- Email: [email protected]
Multiselective visual gas sensor using nickel oxide nanowires as chemiresistor
Accepted Manuscript Title: Multiselective visual gas sensor using nickel oxide nanowires as chemiresistor Authors: Matteo Tonezzer, Le T.T. Dang, Huy Q. Tran, Salvatore Iannotta PII: DOI: Reference:
S0925-4005(17)31760-4 http://dx.doi.org/10.1016/j.snb.2017.09.094 SNB 23185
To appear in:
Sensors and Actuators B
Received date: Revised date: Accepted date:
17-3-2017 4-9-2017 13-9-2017
Please cite this article as: Matteo Tonezzer, Le T.T.Dang, Huy Q.Tran, Salvatore Iannotta, Multiselective visual gas sensor using nickel oxide nanowires as chemiresistor, Sensors and Actuators B: Chemicalhttp://dx.doi.org/10.1016/j.snb.2017.09.094 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Multiselective visual gas sensor using nickel oxide nanowires as chemiresistor Matteo Tonezzer a,*, Le T. T. Dang b, Huy Q. Tran c and Salvatore Iannotta d a
IMEM-CNR, sede di Trento - FBK, Via alla Cascata 56/C, Povo - Trento, Italy ITIMS, Hanoi University of Science and Technology, Dai Co Viet 1, Hanoi, Viet Nam c National Institute of Hygiene and Epidemiology, Yersin Street 1, Hanoi, Vietnam d IMEM-CNR, Parco Area delle Science 37/a, I-43100 Parma, Italy b
*
Corresponding author at: IMEM-CNR, sede di Trento - FBK, Via alla Cascata 56/C, Povo - Trento, Italy.
Tel: +39 0416 314828. Email address: [email protected]
KEYWORDS metal oxide, gas sensor, nanowire, nickel oxide, selectivity
Highlights:
nickel oxide nanowires have been fabricated via easy and cheap hydrothermal way
a thermal gradient (spatial or temporal) is used to get different responses
responses are combined in a three-dimensional plot with no overlapping clouds
responses are transformed in RGB components, giving a visual selective response
ABSTRACT:
Nowadays the detection of unwanted volatile compounds in air is increasingly
important in a wide range of fields. Metal oxide nanosensors are extremely small and cheap devices that could be integrated in any application, but their single resistance response make them nonselective. For this reason, sensor arrays are used where pollutant recognition is needed. Unfortunately, these electronic noses, consisting of different active materials, are complex and expensive. Here, we present a simple visual nanosensor that can detect and recognize selectively several volatile compounds at a relatively low temperature (200-300°C). The dynamic resistance of a conductometric NiO nanosensor is simply transformed in a visual output that allows to recognize different gases and their concentration with a quick look. This way, one single nanostructured metal oxide can act as a sensitive and selective electronic nose, using the powerful post-processing given
by the human eyes and brain. The sensor has proven to discriminate 8 different gases diluted in air (1 oxidizing and 7 reducing) and their respective concentration.
Keywords: Gas sensor; Metal oxide; Nikel oxide; Nanowire; Semiconductor, RGB
1. Introduction Today, gas sensors are very important: the increased social sensitivity to environment and safety makes crucial to control the quality of the ambient we are living in. Many are the aspects of human life that would benefit of a strict monitoring: food and beverages production, house and office air quality, industrial processes, terroristic threats, agriculture new technologies, early medical diagnosis and others. A spread network of nanosensors could help monitoring the quality of urban air, allowing a quick intervention and prevention of risks for human health. Metal oxides were some of the first materials used as resistive gas sensors. Numerous metal oxide semiconductors, including single (ZnO, SnO2, CuO, NiO, TiO2 and WO3) and multi-component oxides (BiFeO3, SrTiO3 and MgAl2O4) have been used as active material in gas sensing devices [1,2]. They are being widely used due to their low cost, robustness, and simple measurement electronics. Furthermore, their stability is a great advantage on their organic competitors. They have been used for decades in different forms: as thick films first [3], as thin films later, and as metal oxide (MO) nanostructured sensors in the new generation [4]. Recently, they have been widely studied in order to improve their sensing properties: controlling size [5] and shape [6] and dimensionality [7] of MO nanostructures is a powerful approach to tune their performance. Furthermore, polycrystalline nanowires can rely on two sensing mechanism, showing better performance than monocrystalline nanostructures [8,9]. At first, n-type metal oxides (above all SnO2 and ZnO) have been investigated much more than their p-type counterparts (some recent reviews are given in [10-13]). Recently though, p-type MOs (among them nickel oxide, NiO), are receiving increasing attention [14,15]. NiO is an important ptype semiconductor with wide band gap ( 3.6–4.0 eV) [16], good chemical stability and excellent electrical properties. Its potential applications extend over a wide variety of fields, including photovoltaics [17], fuel cells [18], batteries [19] and gas sensors [20-22]. So far, an assortment of NiO nanostructures have been grown, including nanoparticles [23], nanorods [24], nanowires [25], nanosheets [26] and nanotubes [27]. NiO nanowires have been grown by vapor-based oxidation method [28], electrochemical deposition [29,30], dehydration method [31] and porous anodic alumina templates [32], but unfortunately usual deposition techniques need high temperature, high vacuum, or complex
reactions, that make them power-consuming and expensive. The hydrothermal growth method [33] has been used in this work due to its easier operation and lower power consumption. Metal oxide resistive nanosensors are simple and cheap, making possible a network that easily monitors any ambient. Unfortunately, their very high sensitivity to most gases is also one of their weak points, as they are often unselective. As their n-type counterpart, p-type conductometric MO sensors have a single output: resistance, increasing (or decreasing) if the tested gas is reducing (or oxidizing) type one. Therefore, a common MO sensor can only discriminate the gas family (reducing or oxidizing). This very poor selectivity makes MO sensors insufficient when different gases from the same family are present. On the other side electronic noses (arrays of several different sensing materials including metal oxides [34,35]) are too complex and expensive to be used in wide networks. In this paper we focus on a single nanostructured metal oxide, used as a visual sensor, displaying a coloured spot, whose RGB (Red Green Blue) components are given by the sensor response at different temperatures. This simple approach (the responses are used as they are, without any processing) is very powerful because the human eye can recognize very small shade changes. The system demonstrated not only to be able to very well distinguish oxidizing from reducing gases, as well as to discriminate among 7 different reducing gases. We believe that such a simple and powerful nanosensor could be improved and applied in several practical applications.
2. Experimental 2.1. Synthesis of nickel oxide nanowires NiO nanowires were grown by means of a two-step procedure: hydrothermal growth followed by heat treatment at 500°C. All chemicals for these experiments were analytical reagent grade and were used directly, without any additional purification. In a standard process [36], 474 mg of nickel chloride [NiCl2∙6H2O] (Sigma-Aldrich, Germany) were added to a mixture of 32 mL ethylene glycol (Sigma-Aldrich, Germany) and 18 mL deionized water (stirred for 15 minutes): this solution was kept under constant magnetic stirring for 30 minutes. 120.6 mg of Na2C2O4 were then added, followed by 30 minutes of continuous stirring (to guarantee a good dispersion of Ni2+ ions). After that, the solution was moved into a teflon-lined stainless steel autoclave, sealed and heated up to 200oC and maintained at that temperature for 24 hours. Then it was let to cool down to room temperature after what the powder was collected by centrifugation, washed three times using absolute ethanol and deionized water, and then dried in air. The intermediate result at this point was
a green powder which, calcinated for 2 hours at 500°C (heating rate of 100°C/hour, and then let cool down naturally), transformed into a white powder. As mentioned in the introduction, shape and size of MO nanostructures are very important parameters when addressing gas sensing applications. Therefore we optimized the growth process in order to obtain long and thin polycrystalline nanowires. Monocrystalline nanowires indeed use only the depletion layer modulation mechanism [3], whereas granular nanowires take advantage also of the succession of potential barriers along the nanostructure axis [37,38]. Hence, the use of grainy nanowires is important for improving the nanosensors performance. Both powders (before and after the calcination) were characterized using scanning electron microscopy (SEM), X-ray diffraction (XRD), transmission electron microscopy (TEM) and selected area electron diffraction (SAED). A Bruker D5005 X-ray diffractometer with CuK1 radiation (=1.5406 Å) operated at 40 kV and 40 mA was used to carry out the XRD analysis. A JEOL7600 scanning electron microscope operated at 10 kV took the SEM images. Transmission electron microscopy and selected-area electron diffraction were performed by way of a JEOL JEM-2100 transmission electron microscope operated at 200 kV.
2.2. Nanosensors fabrication The nanowires obtained after the two-steps procedure were dissolved in ethanol by sonication (3 min). Conductometric sensors were realized dropping NiO nanowires on two interdigitated Pt/Ti electrodes on a SiO2/Si substrate. The interdigitated comb contacts were deposited on the substrate by sputtering and UV lithography. The electrodes were composed by 18 pairs of interdigitated fingers, each 800 μm long and 20 μm wide. The space between two adjacent fingers is 50 μm. Once the NiO nanowires were dropped, the device was heated up at a rate of 1.0oC/min and kept for 2 h at 500oC to facilitate the adhesion between the nanowires and metal electrodes. 2.3. Gas-sensors measurements The NiO nanowires were tested as gas sensors in dynamic conditions: the dilution (dry synthetic air) and the analyte gas (7 reducing gases, hydrogen, ethanol, H2S, LPG, NH3, CO, CO2 and one oxidizing gas, NO2, all diluted in dry synthetic air) were flown through the system at a total flow rate of 500 sccm. The dynamic response of the sensors was measured under different concentration values for each gas, as reported in Table 1.
