Study on the Ozone Gas Sensing Properties of rf-Sputtered Al-Doped NiO Films

Featured Application: Authors are encouraged to provide a concise description of the speciﬁc application or a potential application of the work. This section is not mandatory. Abstract: Al-doped NiO (NiO:Al) has attracted the interest of researchers due to its excellent optical and electrical properties. In this work, NiO:Al ﬁlms were deposited on glass substrates by the radio frequencies (rf) sputtering technique at room temperature and they were tested against ozone gas. The Oxygen content in (Ar-O 2 ) plasma was varied from 2% to 4% in order to examine its effect on the gas sensing performance of the ﬁlms. The thickness of the ﬁlms was between 160.3 nm and 167.5 nm, while the Al content was found to be between 5.3 at% and 6.7 at%, depending on the oxygen content in plasma. It was found that NiO:Al ﬁlms grown with 4% O 2 in plasma were able to detect 60 ppb of ozone with a sensitivity of 3.18% at room temperature, while the detection limit was further decreased to 10 ppb, with a sensitivity of 2.54%, at 80 ◦ C, which was the optimum operating temperature for these ﬁlms. In addition, the ﬁlms prepared in 4% O 2 in plasma had lower response and recovery time compared to those grown with lower O 2 content in plasma. Finally, the role of the operating temperature on the gas sensing properties of the NiO:Al ﬁlms was investigated.


Introduction
Ozone (O 3 ) is a well-known harmful gas existing in the atmosphere as a product of photochemical reactions of Nitrogen dioxide (NO 2 ) and/or Volatile Organic Compounds (VOCs), which are very common environmental pollutants that come from industrial activity, cars, etc. [1]. As a result, over 80% of people live in cities where air pollution is higher than the safety limits of the World Health Organization (WHO) [2]. Taking into account that ozone is associated with various respiratory symptoms, including dyspnea, upper airway irritation, coughing, and chest tightness [1], the need of its detection in outdoor as well as indoor environments becomes of paramount importance.
Various kinds of materials, such as Metal Oxide Semiconductors (MOS) [3][4][5], inorganic perovskites [6], as well as hybrid perovskites [7] have been examined as gas sensing elements for ozone detection during the last decades. Among them, MOS are by far the most well studied materials for ozone gas sensing applications due to their excellent electrical and optical properties as well as the fact that they can be grown by a number of methods even at large scale [8,9]. Nickel oxide (NiO) is a p-type metal oxide semiconductor with a wide energy band gap of 3.4-3.8 eV that has interesting optoelectronic properties [10]. Thus, it can be used in UV photo-detectors [11] as an active material in photovoltaics, such as perovskite or dye-sensitized solar cells [12][13][14][15][16], as well as in electrochromic [17], thermoelectric [18], and gas sensing devices for the detection of gases, such as H 2 [19,20], CH 4 [21], NO 2 [22], ethanol [23], O 3 [24], etc. Furthermore, the introduction of Aluminum (Al) atoms in the NiO structure seems to enhance its electrical as well as optical properties [25], while not affecting its gas sensing performance. As a result, Al-doped NiO films have been already tested against NO 2 [26,27], ethanol [28], CO [29], H 2 , and CH 4 [30]. However, NiO:Al films have not been examined for ozone detection, according to our literature research.
In this work, rf-sputtered Al-doped NiO films were tested, for first time, against ozone. The effect of Oxygen content in Ar-O 2 plasma as well as the operating temperature on the gas sensing performance were studied. It was found that NiO:Al films have the ability to detect ozone even at room temperature, while those prepared with 4% O 2 in plasma successfully detected ozone at an ultra-low concentration of 10 ppb at 80 • C. The latter is one of the lower operating temperatures for metal oxide gas sensors.

Deposition Conditions
A Nordiko RFG2500 rf sputtering system was employed to grow NiO:Al films using a Ni metal target and Ar-O 2 air mixture in plasma. Two (2) Al pellets were placed on the Ni target surface for Al-doping. Both pressure and sputtering power were kept constant during deposition, being 5 mTorr and 300 W, respectively. The films were deposited on three different substrates, namely the commercial (Metrohm/DropSens) InterDigitated Transducers (glass substrate, Pt electrodes, bands/gaps = 5 µm) for sensing measurements, a piece of silicon wafer for structural characterization, and microscope glass for optical measurements. All depositions were done at room temperature (RT). The O 2 content in plasma was varied, being 2.0%, 2.8%, and 4.0% in order to examine its effect on the gas sensing characteristics of the films. After deposition, the films underwent thermal annealing at 400 • C under air for 24 h, in order to improve both their crystallinity and optical properties.

