Tungsten oxide thin film for room temperature nitrogen dioxide gas sensing.

. Tungsten oxide (WO 3 ) thin films for gas sensing have been successfully deposited using reactive direct current (DC) magnetron sputtering at different deposition temperatures (300 ºC, 400 ºC and 500 ºC). The structural, morphological properties, thickness and composition have been investigated using X-ray diffraction (XRD), scanning electron microscopy (SEM) and Rutherford backscattering spectrometry (RBS) techniques. To investigate the effect of deposition temperature on the gas sensing properties of deposited thin films on alumina substrates, was conducted using the Kenosistec gas sensing unit. WO 3 thin film deposited at 500 ºC exhibited a higher response when sensing Nitrogen dioxide (NO 2 ) at room temperature as compared to the thin films prepared at 300 ºC and 400 ºC, respectively. However, as deposited WO 3 thin films exhibited low sensitivity when sensing reducing gases such as hydrogen (H 2 ) and ammonia (NH 3 ), which was an indication of good selectivity properties of WO 3 related sensors.


Introduction
There is a growing demand of metal oxide thin films for numerous important technological applications such as gas sensing, solar cells, solar absorbers, etc. The fast-growing industries and increase in exhaust emissions of gases from industries has put a large demand on the sensitive and selective detection of hazardous gases for environmental pollution monitoring, process control and safety for human health [1,2,3]. This has attributed to the fabrication of gas sensor devices [1,2,3,4,5]. NO2 is one of the most potentially toxic gas which can lead to respiratory problems to humans. This gas leads to the formation of acid rains when excreted to air in its high concentrations above 0.65 ppm [1]. Therefore, the ability to accurately monitor or detect NO2 concentrations in air is very crucial. A variety of metal oxide semiconductor nanostructures, such as ZnO [6], NiO-Nb2O5 [7], SnO [8], CuO [9], and VO2 [10] and WO3 [1] have been extensively investigated as gas sensing materials, however their responses towards NO2 detection at room temperature is still a challenge. Many previously reported metal oxide gas sensors are unable to detect NO2 gas at low temperatures, hence an elevated temperatures e.g., 250 ºC was used when detecting NO2. This is not good since high temperatures has resulted in high power consumption, which also reducing the gas sensor response of. The high concentration of NO2 gas has been used previously when detecting NO2, however a gas sensor that will detect and monitor NO2 gas at room temperature and at low concentration is still needed to be fabricated since it is important to also detect NO2 gas at low concentration. WO3 is a promising material for detection of NO2 [11]. Therefore, in this study WO3 was prepare using different deposition temperatures (300, 400, and 500 ͦ C) to achieve different structures of WO3 which play an important role in its gas sensing. Tungsten oxide nanostructures prepared by various methods are good candidates for detecting NO2 because of their high crystallinity and large specific surface areas [1,11,12]. However, WO3 based sensors require high operating temperatures (above 200 ºC) during gas detection because of high activation energy of reaction with gas molecules [12,13]. The high operating temperature during gas sensing is unfavorable for low power consumption and device integration [1]. Various attempt to reduce the operating temperature of WO3 nanostructures based sensors such as WO3 combine with other materials to form heterostructures or composite sensors have been performed [1,12]. Tungsten oxide (WO3) is an n-type metal oxide semiconducting material with a band gap of about 2.7 eV [3]. WO3 is considered to be a highly promising material for a broad range of applications such as in electronic devices, super-capacitors, optical aspects, gas sensors, solar energy, photocatalyst and biosensors [4,5,6,14,15,16,17]. This is because of its unique structural and physical properties, non-toxicity, and high chemical stability [4,5,6,14,15]. WO3 exhibit various phases at temperature for example, it is tetragonal at temperatures above 740 ºC, orthorhombic between 330 to 740 ºC, monoclinic in the range of 17 ºC to 330 ºC and triclinic between -50 ºC and 17 ºC [5,6,14,15,16]. The most common structure of WO3 is monoclinic [5]. Some previous studies have reported that n-type metal oxide nanostructures (e.g. WO3) are promising materials for gas sensing since they have good electrical properties and excellent stability in various environments [6][7][8][9][10][11][12][13][14][15]. In this work, WO3 thin films was deposited using DC magnetron sputtering for room temperature gas sensing of NO2 gas. Different deposition temperatures was used to obtain different morphological structures which results in different gas sensing performance, this was done for the purpose of achieving a suitable WO3 based sensor which is highly sensitive to low concentration of NO2 and at room temperature. The DC sputtering deposition technique was used because of its fast deposition rate and good adhesion of clean, thin films on the substrates having no unwanted impurities [18].