Gas H2S NO2 CO NH3 C2H5OH H2 CO2 LPG
0.1 5 5 10 25 50 1000 2500
Concentrations tested [ppm] 0.25 0.5 1 10 25 50 10 25 50 25 50 100 50 100 250 100 250 500 2500 5000 10000 5000 10000 20000
100 500 1000
Table 1: Concentration values at which the NiO nanosensors were measured at. The values are in ppm (parts per million).
The test system was handcrafted, including a sensor chamber and a sensor holder that can be heated up to 550°C. Three Aalborg GFC17 mass flow controllers, connected to high purity gas bottles, were used to create the different dilutions, with a total flow of 400 sccm. The sensors resistance was measured by means of a home-built data acquisition system (LabView, National Instruments) connected to a Keithley 2400 multimeter. A DC voltage of 1 V was applied to the devices, that showed a good ohmic behaviour with a negligible interface resistance. The definition of sensor response S in this paper is S = Rgas/Rair, where Rgas and Rair are the resistance of the sensor with analyte gas or without it. Response and recovery times are identified as the time required to reach 90% of the maximum response and to get down to 10% of it. All the parameters were measured on four devices and the measured values were averaged.
3. RESULTS and DISCUSSION 3.1. Nanowires characterization The first step (hydrothermal growth) produced a green powder. Figure 1a shows a SEM image of the nanostructures: they are smooth nanowires with an average diameter of about 60 nm. The aspect ratio (ratio between length and diameter) can be tuned setting the processing temperature and duration. Once the powder is calcinated (calcination is made at 500°C and lasts for 2 hours), it becomes grey and changes its micro- and nano-morphology. Figure 1b and c show that the treatment at high temperature transforms the smooth nanostructures agglomerating their material in nanoparticles still keeping the linear shape of nanowires. The
particles have diameters around 20-50 nm, are slightly polyhedral, and are linked in the direction of former nanowires. Kim and Wang obtained similar results with their groups [39,40].
Figure 1: SEM images of the nanowires (a) after the hydrothermal growth and (b,c) after calcination for 2 hours at 500°C. The nanowires are smooth before the heat treatment, whilst polycrystalline after it.
The samples microstructure have then been investigated with X-ray diffraction (spectra shown in Figure 2), proving the composition of the nanostructures before and after the high temperature treatment. The many diffraction peaks present in the top plot were obtained for non-calcinated nanowires and can all be indexed as reflections from monoclinic nickel oxalate hydrate (NiC 2O4·2H2O, JCPDS 250581). The spectrum of the powder after the annealing procedure is exhibited in the bottom plot (blue online), showing well crystallized cubic NiO phase (JCPDS 47-1049). The three strong peaks at 37.3, 43.4 and 62.9° can be assigned to the cubic unit cell of nickel oxide, with a lattice parameter of 0.41667 nm. No additional peaks are present in the two spectra of Figure 2, confirming that the nanowires are crystalline without impurities or amorphous layers. The plotted spectra prove that the NiC2O4·2H2O precursor was entirely transformed into NiO during the calcination process.
Figure 2: XRD spectra of nanowires (top) before and (bottom, blue online) after the 2 h calcination at 500 °C. Both spectra agree with the JCPDS sheets of the relative material (NiC2O4 and NiO).
Morphology and crystalline structure of the nanowires before and after the thermal annealing were investigated also by TEM analysis and selected area electron diffraction. In Figure 3a is presented a TEM image of the smooth nickel oxalate hydrate nanowires: the average diameters are about 50-70 nm and the wires tend to aggregate in bundles. The two diffused rings of spots in the SAED pattern of figure 3b can be assigned to (332) and (510) diffraction lines of monoclinic NiC2O4·2H2O, confirming its crystalline structure. The SAED pattern of the NiO nanowires is given in Figure 3c, displaying four rings of bright spots. The four rings can be assigned respectively to (111), (200), (220) and (222) diffraction lines of cubic NiO. Figure 3d shows a TEM image of the polycrystalline NiO nanowires. The bundle in Figure 3d consists of some nanowires which are clearly composed of tiny nanoparticles. These morphological and structural characterizations confirm that the final product of the 2-steps growth procedure are polycrystalline NiO nanowires.
Figure 3: TEM images of (a) nickel oxalate hydrate nanowires (produced with the hydrothermal growth) and (d) NiO polycrystalline nanowires (obtained after the calcination at 500°C for 2 hours). SAED patterns relative to (b) nickel oxalate hydrate nanowires and (c) NiO polycrystalline nanowires. The bright points in both SAED patterns confirm the crystallinity of the nanowires.
3.2. Dynamic gas sensing At first a linear voltage from -5 V to +5 V was applied to the sensors in order to test the electrical contact between the nanostructured NiO and the metal electrodes. The I-V plots confirmed a very good ohmic behaviour. The devices showed a resistance in the range 3.5 MΩ - 22 kΩ as the temperature was increased from 200 to 400°C. The NiO nanodevices' resistance was constant at all temperatures and increased rapidly when a reducing gas was flown into the system. The resistance value decreased quickly to its previous value as soon as the gas flow was suspended and air was brought back into the measuring chamber. This fully matches the expected behaviour of p-type semiconductor materials. Nickel oxide is a renowned p-type semiconductor when it is grown in standard conditions. When a NiO nanowire is exposed to air, its surface will adsorb oxygen in the form of O and O2, depleting electrons from the core. This increases the number of electrical charge carriers (holes) rising the conductivity of the nanostructure [5]. The molecules of a reducing gas react with the adsorbed oxygen, releasing electrons from the chemical bonds back to the polycrystalline nanowires. This decreases the holes concentration and increases the device resistance. At all temperature values the response intensity improves with increasing gas concentration, starting with a linear behaviour that tends to saturate at higher concentrations. 3.3. Sensor response in 3D
As already mentioned, conductometric (or resistive) sensors provide only one output, in the form of a resistance increase or decrease. This means that they can only discriminate the two big families of reducing or oxidizing gases. Several papers can be found in literature considering as "highly selective" resistive sensors whose response to a specific gas is 6 or even 2.5 times higher than that to other gases [41,42]. Even if these are very good papers, we want to question this idea of selectivity, because a higher concentration of another gas would give exactly the same response of a lower concentration of the specific gas for that sensor, making them indistinguishable. For this reason we decided to use the response measurements obtained at three temperatures (200, 250 and 300°C) as coordinates to better separate the points relative to different gases. This can be achieved by testing one nanodevice at different temperature values, or by placing three nanosensors along a temperature gradient. Figure 4a shows a three-dimensional plot in which the different gases are plotted with different colour. The coordinates of each point are given by the response values of the sensor at different temperatures (200, 250 and 300°C). As can be seen, the points with only air are obviously close to the origin (response around 1 at all temperatures), while the only oxidizing gas (NO2) is in the opposite quadrant compared to all the reducing gases (quadrants projections are highlighted with light blue and pink colours). It is clear in Figures 4b and c, which are orthogonal projections of the 3D plot, that each series of points belonging to a single gas is positioned on a precise direction (roughly a straight line). As the gas concentration increases, the corresponding point in the plot is further away from the origin. This means that the threedimensional positioning could be used to selectively discriminate each gas and also (at least qualitatively) its concentration.
Figure 4: a) Three-dimensional plot in which each coordinate shows one sensor response, respectively at 200, 250 and 300°C. Orthogonal projections showing the responses at b) 200 and 250°C; c) 200 and 300°C. The points relative to each gas are drawing a different line in the plots.
In order to demonstrate that all the gases are distinguishable, we report respective zooms in Figure 5. This zoom demonstrates that also the closest points in Figure 4 (obtained measuring NH3, LPG, H2S, CO and CO2) are positioned on "lines" and far enough from each other to be discriminated.
Figure 5: Zooms of the previous plots (Figure 4) showing that the NiO nanosensor can distinguish also NH3, LPG, H2S, CO and CO2, even if their points are closer to each other.
Even if these plots demonstrate that the different gases can be discriminated by appropriate algorithms, they are not easy to be understood at a glance, especially if you consider that the points come without colours. This is the reason that lead us to the visual approach, through RGB encoding, that will be presented in Section 3.5.
3.4. Response and recovery time The response is quick and the recovery is complete for all the gases, with negligible drifts. The device response and recovery speeds depend much on the gas and on the working temperature. Figure 6 shows sensors response and recovery times at different temperatures for different analytes. As can be seen, the NiO nanosensors response and recovery times are quite long at low temperature, and become much shorter already at 250 or 300°C. This is a well-known weak point of metal oxides [43-45]. As can be seen in Figure 6, the gas with longer responses and recoveries at low temperature is CO2. Each gas shows different behaviours for response and recovery times as a function of the temperature, and only a general trend can be noticed: while increasing the device temperature, both response and recovery time decrease very much. Most gases present response times between 100 and 1000 seconds at 200°C while the range decreases to 40-200 seconds at 250°C down to 15-80 seconds at 300°C. The same happens for recovery times. The different response and recovery times behaviours among the different gases can be important, and are due to specific reactions that gaseous species have at the surface of the NiO nanowires.