Characterization
The structure of the Al-doped NiO films was examined by the X-ray Diffraction (XRD) technique, using a PANalyticalEmpyrean diffractometer equipped with Cu-LFF as an X-Ray source, at λ = 0.15406 nm. The measurement mode was 2θ/θ varying from 20 • to 90 • , with a step of 0.013 • /s. Using the XRD pattern the crystallite size (D) of the films was calculated according to Scherrer's Equation (1) where λ is the X-ray wavelength which equals to 0.154 nm. B is the Full Width at Half Maximum (FWHM) of the corresponding peak at an angle 2θ. Moreover, the lattice constant (a 0 ) for the cubic structure of NiO was calculated according to Equation (2) a = (h 2 + k 2 + 2 ) 1/2 /d (2) where d is the distance between the adjacent planes in the set (hk ). Field Emission Surface Electron Microscopy (FESEM) was used to investigate the surface morphology of the films using a Hitachi S570 microscope, equipped with an Energy Dispersive X-ray (EDX) spectrometry system, which was employed to determine the Al at% of the NiO:Al films. Optical properties of the films were studied with a Perkin Elmer Lambda 950 UV/Vis/NIR spectrophotometer in the wavelength range of 250-2500 nm. The transmittance spectra and the corresponding Tauc plot were used in order to calculate the optical energy bandgap of the films assuming that direct transitions are permitted according to Equation (3) where α is the absorption coefficient, hv the photon energy, A is a constant and E g the optical energy bandgap. The thickness of the films was measured by using a Veeco Dektak 150 stylus profilometer.

Gas Sensing
In order to investigate the gas sensing performance in ozone, the NiO:Al films deposited on IDTs were placed in a homemade stainless steel test chamber. A mechanical pump was used to initially evacuate the chamber, while an FP-400 temperature controller was employed to regulate the operating temperature. Ozone of different concentrations was produced by an ozone analyzer (Thermo, Model 49i, Thermofisher scientific, Waltham, MA, USA). A Keithley 6517A electrometer was used to apply a constant voltage on the films and measure the electrical current variations upon the interaction between the film and the target gas. The chamber was initially evacuated at a pressure of 1 mbar. Prior to the ozone exposure, the films were treated with synthetic (dry) air for 45 min in order to be stabilized in the testing environment. After that, the films were exposed to ozone of specific concentrations for a constant time duration in which the electrical current was increased, while the recovery of the films in its initial state was done by exposing them in synthetic air for a specific time period. Both ozone and synthetic air were introduced in the chamber with a flow of 200 sccm (standard cubic centimeters per minute) regulated by mass flow controllers, while the pressure in the chamber was kept constant of 700 mbar. The whole procedure was monitored by a PC through an appropriate LabVIEW program. The sensitivity (S) of the sensor is defined by the following Equation (4) S(%) = [(I gas − I air )/I air ] · 100% (4) where I gas is the maximum value of electrical current in the presence of ozone, while I air is the minimum value of electrical current in the presence of synthetic air (absence of ozone), as depicted in Figure 1. Moreover, the response time (t resp ) is defined as the time that is needed for the electrical current to become equal to the 90% of the value of I gas in the presence of ozone, while the recovery time (t rec ) is defined as the required time for the electrical current to be equal to 10% of the value of I gas in the absence of ozone (in the presence of synthetic air).
where α is the absorption coefficient, hv the photon energy, A is a constant and Eg the optical energy bandgap. The thickness of the films was measured by using a Veeco Dektak 150 stylus profilometer.