Sample preparation
WO3 thin films were synthesized on alumina substrates using a DC reactive magnetron sputtering technique. The 99.9% pure Tungsten target as purchased from AJA was sputtered at a pressure of 3 × 10 −3 Torr for 50 minutes. Optimum deposition conditions for WO3 thin films were reached through by varying the deposition power between 100 W to 200 W in steps of 50 W and the oxygen flow between 2 to 8 sccm. RBS and XRD characterizations techniques revealed that the best deposition parameters to obtain well crystalline stoichiometric WO3 thin films with high purity are a deposition power of 150W, an argon gas flow rate of 8 sccm and oxygen gas flow rate of 6 sccm, at a deposition pressure of 3x10 -3 Torr for a deposition time of 50 minutes. These optimized deposition parameters were then used to synthesize different samples of WO3 by varying the deposition temperature between 300 o C to 500 o C. Subsequent in attaining WO3 thin films at 300 o C, 400 o C and 500 o C. Figure 1 (a-c) shows the images of SEM surface morphologies of WO3 thin films that were deposited on silicon substrates using DC reactive magnetron sputtering at different deposition temperatures. Figure 1 (a), (b) and (c) are surface morphologies of WO3 films prepared at 300, 400 and 500 ℃ respectively. During the deposition of WO3 films, all other parameters were kept constant except the deposition temperature which was varied (300, 400 and 500 ℃). Therefore, it can be concluded that as the deposition temperature changes the surface morphology of WO3 film also changes. At a deposition temperature of 300 ℃, the WO3 surface morphology exhibits an evenly distributed small nano-spherical particles. At 400 ℃, a morphological structure of none-spherical granules was observed, and WO3 prepared at 500 ℃, exhibited the morphology of cornflakes like in shape. The different morphological structures of WO3 samples resulted on different gas sensing responses of WO3, as it is explained under gas sensing results section of this document.  Figure 2 (a-c) shows the XRD (Cukα 1, λ = 0.15406 nm) spectra of the WO3 thin films that were prepared by DC magnetron sputtering at different temperatures (300, 400 and 500 ℃). The Xray diffraction patterns indicated by letters a, b and c are of the WO3 thin films prepared at 300, 400 and 500 ℃ respectively and they show that WO3 thin films obtained are highly crystalline. An increase of the crystalline quality can be envisaged by pattern of XRD spectra. The significant increase in the crystallinity quality could be attributed to the increase of deposition temperature.

Results and discussion
The XRD pattern of WO3 samples correspond to PDF card (01-071-0131), WO3 deposited at different temperatures were identified to have an orthorhombic structure with units' cell lattice parameters of a = 7.341 Å, b = 7.570 Å, c = 7.754 Å and angle of 90º (for α, β and γ) and space group Pmnb (62). The strongest diffraction peak for all WO3 films deposited appears at 2 = 23º in (002) plane, indicating a fine preferential growth in the (002) direction. It was found that the deposition temperature also affects the crystallite size of the WO3 nanoparticles. The crystallite size was calculated using Scherrer's formula [12] from the half width of the (002) peak for all WO3 films and the results are tabulated in Table 1. It was observed that as the deposition temperature increases from 300 ºC to 500 ºC, the crystallite size also increased from 25 nm to 34 nm.