Figure 6: a) Response and b) recovery times of the NiO nanosensors at different temperature values (200, 250 and 300°C). For all gases, both response and recovery times decrease as the working temperature increases. Please note that Y axis is logarithmic, therefore the decrease spans almost two orders of magnitude.
3.5. Visual multi-selective sensor The three-dimensional plot in figure 4 is not quick to understand in all its aspects within a quick look. Indeed, to see that all the gases are distinguishable, we must zoom it and rotate the coordinate axes. For this reason a different method has been used to visualize in a simple and fast way the three-dimensional response, making immediately clear that each gas has its specificity and can be selectively discriminated. The response value at 200, 250 and 300°C are converted respectively in the Red, Green and Blue components of a colour. Transforming the response range at each temperature in a 0-255 scale, the visual response in Figure 7 is realized. Each component of RGB signal goes from zero to 255, hence the normalization consists in transposing the minimummaximum response interval to the integers from 0 to 255. Response equal to 1 (no response) was transformed in RGB zero value. In this first normalization, all data were considered in order to find the maximum response, which was then set as 255 (the maximum value of one component in the whole dataset). This means that the gases whose response vary the most, will show a larger range of brightness. For example, hydrogen gives the highest response at every working temperature, thus reaching a pure white at the higher tested concentration (1000 ppm). On the other side, CO, NH3 and LPG responses do not change much with concentration, at least when compared with the response to hydrogen. For this reason a first information is immediately clear in Figure 7: "the colour of each gas". This visualization method is much better than the 3D plot in Figure 4, in which the colours of the points are given knowing a priori which gas originated that point. If all the points were plotted with the same colour, some gases could not be distinguished easily, and others would need at least some zooming and rotating of the plot. Figure 7, on the contrary, makes instantly evident that a spot belongs to a gas: even the most similar colours (NH3 and LPG) are easily discriminated by the human eye.
Figure 7: RGB visualization of the 3D plot in Figure 4. As can be seen, the different trend of response as a function of the sensor temperature (lines in the 3D space in Figure 4) is reflected in a different colour for each gas. Different shades of the same colour mean different concentrations of the same gas.
Nonetheless, Figure 7 makes it hard to evaluate the concentration of a certain gas. This problem is due to the general normalization applied at the whole set of measurements. Once the gas has been recognized (by the colour or by any other mean, like the position in the three-dimensional plot), a different normalization can be applied. This second normalization consists in eight separate normalizations: once a group of points was recognized as belonging to the same gas, only that group was used to find the minimum and maximum response values to transform into RGB 0 and 255. This separate normalization obviously got brighter colours (each group gets at least a 255 value), as can be seen in Fig. 8. It is clear that this new image allows to get more information at a glance: inside each subset of coloured spots, it is clear the trend along the gas concentration. Obviously, the visualization of each gas is affected in a different way: for example the red spots of H2S are enhanced in a minor way, and the same happens for NO2, hydrogen and CO2. On the other hand, the appearance of NH3 and ethanol spots is greatly enhanced, allowing to instantly recognize different gas concentrations. The only gas that still shows a very small variation along the subset is CO. This is due to the fact that CO response doesn't raise very much with the concentration, at any sensor temperature. We can basically see the colour difference along a subset as a "three-dimensional sensitivity" (usually defined as the slope of the response as a function of the gas concentration).
Figure 8: Enhanced RGB visualization of the 3D plot in Figure 4. The process is identical to that used to get Figure 7, but the colours are obtained normalizing each group (relative to a different gas) separately.
CO presents a very small sensitivity compared to all the other gases. Nonetheless, the human eye can see that the light blue becomes brighter while the concentration is increasing.
4. Conclusions A visual multi-selective sensor has been fabricated using NiO polycrystalline nanowires. Exploiting three different working temperatures, the usual resistive response has been transformed in a threedimensional plot that discriminates all different gases, even belonging to the same reducing family. Converting the response values at 200, 250 and 300°C into an RGB chart, it is possible to immediately recognize the detected gas, and a re-normalization of single subsets can even allow to qualitatively discriminate different concentration of the same gas. These results, achieved with one single nanostructured metal oxide, are very promising for a new generation of cheap and simple nanosensors.
5. Acknowledgements The authors acknowledge the Italian Ministry of Foreign Affairs and International Cooperation (MAECI) for funding the bilateral project HyMN.
References [1] G. Korotcenkov, Metal oxides for solid-state gas sensors: What determines our choice? Mater. Sci. Eng. B 139 (2007) 1–23. [2] G. Eranna, B.C. Joshi, D.P. Runthala, R.P. Gupta, Oxide Materials for Development of Integrated Gas Sensors—A Comprehensive Review. Crit. Rev. Solid State Mater. Sci. 29 (2004) 111–188. [3] G. Pfaff, Effect of powder preparation and sintering on the electrical properties of tin dioxide-based ceramic gas sensors, Sensor Actuat. B-Chem., 20 (1994) 43-48. [4] I. Paulowicz, V. Hrkac, S. Kaps, V. Cretu, O. Lupan, T. Braniste, V. Duppel, I. Tiginyanu , L. Kienle, R.Adelung and Y.K. Mishra, Three-dimensional SnO2 nanowire networks for multifunctional applications: from hightemperature stretchable ceramics to ultraresponsive sensors, Adv. Elec. Mater. 1 (2015) 1500081. [5] M. Tonezzer and N. V. Hieu, Size-dependent response of single-nanowire gas sensors, Sensor Actuat. B-Chem. 163 (2012) 146–152. [6] A. Sanger, P.K. Jain, Y.K. Mishrac, R. Chandraa, Palladium decorated silicon carbide nanocauliflowers for hydrogen gas sensing application, Sensors Actuat. B-Chem. 242 (2016) 694-699. [7] M. Tonezzer, T. T. L. Dang, N. Bazzanella, V. H. Nguyen, S. Iannotta, Comparative gas-sensing performance of 1D and 2D ZnO nanostructures, Sensor Actuat. B-Chem. 220 (2015) 1152–1160. [8] X. Jiang, Y. Wang, T. Herricks and Y. Xia, Ethylene glycol-mediated synthesis of metal oxide nanowires, J. Mater. Chem. 14 (2004) 695-703. [9] Y. Wang, X. Jiang, and Y. Xia, A Solution-Phase, Precursor route to polycrystalline SnO2 nanowires that can be used for gas sensing under ambient conditions, J. Am. Chem. Soc. 125 (2003) 16176-16177. [10] J.P. Cheng, Jiao Wang, Q.Q. Li, H.G. Liu, Y. Li, A review of recent developments in tin dioxide composites for gas sensing application, Journal of Industrial and Engineering Chemistry 44 (2016) 1–22. [11] S.J. Patil, A.V. Patil, C.G. Dighavkar, K.S. Thakare, R.Y. Borase, S.J. Nandre, N.G. Deshpande and R.R. Ahire, Semiconductor metal oxide compounds based gas sensors: A literature review, Front. Mater. Sci. 9 (2015) 14-37. [12] B. Yuliarto, G. Gumilar, and N.L.W. Septiani, SnO 2 Nanostructure as Pollutant Gas Sensors: Synthesis, Sensing Performances, and Mechanism, Advances in Materials Science and Engineering 2015 (2015) 694823. [13] R. Kumar, O. Al-Dossary, G. Kumar, A. Umar, Zinc Oxide Nanostructures for NO2 Gas–Sensor Applications: A Review, Nano-Micro Lett. 7 (2015) 97-120. [14] E.S. Hassan, A.A. Saeed, A.K. Elttayef, Doping and thickness variation influence on the structural and sensing properties of NiO film prepared by RF-magnetron sputtering, J. Mater. Sci.: Mater. Electron. 27 (2016) 1270– 1277. [15] Y. Lu, Y.H. Ma, S.Y. Ma, W.X. Jin, S.H. Yan, X.L. Xu, X.H. Jiang, T.T. Wang, H.M. Yang, H. Chen, Z. Qiang, Synthesis of cactus-like NiO nanostructure and their gas-sensing properties, Mater. Lett. 164 (2016) 48–51. [16] R. LoNigro, S. Battiato, G. Greco, P. Fiorenza, F. Roccaforte, G. Malandrino, Metal Organic Chemical Vapor Deposition of nickel oxide thin films for wide band gap device technology, Thin Solid Films 563 (2014) 50–55. [17] M-W. Lin, K-C. Wang, J-H, Wang, M-H. Li, Y-L. Lai, T. Ohigashi, N. Kosugi, P. Chen, D.H. Wei, T.F. Guo, Y-J. Hsu, Improve Hole Collection by Interfacial Chemical Redox Reaction at a Mesoscopic NiO/CH3NH3PbI3 Heterojunction for Efficient Photovoltaic Cells, Adv. Mater. Interfaces 3 (2016) 1600135. [18] D. Xu, K. Li, Y. Zhou, Y. Gao, D. Yan, S. Xu, The effect of NiO addition on the grain boundary behavior and electrochemical performance of Gd-doped ceria solid electrolyte under different sintering conditions, J. Eur. Ceram. Soc. 37 (2017) 419-425. [19] B. Wu, M-Q. Sheng, S-N. Gao, Y. Wang, F. Liao, Single-source precursor to Ag/NiO composite for rechargeable charge storage, J. Alloy Compd. 692 (2017) 34-39. [20] R-Y. Miao, W. Zeng, Q. Gao, SDS-assisted hydrothermal synthesis of NiO flake-flower architectures with enhanced gas-sensing properties, Appl. Surf. Sci. 384 (2016) 304-310. [21] M. Tonezzer, D.T.T. Le, T. Q. Huy, S. Iannotta, Dual-selective hydrogen and ethanol sensor for steam reforming systems, Sensor Actuat. B-Chem. 236 (2016) 1011–1019. [22] N. Kaur, E. Comini, D. Zappa, N. Poli, G. Sberveglieri, Nickel oxide nanowires: vapor liquid solid synthesis and integration into a gas sensing device, Nanotechnology 27 (2016) 205701. [23] W-C. Huang, J-D. Kim, Nickel oxide nanoparticle-based method for simultaneous harvesting and disruption of microalgal cells, Bioresource Technol. 218 (2016) 1290-1293. [24] H-C. Pang, B-Q. Guan, W-W. Sun, Y. Wang., Metal-Organic-Frameworks Derivation of Mesoporous NiO Nanorod for High-Performance Lithium Ion Batteries, Electrochim. Acta, 213 (2016) 351-357.