Gas sensing
In order to investigate the gas sensing performance in ozone, the NiO:Al films deposited on IDTs were placed in a homemade stainless steel test chamber. A mechanical pump was used to initially evacuate the chamber, while an FP-400 temperature controller was employed to regulate the operating temperature. Ozone of different concentrations was produced by an ozone analyzer (Thermo, Model 49i, Thermofisher scientific, Waltham, MA, USA). A Keithley 6517A electrometer was used to apply a constant voltage on the films and measure the electrical current variations upon the interaction between the film and the target gas. The chamber was initially evacuated at a pressure of 1 mbar. Prior to the ozone exposure, the films were treated with synthetic (dry) air for 45 min in order to be stabilized in the testing environment. After that, the films were exposed to ozone of specific concentrations for a constant time duration in which the electrical current was increased, while the recovery of the films in its initial state was done by exposing them in synthetic air for a specific time period. Both ozone and synthetic air were introduced in the chamber with a flow of 200 sccm (standard cubic centimeters per minute) regulated by mass flow controllers, while the pressure in the chamber was kept constant of 700 mbar. The whole procedure was monitored by a PC through an appropriate LabVIEW program. The sensitivity (S) of the sensor is defined by the following Equation (4) S where Igas is the maximum value of electrical current in the presence of ozone, while Iair is the minimum value of electrical current in the presence of synthetic air (absence of ozone), as depicted in Figure 1. Moreover, the response time (tresp) is defined as the time that is needed for the electrical current to become equal to the 90% of the value of Igas in the presence of ozone, while the recovery time (trec) is defined as the required time for the electrical current to be equal to 10% of the value of Igas in the absence of ozone (in the presence of synthetic air). All measurements were taken under the room relative humidity (RH) which was about 60%.  All measurements were taken under the room relative humidity (RH) which was about 60%.

Characterization
The XRD patterns of rf-sputtered Al-doped NiO films deposited on Si substrates at room temperature with different O 2 content in plasma that underwent thermal annealing at 400 • C for 24 h are presented in Figure 2. It can be seen that all films had a polycrystalline structure with a peak of high intensity at 2θ = 43.3 • corresponding to a reflection from the (200) planes and two low intensity peaks at 2θ = 36.5 • and 62.9 • corresponding to reflections from the (111) and (220) planes, respectively, according to JCPDS Card No. 047-1049. No peaks corresponding to other materials were found, indicating that all Al atoms were introduced in the NiO structure. Using Equation (1) for the diffraction peak with the highest intensity (200), the crystallite size of the films was calculated (Table 1) and found to be between 8.68 nm for films prepared with 2% and 2.8% O 2 in plasma and 7.89 nm for those prepared with 4% O 2 in plasma. In the case of the films prepared with 4% O 2 , the decrease of the crystallite size may lead to a further increase of defect states inside the crystal structure, which can also be observed through the intensity of the main peak. In addition, the lattice constant (a 0 ) of the films was calculated from Equation (2) and was varied between 0.417 nm and 0.421 nm (

Characterization
The XRD patterns of rf-sputtered Al-doped NiO films deposited on Si substrates at room temperature with different O2 content in plasma that underwent thermal annealing at 400 °C for 24 h are presented in Figure 2. It can be seen that all films had a polycrystalline structure with a peak of high intensity at 2θ = 43.3° corresponding to a reflection from the (200) planes and two low intensity peaks at 2θ = 36.5° and 62.9° corresponding to reflections from the (111) and (220) planes, respectively, according to JCPDS Card No. 047-1049. No peaks corresponding to other materials were found, indicating that all Al atoms were introduced in the NiO structure. Using Equation (1) for the diffraction peak with the highest intensity (200), the crystallite size of the films was calculated (Table 1) and found to be between 8.68 nm for films prepared with 2% and 2.8% O2 in plasma and 7.89 nm for those prepared with 4% O2 in plasma. In the case of the films prepared with 4% O2, the decrease of the crystallite size may lead to a further increase of defect states inside the crystal structure, which can also be observed through the intensity of the main peak. In addition, the lattice constant (a0) of the films was calculated from Equation (2) and was varied between 0.417 nm and 0.421 nm (Table 1). Both the crystallite size and lattice constant are in agreement with the values reported in the literature for Al-doped NiO films prepared by various techniques [25,31]. The surface morphology of the NiO:Al films is presented in Figure 3a-c. It can be seen that all films had a smooth and uniform surface consisting of grains with a diameter of around 40 nm The increased defects on the surface of the film prepared with 4% O2 can be also be observed.  The surface morphology of the NiO:Al films is presented in Figure 3a-c. It can be seen that all films had a smooth and uniform surface consisting of grains with a diameter of around 40 nm The increased defects on the surface of the film prepared with 4% O 2 can be also be observed. The transmittance spectra of the Al-doped NiO films are shown in Figure 4. All films were highly transparent (>70%) in the visible region as well as in the near infrared one. Using Equation (3) and the Tauc plot (inset of Figure 4), the optical energy bandgap was calculated and found to be around 3.66 eV (Table 1)   The transmittance spectra of the Al-doped NiO films are shown in Figure 4. All films were highly transparent (>70%) in the visible region as well as in the near infrared one. Using Equation (3) and the Tauc plot (inset of Figure 4), the optical energy bandgap was calculated and found to be around 3.66 eV (Table 1) The transmittance spectra of the Al-doped NiO films are shown in Figure 4. All films were highly transparent (>70%) in the visible region as well as in the near infrared one. Using Equation (3) and the Tauc plot (inset of Figure 4), the optical energy bandgap was calculated and found to be around 3.66 eV (Table 1)