The samples were further characterized and analyzed for composition and thickness using RBS technique which was carried out using a beam of 3.6 MeV alpha particles ( 4 He ++ ). In Figure 3, 4 and 5 are the RBS spectra of 3.6 MeV ( 4 He ++ ) which was incident on the samples. It was observed from the RBS as indicated on Table 2 that the composition found is indeed WO3 compound which agree with the XRD results. It was also observed from the RBS that the thickness of the samples deposited at different temperatures (300 ºC, 400 ºC and 500 ºC) were not the same. The thickness seemed to increase with an increase in the deposition temperature. The sample deposited at 300 ºC was found to have thickness of 120 nm, the sample deposited at 400 ºC was found to have a thickness of 360 nm. However, when compared with the sample deposited at 500 ºC, not much difference in thickness was observed since this sample was found to have a thickness of 370 nm. This increase in thickness might perhaps attributed by the increase in grain size or crystallite size of the samples as it was observed from the respective results of XRD. It should also be noted from SEM results that the morphology of the samples was not the same due to the change in deposition temperatures.    Figure 6 shows the WO3 based sensor film deposited on the alumina strip with platinum electrodes. The alumina strip was used because of its good electrodes for current flow and its melting point is high (melting point is 1768 ºC) [12]. Some previous studies [19] have shown WO3 as a good sensor for NO2 gas detection. However, in those studies high operating temperatures (above 200 ºC) were used during NO2 detection to try to enhance the sensitivity of WO3 based sensor and it has also been reported that WO3 is not a good sensor at room temperature since it has high activation energy. Therefore, high operating temperature is required to enhance the sensitivity of WO3 [20,21]. However, a good sensor of NO2 which will be able to operate at room temperature is also needed since NO2 can also be found at room temperature environment. WO3 based sensors which were prepared at different deposition temperatures were each tested for their properties at room temperature on a gas sensing machine called Kenosistec at University of Zululand, but before testing the response of the sensors towards NO2 gas, the sensors were first exposed into different relative humidity (RH%) to check their optimum humidity at room temperature. Figure 7 represents the response of the sensors towards 20 ppm NO2 at room temperature at various relative humidity. It was observed that 70%RH is the optimum humidity for WO3 based sensors and therefore the gas sensing properties of WO3 based sensors were tested at 70%RH. It can be noted from Figure 7 that the humidity does affect the sensor response of WO3 as NO2 gas of 20 ppm is introduced to the surface of WO3. It can also be seen that from 0%RH to 30%RH the sensor does respond towards the incoming NO2 gas, but it does not show even a small recovery. However, the sensor started to show some recovery when the humidity was raised to 70%RH. At 90%RH the sensor did show some recovery however the response was negatively affected as it seems to drop when the humidity is above 70%RH, thus it can be concluded that the optimum humidity for WO3 based sensor is 70%RH since the response and recovery of the sensors towards NO2 gas was good.  Fig. 7. The sensor response curve of WO3 based sensor towards 20 ppm concentration of NO2 gas at room temperature at different relative humidity.
When WO3 based sensor film was exposed to the NO2 gas, the resistance of the sensor increased to a certain value and then decreased to the original value once the NO2 is switched off. This behavior is normal as reported in previous studies [11,12,17,19] since WO3 is the n-type metal oxide semiconductor and NO2 is the oxidizing gas. This type of behavior corresponds with gas sensing response of Rg/Ra where Ra is the resistance of the sensor, when no gas present on the sensor, and Rg is the resistance of the sensor when they are a gas introduced on the sensor. It shows the chemisorption of NO2 to the WO3 surface. Figure 8 shows the sensor response curve of the WO3 metal oxide sensors deposited at 300 ºC, 400 ºC and 500 ºC. These sensors were exposed to different NO2 concentrations at room temperature and at 70%RH. The NO2 gas was switched on for 1200 seconds to check how fast the sensor is responding to the gas and switched off for 600 seconds to check how fast it is going to recover. It can be seen from Figure 8 that the sensors responded immediately to NO2 gas at all concentrations, but they took some time to recover, and all of them appear not to be able to recover completely to their original resistance. However, the response and recovery time of the WO3 based sensor appear to be slowly improving as the temperature was raised to 500 ºC and this can clearly be seen from Figure 10 where response and recovery time of each the sensor was calculated. Figure 9 shows the response curve of WO3 based sensor deposited at 500 ºC. It can be noted from Figure 9 that when the relative humidity was still at 70%RH the sensor response of the sensor was good, but when it was raised to 90%RH during the detection of 150 ppm NO2, the sensor response behavior changed, and this change shows that the sensor does not perform well at 90%RH. Figure 8 and Figure 9 shows that the initial resistance of WO3 deposited at 500 ºC changed from starting at around 112 kΩ to start around 162 kΩ. This was attributed by the fluctuating factors during gas sensing e.g., humidity which was 70%RH±2%RH and the room temperature which was 17.6 ºC± 0.4 ºC.  