[25] C. Liu, C-L. Li, K. Ahmed, Z. Mutlu, C.S. Ozkan, M. Ozkan, Template Free and Binderless NiO Nanowire Foam for Li-ion Battery Anodes with Long Cycle Life and Ultrahigh Rate Capability, Sci. Rep. 6 (2016) 29183. [26] J. Min, J. Liu, M. Lei, W-J. Wang, Y-K. Lu, L-Y. Yang, Q. Yang, G. Liu, N. Su, Self-Assembly of Parallelly Aligned NiO Hierarchical Nanostructures with Ultrathin Nanosheet Subunits for Electrochemical Supercapacitor Applications, ACS Appl. Mater. Inter. 8 (2016) 780-791. [27] A-F. Liu, H-W. Che, Y-Z. Mao, Y-Q. Wang, J-B. Mu, C-X. Wu, Y-M. Bai, X-L. Zhang, G-S. Wang, Templatefree synthesis of one-dimensional hierarchical NiO nanotubes self-assembled by nanosheets for high-performance supercapacitors, Ceram. Int. 42 (2016) 11435-11441. [28] Z.P. Wei, M. Arredondo, H.Y. Peng, Z. Zhang, D.L. Guo, G.Z. Xing, Y.F. Li, L.M. Wong, S.J. Wang, N. Valanoor, T. Wu, A template and catalyst-free metal-etching-oxidation method to synthesize aligned oxide nanowire arrays: NiO as an example, ACS Nano 4 (2010) 4785–4791. [29] Y. Lin, T. Xie, B. Cheng, B. Geng, L. Zhang, Ordered nickel oxide nanowire arrays and their optical absorption properties, Chem. Phys. Lett. 380 (2003) 521–525. [30] K. Nielsch, F. Müller, A.-P. Li, U. Gösele, Uniform nickel deposition into ordered alumina pores by pulsed electrodeposition, Adv. Mater. 12 (2000) 582–586. [31] H. Pang, Q. Lu, Y. Zhang, Y. Li, F. Gao, Selective synthesis of nickel oxide nanowires and length effect on their electrochemical properties, Nanoscale 2 (2010) 920–922. [32] A. Kolmakov, Y. Zhang, G. Cheng and M. Moskovits, Detection of CO and O2 using tin oxide nanowire sensors, Adv. Mater. 15 (2003) 997-1000. [33] B. Miao, W. Zeng, L. Lin, S. Xu, Characterization and gas-sensing properties of NiO nanowires prepared through hydrothermal method, Physica E 52 (2013) 40–45. [34] F. Röck, N. Barsan, U. Weimar, Electronic Nose: Current Status and Future Trends, Chem. Rev. 108 (2008) 705725. [35] E.N. Carmona, V. Sberveglieri, A. Ponzoni, V. Galstyana, D. Zappa, A. Pulvirenti, E. Comini, Detection of food and skin pathogen microbiota by means of an electronic nose based on metal oxide chemiresistors, Sensor Actuat. B-Chem. 238 (2017) 1224. [36] T.T.L. Dang, M. Tonezzer, V.H. Hieu, Hydrothermal Growth and Hydrogen Selective Sensing of Nickel Oxide Nanowires, J. Nanomater. 2015 (2015) 785856. [37] A. Katoch, S-W. Choi, H.W. Kim, S.S. Kim, Highly sensitive and selective H2 sensing by ZnO nanofibers and the underlying sensing mechanism, J. Hazard. Mater. 286 (2015) 229–235. [38] G.K. Mani, J.B.B. Rayappan, A highly selective and wide range ammonia sensor—Nanostructured ZnO:Co thin film, Mater. Sci. Eng. B 191 (2015) 41–50. [39] J-M. Choi, J-H. Byun, S.S. Kim, Influence of grain size on gas-sensing properties of chemiresistive p-type NiOnanofibers, Sensor Actuat. B-Chem. 227 (2016) 149–156. [40] C. Wang, T. Wang, B. Wang, X. Zhou, X. Cheng, P. Sun, J. Zheng, G. Lu, Design of α-Fe2O3 nanorods functionalized tubular NiO nanostructure for discriminating toluene molecules, Sci. Rep. 6 (2016) 26432. [41] H-S. Woo, C.W. Na, I.-D. Kim, J-H. Lee, Highly sensitive and selective trimethylamine sensor using onedimensional ZnO–Cr2O3 hetero-nanostructures, Nanotechnology 23 (2012) 245501. [42] A. Sanger, A. Kumar, A. Kumar, R. Chandra, Highly sensitive and selective hydrogen gas sensor using sputtered grown Pd decorated MnO2 nanowalls, Sensor Actuat. B-Chem. 234 (2016) 8-14. [43] E. Oh, H-Y. Choi, S-H. Jung, S. Cho, J.C. Kim, K-H. Lee, S-W. Kang, J. Kim, J-Y. Yun, S-H. Jeong, Highperformance NO2 gas sensor based on ZnO nanorod grown by ultrasonic irradiation, Sens. Actuat. B-Chem. 141 (2009) 239–243. [44] B. Urasinska-Wojcik, T.A. Vincent, M.F. Chowdhury, J.W. Gardner, Ultrasensitive WO3 gas sensors for NO2 detection in air and low oxygen environment, Sens. Actuat. B-Chem. 239 (2017) 1051–1059. [45] A. Gaiardo, B. Fabbri, A. Giberti, V. Guidi, P. Bellutti, C. Malagù, M. Valt, G. Pepponi, S. Gherardi, G. Zonta, A. Martucci, M. Sturaro, N. Landini, ZnO and Au/ZnO thin films: Room-temperature chemoresistive properties for gas sensing applications, Sens. Actuat. B-Chem. 237 (2016) 1085–1094.
Biographies
Matteo Tonezzer graduated “cum laude” in Physics of the Matter and received hisPhD degree “with honor” from the Faculty of Physics at the University of Trento, Italy, in 2011. His thesis was the optimization of inorganic and organic nanostructured materials toward gas sensing. In 2011, he won the Young Scientist Award from the European Materials Research Society (EMRS). He worked in research centers in France (ESRF), Brazil (UFMG), Vietnam (HUST), and USA (GaTech). He is currently working for IMEM at the Italian National Research Council. Major research interest is synthesis and characterization of nanostructured materials. Dang Thi Thanh Le obtained her MSc and PhD degrees in Materials Science from International Training Institute for Materials Science (ITIMS)-Hanoi University of Science and Technology (HUST), Hanoi, Vietnam in 2001 and 2011. She worked as a visiting postdoc at The Angstrom Laboratory-Uppsala University, Sweden in the academic year of 2011–2012. She is working as a researcher/lecturer at ITIMS. Her current interests include synthesis, characterization and application of nanomaterials for gas-sensing and bio-sensing. Tran Quang Huy received his PhD in Materials Science at Hanoi University of Science and Technology in 2012. He is responsible for the Nanobiomedicine Group at the National Insitute of Hygiene and Epidemiology in Hanoi and principal investigator of Nano Energy & NanoBiomedical Research Group. He coordinated several national and international projects. He published more than 40 articles in peer-reviewed journals and contributed 6 chapters for other books. He is a reviewer for several international scientific journals, editorial board member of the Vietnam Journal of Preventive Medicine. His current research interests are bionanomaterials, biological ultrastructures and technology solutions for environmental treatments. Salvatore Iannotta graduated in Physics at the University of Bologna, then in August1984 the PhD in Chemistry at the (GWC) (Guelph Waterloo Centre for Graduate Work in Chemistry-Ontario Canada). He is presently Director of CNR-IMEM of materials for electronic and magnetism. Major research interests are: synthesis and characterization of nanostructured and molecular materials; supersonic beams of metallic and semiconductor clusters as well as of molecular aggregates, organic electronics, gas a VOC sensors, memristive materials devices and systems, active biosensors and bio-electronics. He is author of more than 130 papers and invited speaker to a wide number of international conferences.
