of the experiment described in Section 2.3, under the condition that the sensor is under a 200 °C working temperature, the gas sensitivity (response curve) of the TiO₂ nanotube array sensor is measured at SO₂ gas concentrations of 10, 20, 30, 40 and 50 ppm, and the result is shown in Figure 7. The figure illustrates that the bigger the concentration of SO₂ gas, the higher the sensor response (sensitivity).

Therefore, based on the linear fitting, the relation between the concentration of SO₂ gas and sensor sensitivity is a certain linear relation under a low concentration, and the linear correlation coefficient R² is 0.992. As shown in Figure 8, the ability of the sensor to detect low concentrations of SO₂ gas has a certain practical value.

SO₂ is produced by SF₆ decomposition. It is necessary to study the TiO₂ nanotubes sensor response to SF₆ gas decomposition products and background gas SF₆. According to methods and procedures of the experiment in Section 2.3, under the condition that the sensor is under a 200 °C working temperature, the gas sensitivity (response curve) of the TiO₂ nanotube array sensor is measured at 50 ppm SO₂, 50 ppm SOF₂, 50 ppm SO₂F₂ and 99.999% SF₆, and the result is shown in Figure 9. We can see that the sensor responses to 50 ppm SO₂, 50 ppm SOF₂, 50 ppm SO₂F₂ and 99.999% SF₆ are −76%, −7.8%, −5.5% and −7.7%, respectively. This illustrates that the TiO₂ nanotubes sensor has good selectivity for SO₂ gas. It is suitable for checking SO₂ gas, the main component of SF₆ decomposition in the GIS.

According to the experimental method and process in Section 2.3, when the temperature was at 200 °C, the sensor stability test towards the SF6 decomposition component SO₂ was conducted. The test results are shown in Figure 10. Before the sensor stability test, the TiO₂ nanotube array sensor resistance remained basically unchanged when flowing pure N₂ was injected. When injected with 50 ppm SO₂ gas, the sensor resistance changed dramatically and achieved stability immediately. When injected with N₂ after a certain period, the sensor resistance gradually returned to the initial value. The above experiment was tested three times, and we found that the sensor response of SO₂ gas is invariable, and every time through nitrogen cleaning, the resistance of sensor can be returned to the initial value. Thus it is shown that the sensor has good stability. The sensor was tested agian after two months, when we repeated the 50 ppm SO₂ gas experiment once again, and found that the sensor response falls, and resistance can’t return to the initial value, showing that the sensor has experienced chemical poisoning. Ultraviolet light is used for illumination and quick reduction of the resistance of the sensor. When stability is achieved, the resistance becomes smaller than the initial one. By passing nitrogen gas again, the sensor resistance increases gradually, and finally achieves stability. The sensor resistance returns to the initial value; when we performed the 50 ppm SO₂ gas experiment two times, the sensor response reaches the level of the two months ago. That is, through the ultraviolet irradiation, the SO₂ gas molecules adsorbed on the sensor can be washed away completely. Therefore, the TiO₂ nanotube array achieves complete desorption results.

TiO₂ as a sensitive material that detects gases through changes in physical properties, such as electric conductivity, when the gas comes in contact with a TiO₂ molecule surface. Oxygen has a very strong adsorption. Oxygen in the air at room temperature adsorbs physically on the TiO₂ surface. When a certain energy is gained, oxygen adsorbs on the TiO₂ nanotube array sensor surface in the form of chemical adsorption. The common forms of chemical absorption oxygen are O_2ads⁻, O_ads⁻, and O_ads²⁻, which relate to the environmental temperature [15]. The experimental results indicate that under low temperatures, the oxide surface exists in the form of a “molecular ion” O_2ads⁻, and changes into a form of an “atomic ion” O_ads⁻ and O_ads²⁻ with the rise in temperature. At more than 450 K, O_ads⁻ dominates the surface oxygen adsorption. In the current study, the working temperature of the sensor is 200 °C (473.15 K); hence, in the process, the oxygen in the air captures the surface electron of TiO₂, changing into chemical adsorption. The chemical reaction equation is: O 2 + 2 e ′ → 2     O ads −

When the sensor surface has reduced the SO₂ gas, the response between the adsorption of oxygen and the SO₂ gas sensor is as follows: SO 2 + O ads − → SO 2     O ads + e ′

SO₂ gas undergoes an ionic reaction with the surface adsorption oxygen, removes an electron, releases back into the conduction band, and causes the conductivity of the TiO₂ materials to increase, thus causing the resistance to decrease. In this manner, the TiO₂ nanotube sensor plays a sensing function.

This finding is consistent with the experimental results. Although the TiO₂ nanotube array and the TiO₂ thin film semiconductor are different in microstructure and form, the above semiconductor adsorption mechanism is suitable for explaining the gas sensor response of the TiO₂ nanotube array.

As seen in Figure 10, after two months, we repeated the 50 ppm SO₂ gas experiment once again, and found that the sensor response was reduced, and the resistance can't return to the initial value. This phenomenon is due to the residual thermal decomposition of SO₂ molecules fixed in the TiO₂ nanotube arrays as the result of chemical adsorption. The adsorption energy of chemical adsorption is much larger than the physical adsorption capacity; hence, pure nitrogen gas flushing the sensor and low-temperature heating are not sufficient to remove completely the SO₂ molecules left on the sensor by chemical adsorption. Therefore, ultraviolet light is used to desorb SO₂ molecules attached on the sensor. In this manner, the sensor repeatability and service life can be improved because the forbidden bandwidth of the ultraviolet photon energy is almost the same as that of many metal oxide semiconductors. Hence, ultraviolet radiation can be absorbed effectively by the TiO₂ nanotube array. The surface of the film and the inner portion undergo a range of physical and chemical processes. In the case of gas adsorption, ultraviolet radiation can be absorbed by the TiO₂ nanotube array through electron-hole pair excitation, thus increasing the carrier concentration and reducing the grain interface barrier. Through these processes, the TiO₂ nanotube array conductivity can be increased, and the resistor reduced. Ultraviolet radiation can be absorbed directly by gas molecules to produce desorption or stimulate chemical reactions between different types of molecules [16]. This finding indicates that the irradiation of ultraviolet light sensor can remove the remaining SO₂ molecules effectively and thoroughly. This method can improve sensor repeatability and reduce sensor chemical poisoning, thereby increasing the service life of the sensor.