Hydrogen sensing enhancement of zinc oxide nanorods via voltage biasing

The capability of zinc oxide (ZnO) as a hydrogen sensing element has been pushed to its limits. Different methods have been explored to extend its sensing capability. In this paper, we report a novel approach which significantly improves the hydrogen sensing capability of zinc oxide by applying a bias voltage to ZnO nanorods as the sensing elements. Zinc oxide in the form of aligned nanorods was first synthesized on an Au-coated Si(111) substrate using a facile method via the galvanic-assisted chemical process. The sensing performance of the zinc oxide nanorods was investigated in response to the applied biasing voltage. It was found that the sensitivity, response time and detection limit of the ZnO sensing elements were dramatically improved with increasing bias voltage. A 100% increment in sensing response was achieved for the detection of 2000 ppm hydrogen gas when the bias voltage was increased from −2 to −6 V with 70% reduction in response and recovery times. This remarkable sensing performance is attributed to the reaction of hydrogen with chemisorbed oxygen ions on the surface of the ZnO nanorods that served as the electron donors to increase the sensor conductance. Higher reverse bias voltages sweep the electrons faster across the electrodes. This shortened the response time and, at the same time, depleted the electrons in the sensor elements and weakens oxygen adsorption. The oxygen ions could then be readily removed by hydrogen, leading to a higher sensitivity of the sensors. This, therefore, envisages a way for high-speed hydrogen gas sensing with high detection sensitivities.

TFC, 0000-0002-8019-8197 The capability of zinc oxide (ZnO) as a hydrogen sensing element has been pushed to its limits. Different methods have been explored to extend its sensing capability. In this paper, we report a novel approach which significantly improves the hydrogen sensing capability of zinc oxide by applying a bias voltage to ZnO nanorods as the sensing elements. Zinc oxide in the form of aligned nanorods was first synthesized on an Aucoated Si(111) substrate using a facile method via the galvanicassisted chemical process. The sensing performance of the zinc oxide nanorods was investigated in response to the applied biasing voltage. It was found that the sensitivity, response time and detection limit of the ZnO sensing elements were dramatically improved with increasing bias voltage. A 100% increment in sensing response was achieved for the detection of 2000 ppm hydrogen gas when the bias voltage was increased from −2 to −6 V with 70% reduction in response and recovery times. This remarkable sensing performance is attributed to the reaction of hydrogen with chemisorbed oxygen ions on the surface of the ZnO nanorods that served as the electron donors to increase the sensor conductance. Higher reverse bias voltages sweep the electrons faster across the electrodes. This shortened the response time and, at the same time, depleted the electrons in the sensor elements and weakens oxygen adsorption. The oxygen ions could then be readily removed by hydrogen, leading to a higher sensitivity of the sensors. This, therefore, envisages a way for high-speed hydrogen gas sensing with high detection sensitivities.

Materials and methodology
Au-coated silicon substrate was cleaned using standard procedures by first sonicating in acetone for 15 min followed by isopropyl alcohol for 15 min before rinsing with deionized water (DI) and blown dry with air. The edges of the substrate were then wrapped with Al foils so that the difference in the reduction potentials between the two materials provided the driving force for the formation of the ZnO nanostructures by galvanic displacement reaction. An aqueous solution, containing 25 mM zinc nitrate hexahydrate (Zn(NO 3 ) 2 ·6H 2 O, 98%) and 25 mM hexamethylenetetramine (C 6 H 12 N 4 ), was used as the electrolyte. The solution was maintained at 75°C in a water bath on a hotplate. The substrate was placed with growth surface facing downward in the electrolyte and the growth time for ZnO was 4 h. The ZnOcoated silicon substrate was then rinsed with DI water and blown dry with air. The ZnO nanorod coating was characterized using an X-ray diffractometer (PANalytical X'Pert Pro MPD) with Cu Kα radiation. The morphology of the nanorods was examined using a FESEM (Carl Zeiss GeminiSEM 500).
For gas sensing measurements, the substrate with ZnO nanostructures was mounted on a prefabricated printed circuit board (PCB) with the Au-coated Si substrate as the mechanical contact electrode (figure 1a). Figure 1b shows the schematic diagram of the fabricated sensor chip. Gas sensing measurements were performed with constant gas flow across the sensor chip in a sealed custommade acrylic glass chamber of volume 12 cm 3 (figure 1).  (standard cubic centimetres per minute). Gas sensing data were acquired via a customized Labview program interfaced with a Keithley Source Measure Unit 2602A. A range of bias voltages from −2 to −6 V was applied to the sensors. All the experiments were carried out with the sensor chip first exposed to air to obtain the baseline resistance, followed by exposures to the desired concentrations of hydrogen gas before the air was flushed back to complete a cycle. Time interval for each H 2 flow and air purging event was set to 300 s. The sensitivity (R) of the ZnO sensor towards H 2 gas was defined as the percentage of relative resistance change ([(R g -R 0 )/R 0 ]), using the following equation: where R 0 and R g are the sensor resistances in the absence and presence of H 2 gas, respectively. The response time is defined as the time required for reaching 90% of the total change in the electric resistance at a given H 2 concentration, while the recovery time is defined as the time required for reaching 10% of the original baseline value after the removal of H 2 .