S0925-4005(17)31760-4 http://dx.doi.org/10.1016/j.snb.2017.09.094 SNB 23185
To appear in:
Sensors and Actuators B
Received date: Revised date: Accepted date:
17-3-2017 4-9-2017 13-9-2017
Please cite this article as: Matteo Tonezzer, Le T.T.Dang, Huy Q.Tran, Salvatore Iannotta, Multiselective visual gas sensor using nickel oxide nanowires as chemiresistor, Sensors and Actuators B: Chemicalhttp://dx.doi.org/10.1016/j.snb.2017.09.094 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Multiselective visual gas sensor using nickel oxide nanowires as chemiresistor Matteo Tonezzer a,*, Le T. T. Dang b, Huy Q. Tran c and Salvatore Iannotta d a
IMEM-CNR, sede di Trento - FBK, Via alla Cascata 56/C, Povo - Trento, Italy ITIMS, Hanoi University of Science and Technology, Dai Co Viet 1, Hanoi, Viet Nam c National Institute of Hygiene and Epidemiology, Yersin Street 1, Hanoi, Vietnam d IMEM-CNR, Parco Area delle Science 37/a, I-43100 Parma, Italy b
*
Corresponding author at: IMEM-CNR, sede di Trento - FBK, Via alla Cascata 56/C, Povo - Trento, Italy.
Tel: +39 0416 314828. Email address: [email protected]
KEYWORDS metal oxide, gas sensor, nanowire, nickel oxide, selectivity
Highlights:
nickel oxide nanowires have been fabricated via easy and cheap hydrothermal way
a thermal gradient (spatial or temporal) is used to get different responses
responses are combined in a three-dimensional plot with no overlapping clouds
responses are transformed in RGB components, giving a visual selective response
ABSTRACT:
Nowadays the detection of unwanted volatile compounds in air is increasingly
important in a wide range of fields. Metal oxide nanosensors are extremely small and cheap devices that could be integrated in any application, but their single resistance response make them nonselective. For this reason, sensor arrays are used where pollutant recognition is needed. Unfortunately, these electronic noses, consisting of different active materials, are complex and expensive. Here, we present a simple visual nanosensor that can detect and recognize selectively several volatile compounds at a relatively low temperature (200-300°C). The dynamic resistance of a conductometric NiO nanosensor is simply transformed in a visual output that allows to recognize different gases and their concentration with a quick look. This way, one single nanostructured metal oxide can act as a sensitive and selective electronic nose, using the powerful post-processing given
by the human eyes and brain. The sensor has proven to discriminate 8 different gases diluted in air (1 oxidizing and 7 reducing) and their respective concentration.
Keywords: Gas sensor; Metal oxide; Nikel oxide; Nanowire; Semiconductor, RGB
1. Introduction Today, gas sensors are very important: the increased social sensitivity to environment and safety makes crucial to control the quality of the ambient we are living in. Many are the aspects of human life that would benefit of a strict monitoring: food and beverages production, house and office air quality, industrial processes, terroristic threats, agriculture new technologies, early medical diagnosis and others. A spread network of nanosensors could help monitoring the quality of urban air, allowing a quick intervention and prevention of risks for human health. Metal oxides were some of the first materials used as resistive gas sensors. Numerous metal oxide semiconductors, including single (ZnO, SnO2, CuO, NiO, TiO2 and WO3) and multi-component oxides (BiFeO3, SrTiO3 and MgAl2O4) have been used as active material in gas sensing devices [1,2]. They are being widely used due to their low cost, robustness, and simple measurement electronics. Furthermore, their stability is a great advantage on their organic competitors. They have been used for decades in different forms: as thick films first [3], as thin films later, and as metal oxide (MO) nanostructured sensors in the new generation [4]. Recently, they have been widely studied in order to improve their sensing properties: controlling size [5] and shape [6] and dimensionality [7] of MO nanostructures is a powerful approach to tune their performance. Furthermore, polycrystalline nanowires can rely on two sensing mechanism, showing better performance than monocrystalline nanostructures [8,9]. At first, n-type metal oxides (above all SnO2 and ZnO) have been investigated much more than their p-type counterparts (some recent reviews are given in [10-13]). Recently though, p-type MOs (among them nickel oxide, NiO), are receiving increasing attention [14,15]. NiO is an important ptype semiconductor with wide band gap ( 3.6–4.0 eV) [16], good chemical stability and excellent electrical properties. Its potential applications extend over a wide variety of fields, including photovoltaics [17], fuel cells [18], batteries [19] and gas sensors [20-22]. So far, an assortment of NiO nanostructures have been grown, including nanoparticles [23], nanorods [24], nanowires [25], nanosheets [26] and nanotubes [27]. NiO nanowires have been grown by vapor-based oxidation method [28], electrochemical deposition [29,30], dehydration method [31] and porous anodic alumina templates [32], but unfortunately usual deposition techniques need high temperature, high vacuum, or complex
reactions, that make them power-consuming and expensive. The hydrothermal growth method [33] has been used in this work due to its easier operation and lower power consumption. Metal oxide resistive nanosensors are simple and cheap, making possible a network that easily monitors any ambient. Unfortunately, their very high sensitivity to most gases is also one of their weak points, as they are often unselective. As their n-type counterpart, p-type conductometric MO sensors have a single output: resistance, increasing (or decreasing) if the tested gas is reducing (or oxidizing) type one. Therefore, a common MO sensor can only discriminate the gas family (reducing or oxidizing). This very poor selectivity makes MO sensors insufficient when different gases from the same family are present. On the other side electronic noses (arrays of several different sensing materials including metal oxides [34,35]) are too complex and expensive to be used in wide networks. In this paper we focus on a single nanostructured metal oxide, used as a visual sensor, displaying a coloured spot, whose RGB (Red Green Blue) components are given by the sensor response at different temperatures. This simple approach (the responses are used as they are, without any processing) is very powerful because the human eye can recognize very small shade changes. The system demonstrated not only to be able to very well distinguish oxidizing from reducing gases, as well as to discriminate among 7 different reducing gases. We believe that such a simple and powerful nanosensor could be improved and applied in several practical applications.
2. Experimental 2.1. Synthesis of nickel oxide nanowires NiO nanowires were grown by means of a two-step procedure: hydrothermal growth followed by heat treatment at 500°C. All chemicals for these experiments were analytical reagent grade and were used directly, without any additional purification. In a standard process [36], 474 mg of nickel chloride [NiCl2∙6H2O] (Sigma-Aldrich, Germany) were added to a mixture of 32 mL ethylene glycol (Sigma-Aldrich, Germany) and 18 mL deionized water (stirred for 15 minutes): this solution was kept under constant magnetic stirring for 30 minutes. 120.6 mg of Na2C2O4 were then added, followed by 30 minutes of continuous stirring (to guarantee a good dispersion of Ni2+ ions). After that, the solution was moved into a teflon-lined stainless steel autoclave, sealed and heated up to 200oC and maintained at that temperature for 24 hours. Then it was let to cool down to room temperature after what the powder was collected by centrifugation, washed three times using absolute ethanol and deionized water, and then dried in air. The intermediate result at this point was
a green powder which, calcinated for 2 hours at 500°C (heating rate of 100°C/hour, and then let cool down naturally), transformed into a white powder. As mentioned in the introduction, shape and size of MO nanostructures are very important parameters when addressing gas sensing applications. Therefore we optimized the growth process in order to obtain long and thin polycrystalline nanowires. Monocrystalline nanowires indeed use only the depletion layer modulation mechanism [3], whereas granular nanowires take advantage also of the succession of potential barriers along the nanostructure axis [37,38]. Hence, the use of grainy nanowires is important for improving the nanosensors performance. Both powders (before and after the calcination) were characterized using scanning electron microscopy (SEM), X-ray diffraction (XRD), transmission electron microscopy (TEM) and selected area electron diffraction (SAED). A Bruker D5005 X-ray diffractometer with CuK1 radiation (=1.5406 Å) operated at 40 kV and 40 mA was used to carry out the XRD analysis. A JEOL7600 scanning electron microscope operated at 10 kV took the SEM images. Transmission electron microscopy and selected-area electron diffraction were performed by way of a JEOL JEM-2100 transmission electron microscope operated at 200 kV.
2.2. Nanosensors fabrication The nanowires obtained after the two-steps procedure were dissolved in ethanol by sonication (3 min). Conductometric sensors were realized dropping NiO nanowires on two interdigitated Pt/Ti electrodes on a SiO2/Si substrate. The interdigitated comb contacts were deposited on the substrate by sputtering and UV lithography. The electrodes were composed by 18 pairs of interdigitated fingers, each 800 μm long and 20 μm wide. The space between two adjacent fingers is 50 μm. Once the NiO nanowires were dropped, the device was heated up at a rate of 1.0oC/min and kept for 2 h at 500oC to facilitate the adhesion between the nanowires and metal electrodes. 2.3. Gas-sensors measurements The NiO nanowires were tested as gas sensors in dynamic conditions: the dilution (dry synthetic air) and the analyte gas (7 reducing gases, hydrogen, ethanol, H2S, LPG, NH3, CO, CO2 and one oxidizing gas, NO2, all diluted in dry synthetic air) were flown through the system at a total flow rate of 500 sccm. The dynamic response of the sensors was measured under different concentration values for each gas, as reported in Table 1.