Gas Sensing
The gas sensing performance of rf-sputtered Al-doped NiO films was investigated by exposing them to ozone gas with a concentration varying from 10 ppb to 3000 ppb, at operating temperatures between RT and 150 • C. The duration in ozone exposure was 5 min for films prepared with 4% O 2 in plasma and 10 min for the films prepared with 2% and 2.8% O 2 in plasma, while the recovery of the sensors was carried out by exposing them to synthetic air for 10 min and 15 min, respectively. The applied voltage was 10 V when the operating temperature equals to RT, while for the rest of the operating temperatures the applied voltage was 1 V.
A typical measurement of electrical current variation for different O 3 concentrations is presented in Figure 5. In the presence of ozone the electrical current of the film is increased reaching a maximum value, while in the absence of the target gas the electrical current of the film is decreased, as a result of its interaction with synthetic air, reaching its initial value. Moreover, it is noted that the sensor cannot be fully recovered to its initial electrical current value, especially at high concentrations (>600 ppb), probably due to residual oxidation. The gas sensing mechanism for Al-doped NiO films is the typical one for a p-type semiconductor. Thus, upon the exposure to synthetic air, Oxygen species (O, O 2− , etc.) are formed as a result of the O 2 adsorption on the films' surface. This is described through the following Equations (5)-(7).
where the kind of oxygen ions is dependent on the operating temperature (T oper ), being O − 2 for T oper < 100 • C, O − for 100 • C < T oper < 300 • C, and O 2− for T oper > 300 • C. The Oxygen species take out electrons from NiO, increasing the number of holes, leading to an increase of electrical conductivity, thus a depletion layer is formed. As a result, the height of the potential barrier at the grain boundaries is decreased and the electrical current is slightly increased. When NiO:Al films interact with ozone, which is a strongly oxidizing gas, the ozone molecules are adsorbed on the films' surface trapping the free electrons, resulting in an increase of the hole concentration as well as in a decrease of the thickness of the depletion layer. Thus, an extra increase of the electrical current is occurred [32,33]. In Figure 6a-c the variation in sensitivity of the NiO:Al films grown with different O2 content in plasma are shown. This was calculated by Equation (4) as a function of ozone concentration for various operating temperatures and plotted in logarithmic scale. In addition, a linear fitting has been applied confirming that the sensitivity is increased with O3 In Figure 6a-c the variation in sensitivity of the NiO:Al films grown with different O 2 content in plasma are shown. This was calculated by Equation (4) as a function of ozone concentration for various operating temperatures and plotted in logarithmic scale. In addition, a linear fitting has been applied confirming that the sensitivity is increased with O 3 concentration, independent of the Oxygen content in the plasma or the operating temperature. From Figure 6, it becomes clear that all films can detect the ozone gas even at an ultra-low concentration of 10 ppb. Specifically, the films prepared with 2% O 2 in plasma had a lowest detection limit of around 200 ppb with a sensitivity 1.88%, 11.31% and 38.66% at RT, 110 • C and 150 • C, respectively, while at 80 • C the lowest detection limit was 400 ppb with a sensitivity of 5.17%. Moreover, films synthesized with 2.8% O 2 in plasma had a lowest detection limit of 400 ppb [27] (sensitivity 0.84%), 200 ppb (sensitivity 6.25%), 200 ppb (sensitivity 6.94%) and 60 ppb (sensitivity 6.22%) when the operating temperature equals to RT, 80 • C, 110 • C, and 150 • C, respectively. Finally, the NiO:Al prepared with 4% O 2 in plasma appeared to have a lowest detection limit of 60 ppb (sensitivity 3.18%), 10 ppb (sensitivity 2.54%), 60 ppb (sensitivity 2.44%) and 400 ppb (sensitivity 9.02%) when the operating temperature equals to RT, 80 • C, 110 • C, and 150 • C. The detection limit of 10 ppb at 80 • C is one order of magnitude less than the one that has been reported [32] for Co 3 O 4 nanobricks at almost the same working temperature (85 • C). The above mentioned values for all NiO:Al films are presented in Table 2.