The response time and recovery time are one of the most important properties of the gas sensors because they indicate how fast the gas sensor will respond to the target gas and how fast does it recover to its initial resistance when the target gas has been switched on and thereafter switched off during testing in gas sensor machine. This is because the resistance of the sensor changes when it is exposed to the target gas and when it changes it shows that it is able to detect or sense that particular gas exposed to [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17]. In Figure 10 the response and recovery time for gas sensors deposited at 300 ºC, 400 ºC and 500 ºC were calculated, the sensors were exposed to 80 ppm concentration of NO2 at room temperature and 70%RH. It can be noted from Figure 10 that as the temperature which was used to prepare or deposit the WO3 increased from 300 ºC to 500 ºC, the response and recovery time of WO3 based sensor was improving. It was observed that the response time decreased from 142.2 to 108 seconds, and the recovery time decreased from 588.6 to 565.8 seconds. Therefore, it can be concluded from Figure 10 that WO3 based sensor deposited at 500 ºC is a good sensor compared to those deposited at 300 ºC and 400 ºC, this also agree when comparing the sensitivity of the three sensors in Figure 11. The sensitivity is also the most important characteristic of the gas sensor because it tells us the rate at which the gas sensor responds towards the target gas, as the concentration of the gas increases [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17]22,23]. The effectiveness of the gas sensor is determined by its sensitivity. The sensitivity (ΔS/ΔC) is determined from the slope of the linear curve of the sensor response (S) vs concentration (C). In Figure 11, the sensor response with concentration of NO2 gas were plotted for different WO3 based sensors deposited at different temperatures. It can be noted from Figure  12 that the sensitivity of WO3 based sensor deposited at 500 ºC is a bit higher compared to other two sensors. The sensitivity was calculated from 80 ppm to 150 ppm for all three sensors. This is also true when looking at other concentrations (20 ppm to 40 ppm, 40 ppm to 80 ppm, and 150 ppm to 300 ppm) that the sensitivity of WO3 deposited at 500 ºC is high compared to those deposited at 300 ºC and 400 ºC. Since WO3 based sensor deposited at 500 ºC was the one to show good sensitivity, it was further tested to the varying NH3 gas concentrations and to hydrogen gas at the same gas sensing environment (room temperature and 70%RH). It can be observed from Figure 12 that WO3 based sensor did able to respond towards NH3 gas, however the behavior of this sensing was not clear and the response towards this gas was not good compared to the response towards NO2. In Figure 13 it can be observed also that WO3 was not able to detect hydrogen gas even though the hydrogen concentrations were varied, this was perhaps occurring because NH3 and H2 are both reducing gases, hence WO3 was selective to NO2 gas.

Conclusion
WO3 nanostructured thin films were successfully deposited on alumina strips substrates using DC magnetron sputtering at different deposition temperature (300, 400 and 500 ºC). WO3 thin films were characterized using XRD, SEM and RBS. From the XRD it was found that the samples (WO3 thin films) were highly crystalline and that as the deposition temperature increases, the samples become more crystalline, and this was observed since the diffraction peak intensity of the samples increased with the increase in deposition temperature. The increase in crystallite size of the samples with an increase in deposition temperature was also observed. From the SEM it was also found that the deposition temperature has more effect on the sample morphology since the morphology of the samples was completely different from each other. RBS confirmed that thin films of WO3 were deposited on silicon and alumina strip substrates as indicated by the sample composition obtained from RBS which also agree with the XRD results. When increasing the deposition temperatures thickness of the samples also increases, this was thought as being attributed by the increase in grain size or crystallite size of the samples as this was observed from XRD results. The samples were tested on a gas-sensing machine (Kenosistec) for their gas sensing performance. It was found that the sample deposited at 500 ºC has a good gas sensing performance when compared to those that were deposited at 300 ºC and 400 ºC, and that it can be effectively used to sense NO2 gas at room temperatures under the humidity of 70%RH. The NO2 gas sensor based on WO3 film deposited at 500 ºC showed the n-type behavior of sensing since the exposure to NO2 gas resulted in the increase in its resistance, this is a normal behavior since NO2 is an oxidizing gas and WO3 is an n-type metal oxide. The WO3 based sensor showed the poor response towards NH3 and H2 gases, which shows that WO3 gas sensor is a selective to NO2.