Gas H2S NO2 CO NH3 C2H5OH H2 CO2 LPG
0.1 5 5 10 25 50 1000 2500
Concentrations tested [ppm] 0.25 0.5 1 10 25 50 10 25 50 25 50 100 50 100 250 100 250 500 2500 5000 10000 5000 10000 20000
100 500 1000
Table 1: Concentration values at which the NiO nanosensors were measured at. The values are in ppm (parts per million).
The test system was handcrafted, including a sensor chamber and a sensor holder that can be heated up to 550°C. Three Aalborg GFC17 mass flow controllers, connected to high purity gas bottles, were used to create the different dilutions, with a total flow of 400 sccm. The sensors resistance was measured by means of a home-built data acquisition system (LabView, National Instruments) connected to a Keithley 2400 multimeter. A DC voltage of 1 V was applied to the devices, that showed a good ohmic behaviour with a negligible interface resistance. The definition of sensor response S in this paper is S = Rgas/Rair, where Rgas and Rair are the resistance of the sensor with analyte gas or without it. Response and recovery times are identified as the time required to reach 90% of the maximum response and to get down to 10% of it. All the parameters were measured on four devices and the measured values were averaged.
3. RESULTS and DISCUSSION 3.1. Nanowires characterization The first step (hydrothermal growth) produced a green powder. Figure 1a shows a SEM image of the nanostructures: they are smooth nanowires with an average diameter of about 60 nm. The aspect ratio (ratio between length and diameter) can be tuned setting the processing temperature and duration. Once the powder is calcinated (calcination is made at 500°C and lasts for 2 hours), it becomes grey and changes its micro- and nano-morphology. Figure 1b and c show that the treatment at high temperature transforms the smooth nanostructures agglomerating their material in nanoparticles still keeping the linear shape of nanowires. The
particles have diameters around 20-50 nm, are slightly polyhedral, and are linked in the direction of former nanowires. Kim and Wang obtained similar results with their groups [39,40].
Figure 1: SEM images of the nanowires (a) after the hydrothermal growth and (b,c) after calcination for 2 hours at 500°C. The nanowires are smooth before the heat treatment, whilst polycrystalline after it.
The samples microstructure have then been investigated with X-ray diffraction (spectra shown in Figure 2), proving the composition of the nanostructures before and after the high temperature treatment. The many diffraction peaks present in the top plot were obtained for non-calcinated nanowires and can all be indexed as reflections from monoclinic nickel oxalate hydrate (NiC 2O4·2H2O, JCPDS 250581). The spectrum of the powder after the annealing procedure is exhibited in the bottom plot (blue online), showing well crystallized cubic NiO phase (JCPDS 47-1049). The three strong peaks at 37.3, 43.4 and 62.9° can be assigned to the cubic unit cell of nickel oxide, with a lattice parameter of 0.41667 nm. No additional peaks are present in the two spectra of Figure 2, confirming that the nanowires are crystalline without impurities or amorphous layers. The plotted spectra prove that the NiC2O4·2H2O precursor was entirely transformed into NiO during the calcination process.
Figure 2: XRD spectra of nanowires (top) before and (bottom, blue online) after the 2 h calcination at 500 °C. Both spectra agree with the JCPDS sheets of the relative material (NiC2O4 and NiO).
Morphology and crystalline structure of the nanowires before and after the thermal annealing were investigated also by TEM analysis and selected area electron diffraction. In Figure 3a is presented a TEM image of the smooth nickel oxalate hydrate nanowires: the average diameters are about 50-70 nm and the wires tend to aggregate in bundles. The two diffused rings of spots in the SAED pattern of figure 3b can be assigned to (332) and (510) diffraction lines of monoclinic NiC2O4·2H2O, confirming its crystalline structure. The SAED pattern of the NiO nanowires is given in Figure 3c, displaying four rings of bright spots. The four rings can be assigned respectively to (111), (200), (220) and (222) diffraction lines of cubic NiO. Figure 3d shows a TEM image of the polycrystalline NiO nanowires. The bundle in Figure 3d consists of some nanowires which are clearly composed of tiny nanoparticles. These morphological and structural characterizations confirm that the final product of the 2-steps growth procedure are polycrystalline NiO nanowires.
Figure 3: TEM images of (a) nickel oxalate hydrate nanowires (produced with the hydrothermal growth) and (d) NiO polycrystalline nanowires (obtained after the calcination at 500°C for 2 hours). SAED patterns relative to (b) nickel oxalate hydrate nanowires and (c) NiO polycrystalline nanowires. The bright points in both SAED patterns confirm the crystallinity of the nanowires.
3.2. Dynamic gas sensing At first a linear voltage from -5 V to +5 V was applied to the sensors in order to test the electrical contact between the nanostructured NiO and the metal electrodes. The I-V plots confirmed a very good ohmic behaviour. The devices showed a resistance in the range 3.5 MΩ - 22 kΩ as the temperature was increased from 200 to 400°C. The NiO nanodevices' resistance was constant at all temperatures and increased rapidly when a reducing gas was flown into the system. The resistance value decreased quickly to its previous value as soon as the gas flow was suspended and air was brought back into the measuring chamber. This fully matches the expected behaviour of p-type semiconductor materials. Nickel oxide is a renowned p-type semiconductor when it is grown in standard conditions. When a NiO nanowire is exposed to air, its surface will adsorb oxygen in the form of O and O2, depleting electrons from the core. This increases the number of electrical charge carriers (holes) rising the conductivity of the nanostructure [5]. The molecules of a reducing gas react with the adsorbed oxygen, releasing electrons from the chemical bonds back to the polycrystalline nanowires. This decreases the holes concentration and increases the device resistance. At all temperature values the response intensity improves with increasing gas concentration, starting with a linear behaviour that tends to saturate at higher concentrations. 3.3. Sensor response in 3D
As already mentioned, conductometric (or resistive) sensors provide only one output, in the form of a resistance increase or decrease. This means that they can only discriminate the two big families of reducing or oxidizing gases. Several papers can be found in literature considering as "highly selective" resistive sensors whose response to a specific gas is 6 or even 2.5 times higher than that to other gases [41,42]. Even if these are very good papers, we want to question this idea of selectivity, because a higher concentration of another gas would give exactly the same response of a lower concentration of the specific gas for that sensor, making them indistinguishable. For this reason we decided to use the response measurements obtained at three temperatures (200, 250 and 300°C) as coordinates to better separate the points relative to different gases. This can be achieved by testing one nanodevice at different temperature values, or by placing three nanosensors along a temperature gradient. Figure 4a shows a three-dimensional plot in which the different gases are plotted with different colour. The coordinates of each point are given by the response values of the sensor at different temperatures (200, 250 and 300°C). As can be seen, the points with only air are obviously close to the origin (response around 1 at all temperatures), while the only oxidizing gas (NO2) is in the opposite quadrant compared to all the reducing gases (quadrants projections are highlighted with light blue and pink colours). It is clear in Figures 4b and c, which are orthogonal projections of the 3D plot, that each series of points belonging to a single gas is positioned on a precise direction (roughly a straight line). As the gas concentration increases, the corresponding point in the plot is further away from the origin. This means that the threedimensional positioning could be used to selectively discriminate each gas and also (at least qualitatively) its concentration.
Figure 4: a) Three-dimensional plot in which each coordinate shows one sensor response, respectively at 200, 250 and 300°C. Orthogonal projections showing the responses at b) 200 and 250°C; c) 200 and 300°C. The points relative to each gas are drawing a different line in the plots.
In order to demonstrate that all the gases are distinguishable, we report respective zooms in Figure 5. This zoom demonstrates that also the closest points in Figure 4 (obtained measuring NH3, LPG, H2S, CO and CO2) are positioned on "lines" and far enough from each other to be discriminated.
Figure 5: Zooms of the previous plots (Figure 4) showing that the NiO nanosensor can distinguish also NH3, LPG, H2S, CO and CO2, even if their points are closer to each other.
Even if these plots demonstrate that the different gases can be discriminated by appropriate algorithms, they are not easy to be understood at a glance, especially if you consider that the points come without colours. This is the reason that lead us to the visual approach, through RGB encoding, that will be presented in Section 3.5.
3.4. Response and recovery time The response is quick and the recovery is complete for all the gases, with negligible drifts. The device response and recovery speeds depend much on the gas and on the working temperature. Figure 6 shows sensors response and recovery times at different temperatures for different analytes. As can be seen, the NiO nanosensors response and recovery times are quite long at low temperature, and become much shorter already at 250 or 300°C. This is a well-known weak point of metal oxides [43-45]. As can be seen in Figure 6, the gas with longer responses and recoveries at low temperature is CO2. Each gas shows different behaviours for response and recovery times as a function of the temperature, and only a general trend can be noticed: while increasing the device temperature, both response and recovery time decrease very much. Most gases present response times between 100 and 1000 seconds at 200°C while the range decreases to 40-200 seconds at 250°C down to 15-80 seconds at 300°C. The same happens for recovery times. The different response and recovery times behaviours among the different gases can be important, and are due to specific reactions that gaseous species have at the surface of the NiO nanowires.