Furthermore, in Figure 7 the effect of the operating temperature on gas sensing performance of the rf-sputtered NiO:Al films is shown. The films tested against 400 ppb ozone at various operating temperatures, namely RT, 80 • C, 110 • C, and 150 • C. It can be seen that films prepared with 4% O 2 in plasma appeared to have the best sensitivity at 80 • C, while the rest of them showed a maximum sensitivity at 150 • C. This can be attributed to both less Al content and smaller crystallite size of the films grown with 4% O 2 in plasma compared to the others, which also leads to a further decrease of the defect states that can reduce the adsorption energy of the O 3 gas molecules [34]. In addition, the operating temperature of 80 • C is one of the lowest that has been reported for ultra-low ozone concentration (10 ppb) [4,32].
Finally, the response and recovery time of the NiO:Al films for each operating temperature were calculated and presented in Figure 8a-c. Films grown with 2% and 2.8% O 2 in plasma had a response time between 400 s and 800 s, depending on both operating temperature and ozone concentration, while for those prepared with 4% O 2 in plasma the response time was ranging from 150 s to 500 s. In the same way, the recovery time for the former films was between 400 s and 900 s, while it was varied from 200 s to 600 s for films grown with 4% O 2 in plasma. It should be noticed that the Al: NiO films grown with 4% O 2 in plasma showed the best response and the recovery time equals to 189.6 s and 243.6 s, respectively, against 10 ppb ozone at an operating temperature of 80 • C. The response time values of Al: NiO films are higher than those in other reported works, while the recovery time values are comparable to those reported in the literature [4,32,[34][35][36][37][38]. However, most of these sensors work at higher temperatures and/or have a higher concentration detection limit than those presented in this work.  Furthermore, in Figure 7 the effect of the operating temperature on ga formance of the rf-sputtered NiO:Al films is shown. The films tested ag ozone at various operating temperatures, namely RT, 80 °C, 110 °C, and 15 seen that films prepared with 4% O2 in plasma appeared to have the best s °C, while the rest of them showed a maximum sensitivity at 150 °C. This ca to both less Al content and smaller crystallite size of the films grown with 4% compared to the others, which also leads to a further decrease of the defect reduce the adsorption energy of the O3 gas molecules [34]. In addition, the o perature of 80 °C is one of the lowest that has been reported for ultra-low tration (10 ppb) [4,32]. Finally, the response and recovery time of the NiO:Al films for each o perature were calculated and presented in Figure 8a-c. Films grown with 2 in plasma had a response time between 400 s and 800 s, depending on b temperature and ozone concentration, while for those prepared with 4% O response time was ranging from 150 s to 500 s. In the same way, the recove former films was between 400 s and 900 s, while it was varied from 200 s to grown with 4% O2 in plasma. It should be noticed that the Al: NiO films g O2 in plasma showed the best response and the recovery time equals to 18 s, respectively, against 10 ppb ozone at an operating temperature of 80 °C time values of Al: NiO films are higher than those in other reported works covery time values are comparable to those reported in the literature [4,32 ever, most of these sensors work at higher temperatures and/or have a hig tion detection limit than those presented in this work.

Conclusions
Al-doped NiO films were grown by the rf sputtering technique with different oxygen contents in Ar-O 2 plasma, i.e., 2%, 2.8%, and 4%, at room temperature. All films were polycrystalline as revealed by the XRD technique, while their surface was smooth as was confirmed by Scanning Electron Microscopy. Moreover, all films were highly transparent, showing a transmittance of more than 70% in the visible spectrum. The gas sensing performance of the films against ozone was studied at different operating temperatures, namely 25 • C, 80 • C, 110 • C, and 150 • C. All films were sensitive to ozone of different concentrations. Specifically, films grown with 4% O 2 in plasma were able to detect ozone at an ultra-low concentration of 10 ppb at a low operating temperature of 80 • C. Thus NiO:Al films can be potentially used as a sensing element for ozone gas sensors. Funding: Part of this work supported by the project "Quality of Life" (MIS 5002464) which is implemented under the "Action for the Strategic Development on the Research and Technological Sector", funded by the Operational Programme "Competitiveness, Entrepreneurship and Innovation" (NSRF 2014-2020) and co-financed by Greece and the European Union (European Regional Development Fund).
Institutional Review Board Statement: Not applicable.