Results and discussion
3.1. Characterization of the ZnO nanorods X-ray diffractions performed on the sample show that ZnO nanorods were grown as pure crystalline phase of wurtzite structure with a preferred growth along (002) (figure 2). Figure 3 shows the FESEM micrographs of ZnO nanorods grown on Au-coated Si substrate. Top view FESEM image of the sample in figure 3a reveals that the Au-coated Si substrate was not fully covered by ZnO nanorods. The nanorods formed numerous clusters of different sizes. On close examination, the nanorods in figure 3b exhibit irregular shapes, diameters ranging from 30 to 240 nm with an average of 103 nm. The cross-sectional view micrograph in figure 3c reveals that the ZnO nanorods are perpendicularly oriented to the substrate with an average length of 1.8 µm. The thickness of the Au coating is approximately 240 nm. As shown in figure 3d, the majority of the smaller ZnO nanorods tend to be cylindrical with sharper tips compared to the larger ones.
In a chemical process, the formation of ZnO starts with the reduction reaction of dissolved oxygen (equation (3.1)) followed by the formation of Zn(OH) 2 (equation (3.2)) which subsequently converts to ZnO via dehydration (equation (3.3)). In this work, the difference in electronegativity between Al and the Au-coated Si, provided the driving force for the electrons from Al to move to the substrate so that the reduction reaction of dissolved oxygen occurred more efficiently. When the concentration of Zn 2+ and OH − exceeded supersaturation, ZnO nuclei formed at the interface between the substrate and the electrolyte solution. The substrate orientation in the electrolyte, on the other hand, played an important role in activating the anisotropy growth of ZnO nuclei along the [001] direction to form orthogonally grown ZnO nanorods. The formation of ZnO nanorods can be described by the following reactions: and     shows the typical I-V characteristics of the ZnO nanorod arrays on Au-coated silicon substrate grown at room temperature. The I-V plot shows the rectifying Schottky behaviour due to the presence of ZnO/Au heterojunction. Reverse bias voltages of −2, −4 and −6 V were used to characterize the hydrogen sensing behaviour of ZnO as forward bias gave lower sensing characteristics when exposed to H 2 gas. Yu et al. [25,30] also reported that faster sensing response was obtained in reverse bias compared to forward bias operation.   . Sensitivity plotted against time when the ZnO sensor was exposed to hydrogen at different concentrations at room temperature at different reverse bias voltages.

Hydrogen gas sensing
increasing reverse bias voltage, while the response and recovery times decrease significantly. Resistance change is still prominently visible at hydrogen concentration as low as 200 ppm. However, the sensitivity observed at the concentrations of 1800 and 1600 ppm appeared to be almost similar. The sensitivity observed at 1000 and 800 ppm as well as sensitivity observed between 600 and 200 ppm (figure 6) are also not significantly differentiated. This could be attributed to the incomplete gas mixing deficiency during the measurement. Figure 7 illustrates the differences in hydrogen sensing behaviours of ZnO at different bias voltages. A 100% increment in the sensor's sensitivity was recorded at 2000 ppm of hydrogen gas when the bias voltage increased from −2 to −6 V. Reduction in the response and recovery times of more than 70% was also recorded (figures 8 and 9). Dependency of the sensing characteristics on bias voltage provides a promising route to achieving high-speed and optimum sensing performance. This could be attributed to the effects of the interactions between hydrogen, chemisorbed oxygen ions and electro-migration in the sensor materials. ZnO is a well-known n-type semiconductor with its electrons contributed by oxygen vacancies and Zn interstitials [31]. When exposed to the atmospheric environment, the electrons from the ZnO conduction band ionized the atmospheric oxygen to produce negative oxygen ions at the surface of the ZnO nanorods:   At higher reverse bias, the released electrons moved at a faster rate due to the greater driving force across the electrodes and thus, decreased the response times. At the same time, reverse bias also depleted the electrons in the nanorods and weakened oxygen adsorption resulting in oxygen ions being easily removed by H 2 leading to an improvement in sensor performance. Recovery time of a gas sensor is determined by the rate at which atmospheric oxygen recombines with the electrons on the ZnO surface. Owing to the relatively low number of electrons available at high reverse bias, the recovery time is shorter compared to that at low reverse bias. Figure 10 shows the schematics of the proposed sensing mechanisms involved.
A comparison of sensitivity, response time and recovery time for hydrogen gas sensor in this work (    sensor is also higher when compared to others that used ZnO nanorods as the sensing materials except for Hassan et al. [33]. Despite that, their ZnO sensor showed poorer recovery and response times. A high sensitivity sensor is suitable for quantitative measurement of gas concentrations, whereas a fast response sensor, as demonstrated by the sensor characteristics in this work, will be suitable for fast sensing and detection of gas leakage.

Conclusion
In this paper, we have investigated the room temperature hydrogen gas sensing behaviour of ZnO nanorods grown on Au-coated Si(111) substrate by the galvanic-assisted chemical process. It was found that the hydrogen sensing behaviour and performance of the sensor depended strongly on the applied bias voltage. The sensitivity, response time and recovery time could be improved many folds of magnitude by using large bias voltages. This paves the way for the fabrication of hydrogen gas sensors with good detection sensitivities and low detection limits for high-speed sensing in real time.