Figure 6: a) Response and b) recovery times of the NiO nanosensors at different temperature values (200, 250 and 300°C). For all gases, both response and recovery times decrease as the working temperature increases. Please note that Y axis is logarithmic, therefore the decrease spans almost two orders of magnitude.
3.5. Visual multi-selective sensor The three-dimensional plot in figure 4 is not quick to understand in all its aspects within a quick look. Indeed, to see that all the gases are distinguishable, we must zoom it and rotate the coordinate axes. For this reason a different method has been used to visualize in a simple and fast way the three-dimensional response, making immediately clear that each gas has its specificity and can be selectively discriminated. The response value at 200, 250 and 300°C are converted respectively in the Red, Green and Blue components of a colour. Transforming the response range at each temperature in a 0-255 scale, the visual response in Figure 7 is realized. Each component of RGB signal goes from zero to 255, hence the normalization consists in transposing the minimummaximum response interval to the integers from 0 to 255. Response equal to 1 (no response) was transformed in RGB zero value. In this first normalization, all data were considered in order to find the maximum response, which was then set as 255 (the maximum value of one component in the whole dataset). This means that the gases whose response vary the most, will show a larger range of brightness. For example, hydrogen gives the highest response at every working temperature, thus reaching a pure white at the higher tested concentration (1000 ppm). On the other side, CO, NH3 and LPG responses do not change much with concentration, at least when compared with the response to hydrogen. For this reason a first information is immediately clear in Figure 7: "the colour of each gas". This visualization method is much better than the 3D plot in Figure 4, in which the colours of the points are given knowing a priori which gas originated that point. If all the points were plotted with the same colour, some gases could not be distinguished easily, and others would need at least some zooming and rotating of the plot. Figure 7, on the contrary, makes instantly evident that a spot belongs to a gas: even the most similar colours (NH3 and LPG) are easily discriminated by the human eye.
Figure 7: RGB visualization of the 3D plot in Figure 4. As can be seen, the different trend of response as a function of the sensor temperature (lines in the 3D space in Figure 4) is reflected in a different colour for each gas. Different shades of the same colour mean different concentrations of the same gas.
Nonetheless, Figure 7 makes it hard to evaluate the concentration of a certain gas. This problem is due to the general normalization applied at the whole set of measurements. Once the gas has been recognized (by the colour or by any other mean, like the position in the three-dimensional plot), a different normalization can be applied. This second normalization consists in eight separate normalizations: once a group of points was recognized as belonging to the same gas, only that group was used to find the minimum and maximum response values to transform into RGB 0 and 255. This separate normalization obviously got brighter colours (each group gets at least a 255 value), as can be seen in Fig. 8. It is clear that this new image allows to get more information at a glance: inside each subset of coloured spots, it is clear the trend along the gas concentration. Obviously, the visualization of each gas is affected in a different way: for example the red spots of H2S are enhanced in a minor way, and the same happens for NO2, hydrogen and CO2. On the other hand, the appearance of NH3 and ethanol spots is greatly enhanced, allowing to instantly recognize different gas concentrations. The only gas that still shows a very small variation along the subset is CO. This is due to the fact that CO response doesn't raise very much with the concentration, at any sensor temperature. We can basically see the colour difference along a subset as a "three-dimensional sensitivity" (usually defined as the slope of the response as a function of the gas concentration).
Figure 8: Enhanced RGB visualization of the 3D plot in Figure 4. The process is identical to that used to get Figure 7, but the colours are obtained normalizing each group (relative to a different gas) separately.
CO presents a very small sensitivity compared to all the other gases. Nonetheless, the human eye can see that the light blue becomes brighter while the concentration is increasing.
4. Conclusions A visual multi-selective sensor has been fabricated using NiO polycrystalline nanowires. Exploiting three different working temperatures, the usual resistive response has been transformed in a threedimensional plot that discriminates all different gases, even belonging to the same reducing family. Converting the response values at 200, 250 and 300°C into an RGB chart, it is possible to immediately recognize the detected gas, and a re-normalization of single subsets can even allow to qualitatively discriminate different concentration of the same gas. These results, achieved with one single nanostructured metal oxide, are very promising for a new generation of cheap and simple nanosensors.
5. Acknowledgements The authors acknowledge the Italian Ministry of Foreign Affairs and International Cooperation (MAECI) for funding the bilateral project HyMN.
References [1] G. Korotcenkov, Metal oxides for solid-state gas sensors: What determines our choice? Mater. Sci. Eng. B 139 (2007) 1–23. [2] G. Eranna, B.C. Joshi, D.P. Runthala, R.P. Gupta, Oxide Materials for Development of Integrated Gas Sensors—A Comprehensive Review. Crit. Rev. Solid State Mater. Sci. 29 (2004) 111–188. [3] G. Pfaff, Effect of powder preparation and sintering on the electrical properties of tin dioxide-based ceramic gas sensors, Sensor Actuat. B-Chem., 20 (1994) 43-48. [4] I. Paulowicz, V. Hrkac, S. Kaps, V. Cretu, O. Lupan, T. Braniste, V. Duppel, I. Tiginyanu , L. Kienle, R.Adelung and Y.K. Mishra, Three-dimensional SnO2 nanowire networks for multifunctional applications: from hightemperature stretchable ceramics to ultraresponsive sensors, Adv. Elec. Mater. 1 (2015) 1500081. [5] M. Tonezzer and N. V. Hieu, Size-dependent response of single-nanowire gas sensors, Sensor Actuat. B-Chem. 163 (2012) 146–152. [6] A. Sanger, P.K. Jain, Y.K. Mishrac, R. Chandraa, Palladium decorated silicon carbide nanocauliflowers for hydrogen gas sensing application, Sensors Actuat. B-Chem. 242 (2016) 694-699. [7] M. Tonezzer, T. T. L. Dang, N. Bazzanella, V. H. Nguyen, S. Iannotta, Comparative gas-sensing performance of 1D and 2D ZnO nanostructures, Sensor Actuat. B-Chem. 220 (2015) 1152–1160. [8] X. Jiang, Y. Wang, T. Herricks and Y. Xia, Ethylene glycol-mediated synthesis of metal oxide nanowires, J. Mater. Chem. 14 (2004) 695-703. [9] Y. Wang, X. Jiang, and Y. Xia, A Solution-Phase, Precursor route to polycrystalline SnO2 nanowires that can be used for gas sensing under ambient conditions, J. Am. Chem. Soc. 125 (2003) 16176-16177. [10] J.P. Cheng, Jiao Wang, Q.Q. Li, H.G. Liu, Y. Li, A review of recent developments in tin dioxide composites for gas sensing application, Journal of Industrial and Engineering Chemistry 44 (2016) 1–22. [11] S.J. Patil, A.V. Patil, C.G. Dighavkar, K.S. Thakare, R.Y. Borase, S.J. Nandre, N.G. Deshpande and R.R. Ahire, Semiconductor metal oxide compounds based gas sensors: A literature review, Front. Mater. Sci. 9 (2015) 14-37. [12] B. Yuliarto, G. Gumilar, and N.L.W. Septiani, SnO 2 Nanostructure as Pollutant Gas Sensors: Synthesis, Sensing Performances, and Mechanism, Advances in Materials Science and Engineering 2015 (2015) 694823. [13] R. Kumar, O. Al-Dossary, G. Kumar, A. Umar, Zinc Oxide Nanostructures for NO2 Gas–Sensor Applications: A Review, Nano-Micro Lett. 7 (2015) 97-120. [14] E.S. Hassan, A.A. Saeed, A.K. Elttayef, Doping and thickness variation influence on the structural and sensing properties of NiO film prepared by RF-magnetron sputtering, J. Mater. Sci.: Mater. Electron. 27 (2016) 1270– 1277. [15] Y. Lu, Y.H. Ma, S.Y. Ma, W.X. Jin, S.H. Yan, X.L. Xu, X.H. Jiang, T.T. Wang, H.M. Yang, H. Chen, Z. Qiang, Synthesis of cactus-like NiO nanostructure and their gas-sensing properties, Mater. Lett. 164 (2016) 48–51. [16] R. LoNigro, S. Battiato, G. Greco, P. Fiorenza, F. Roccaforte, G. Malandrino, Metal Organic Chemical Vapor Deposition of nickel oxide thin films for wide band gap device technology, Thin Solid Films 563 (2014) 50–55. [17] M-W. Lin, K-C. Wang, J-H, Wang, M-H. Li, Y-L. Lai, T. Ohigashi, N. Kosugi, P. Chen, D.H. Wei, T.F. Guo, Y-J. Hsu, Improve Hole Collection by Interfacial Chemical Redox Reaction at a Mesoscopic NiO/CH3NH3PbI3 Heterojunction for Efficient Photovoltaic Cells, Adv. Mater. Interfaces 3 (2016) 1600135. [18] D. Xu, K. Li, Y. Zhou, Y. Gao, D. Yan, S. Xu, The effect of NiO addition on the grain boundary behavior and electrochemical performance of Gd-doped ceria solid electrolyte under different sintering conditions, J. Eur. Ceram. Soc. 37 (2017) 419-425. [19] B. Wu, M-Q. Sheng, S-N. Gao, Y. Wang, F. Liao, Single-source precursor to Ag/NiO composite for rechargeable charge storage, J. Alloy Compd. 692 (2017) 34-39. [20] R-Y. Miao, W. Zeng, Q. Gao, SDS-assisted hydrothermal synthesis of NiO flake-flower architectures with enhanced gas-sensing properties, Appl. Surf. Sci. 384 (2016) 304-310. [21] M. Tonezzer, D.T.T. Le, T. Q. Huy, S. Iannotta, Dual-selective hydrogen and ethanol sensor for steam reforming systems, Sensor Actuat. B-Chem. 236 (2016) 1011–1019. [22] N. Kaur, E. Comini, D. Zappa, N. Poli, G. Sberveglieri, Nickel oxide nanowires: vapor liquid solid synthesis and integration into a gas sensing device, Nanotechnology 27 (2016) 205701. [23] W-C. Huang, J-D. Kim, Nickel oxide nanoparticle-based method for simultaneous harvesting and disruption of microalgal cells, Bioresource Technol. 218 (2016) 1290-1293. [24] H-C. Pang, B-Q. Guan, W-W. Sun, Y. Wang., Metal-Organic-Frameworks Derivation of Mesoporous NiO Nanorod for High-Performance Lithium Ion Batteries, Electrochim. Acta, 213 (2016) 351-357.
[25] C. Liu, C-L. Li, K. Ahmed, Z. Mutlu, C.S. Ozkan, M. Ozkan, Template Free and Binderless NiO Nanowire Foam for Li-ion Battery Anodes with Long Cycle Life and Ultrahigh Rate Capability, Sci. Rep. 6 (2016) 29183. [26] J. Min, J. Liu, M. Lei, W-J. Wang, Y-K. Lu, L-Y. Yang, Q. Yang, G. Liu, N. Su, Self-Assembly of Parallelly Aligned NiO Hierarchical Nanostructures with Ultrathin Nanosheet Subunits for Electrochemical Supercapacitor Applications, ACS Appl. Mater. Inter. 8 (2016) 780-791. [27] A-F. Liu, H-W. Che, Y-Z. Mao, Y-Q. Wang, J-B. Mu, C-X. Wu, Y-M. Bai, X-L. Zhang, G-S. Wang, Templatefree synthesis of one-dimensional hierarchical NiO nanotubes self-assembled by nanosheets for high-performance supercapacitors, Ceram. Int. 42 (2016) 11435-11441. [28] Z.P. Wei, M. Arredondo, H.Y. Peng, Z. Zhang, D.L. Guo, G.Z. Xing, Y.F. Li, L.M. Wong, S.J. Wang, N. Valanoor, T. Wu, A template and catalyst-free metal-etching-oxidation method to synthesize aligned oxide nanowire arrays: NiO as an example, ACS Nano 4 (2010) 4785–4791. [29] Y. Lin, T. Xie, B. Cheng, B. Geng, L. Zhang, Ordered nickel oxide nanowire arrays and their optical absorption properties, Chem. Phys. Lett. 380 (2003) 521–525. [30] K. Nielsch, F. Müller, A.-P. Li, U. Gösele, Uniform nickel deposition into ordered alumina pores by pulsed electrodeposition, Adv. Mater. 12 (2000) 582–586. [31] H. Pang, Q. Lu, Y. Zhang, Y. Li, F. Gao, Selective synthesis of nickel oxide nanowires and length effect on their electrochemical properties, Nanoscale 2 (2010) 920–922. [32] A. Kolmakov, Y. Zhang, G. Cheng and M. Moskovits, Detection of CO and O2 using tin oxide nanowire sensors, Adv. Mater. 15 (2003) 997-1000. [33] B. Miao, W. Zeng, L. Lin, S. Xu, Characterization and gas-sensing properties of NiO nanowires prepared through hydrothermal method, Physica E 52 (2013) 40–45. [34] F. Röck, N. Barsan, U. Weimar, Electronic Nose: Current Status and Future Trends, Chem. Rev. 108 (2008) 705725. [35] E.N. Carmona, V. Sberveglieri, A. Ponzoni, V. Galstyana, D. Zappa, A. Pulvirenti, E. Comini, Detection of food and skin pathogen microbiota by means of an electronic nose based on metal oxide chemiresistors, Sensor Actuat. B-Chem. 238 (2017) 1224. [36] T.T.L. Dang, M. Tonezzer, V.H. Hieu, Hydrothermal Growth and Hydrogen Selective Sensing of Nickel Oxide Nanowires, J. Nanomater. 2015 (2015) 785856. [37] A. Katoch, S-W. Choi, H.W. Kim, S.S. Kim, Highly sensitive and selective H2 sensing by ZnO nanofibers and the underlying sensing mechanism, J. Hazard. Mater. 286 (2015) 229–235. [38] G.K. Mani, J.B.B. Rayappan, A highly selective and wide range ammonia sensor—Nanostructured ZnO:Co thin film, Mater. Sci. Eng. B 191 (2015) 41–50. [39] J-M. Choi, J-H. Byun, S.S. Kim, Influence of grain size on gas-sensing properties of chemiresistive p-type NiOnanofibers, Sensor Actuat. B-Chem. 227 (2016) 149–156. [40] C. Wang, T. Wang, B. Wang, X. Zhou, X. Cheng, P. Sun, J. Zheng, G. Lu, Design of α-Fe2O3 nanorods functionalized tubular NiO nanostructure for discriminating toluene molecules, Sci. Rep. 6 (2016) 26432. [41] H-S. Woo, C.W. Na, I.-D. Kim, J-H. Lee, Highly sensitive and selective trimethylamine sensor using onedimensional ZnO–Cr2O3 hetero-nanostructures, Nanotechnology 23 (2012) 245501. [42] A. Sanger, A. Kumar, A. Kumar, R. Chandra, Highly sensitive and selective hydrogen gas sensor using sputtered grown Pd decorated MnO2 nanowalls, Sensor Actuat. B-Chem. 234 (2016) 8-14. [43] E. Oh, H-Y. Choi, S-H. Jung, S. Cho, J.C. Kim, K-H. Lee, S-W. Kang, J. Kim, J-Y. Yun, S-H. Jeong, Highperformance NO2 gas sensor based on ZnO nanorod grown by ultrasonic irradiation, Sens. Actuat. B-Chem. 141 (2009) 239–243. [44] B. Urasinska-Wojcik, T.A. Vincent, M.F. Chowdhury, J.W. Gardner, Ultrasensitive WO3 gas sensors for NO2 detection in air and low oxygen environment, Sens. Actuat. B-Chem. 239 (2017) 1051–1059. [45] A. Gaiardo, B. Fabbri, A. Giberti, V. Guidi, P. Bellutti, C. Malagù, M. Valt, G. Pepponi, S. Gherardi, G. Zonta, A. Martucci, M. Sturaro, N. Landini, ZnO and Au/ZnO thin films: Room-temperature chemoresistive properties for gas sensing applications, Sens. Actuat. B-Chem. 237 (2016) 1085–1094.
Biographies
Matteo Tonezzer graduated “cum laude” in Physics of the Matter and received hisPhD degree “with honor” from the Faculty of Physics at the University of Trento, Italy, in 2011. His thesis was the optimization of inorganic and organic nanostructured materials toward gas sensing. In 2011, he won the Young Scientist Award from the European Materials Research Society (EMRS). He worked in research centers in France (ESRF), Brazil (UFMG), Vietnam (HUST), and USA (GaTech). He is currently working for IMEM at the Italian National Research Council. Major research interest is synthesis and characterization of nanostructured materials. Dang Thi Thanh Le obtained her MSc and PhD degrees in Materials Science from International Training Institute for Materials Science (ITIMS)-Hanoi University of Science and Technology (HUST), Hanoi, Vietnam in 2001 and 2011. She worked as a visiting postdoc at The Angstrom Laboratory-Uppsala University, Sweden in the academic year of 2011–2012. She is working as a researcher/lecturer at ITIMS. Her current interests include synthesis, characterization and application of nanomaterials for gas-sensing and bio-sensing. Tran Quang Huy received his PhD in Materials Science at Hanoi University of Science and Technology in 2012. He is responsible for the Nanobiomedicine Group at the National Insitute of Hygiene and Epidemiology in Hanoi and principal investigator of Nano Energy & NanoBiomedical Research Group. He coordinated several national and international projects. He published more than 40 articles in peer-reviewed journals and contributed 6 chapters for other books. He is a reviewer for several international scientific journals, editorial board member of the Vietnam Journal of Preventive Medicine. His current research interests are bionanomaterials, biological ultrastructures and technology solutions for environmental treatments. Salvatore Iannotta graduated in Physics at the University of Bologna, then in August1984 the PhD in Chemistry at the (GWC) (Guelph Waterloo Centre for Graduate Work in Chemistry-Ontario Canada). He is presently Director of CNR-IMEM of materials for electronic and magnetism. Major research interests are: synthesis and characterization of nanostructured and molecular materials; supersonic beams of metallic and semiconductor clusters as well as of molecular aggregates, organic electronics, gas a VOC sensors, memristive materials devices and systems, active biosensors and bio-electronics. He is author of more than 130 papers and invited speaker to a wide number of international conferences.