Flexible ultra-sensitive and resistive NO2 gas sensor based on nanostructured Zn(x)Fe(1−x)2O4

Low concentration gas detection, rapid response time and low working temperature are anticipated for a varied range of toxic gas detection applications. Conversely, the existing gas sensors suffer mostly from a high working temperature along with a slow response at low concentrations of analytes. Here, we report an ultrasensitive flexible nanostructured Zn(x)Fe(1−x)2O4 (x = 0.1, 0.5 and 0.9) based chemiresistive sensor for nitrogen dioxide (NO2) detection. We evince that the prepared flexible sensor Zn(0.5)Fe(0.5)2O4 has detection potential as low as 5 ppm at a working temperature of 90 °C in a short phase. Further, the Zn(0.5)Fe(0.5)2O4 sensor exhibits excellent selectivity, stability and repeatability. The optimized sensor sensing characteristics can be helpful in tremendous development of foldable mobile devices for environmental monitoring, protection and control.


Introduction
There is a high interest in the discovery of novel nanomaterials in order to develop rapid response and extremely sensitive solidstate gas sensors. Semiconductor metal oxide materials are an alternative to conventional sensing materials due to their exceptional characteristics of easy synthesis, cost-effectiveness and low power consumption. These materials are the best candidates to detect poisonous, toxic, ammable, explosive and harmful gases. The gas sensor working principle involves the surface adsorbed atmospheric oxygen species interaction with analyte gas molecules, leading to redox reactions on the semiconductor, such as ZnO, 1,2 WO 3 , 3,4 In 2 O 3 , 5,6 SnO 2 (ref. [7][8][9][10] and Fe 2 O 3 (ref. 11,12) (n-type), and NiO 13 and CoO 3 O 4 (ref. 14) (p-type) gas sensors. In sensing application crucial roles are performed by properties regarding oxides of semiconductor like pores, grain size, crystalline size, lm thickness, layers, and surface to volume ratio. 15,16 Furthermore, if the prepared sensing material is porous, the targeted analyte molecules can easily penetrate the material and react with the total volume of the material by enhancing the sensor response tremendously. 17 Thus, sufficient attention has been concentrated on controlling the structural and morphological parameters of nanomaterials with high-energy surfaces 18,19 and decreased crystalline size, 20 which are gaining special attention, such as oxide composites, core-shell heterostructure nanotubes, 21,22 3D structures 23,24 and doping. 25 Spinel ferrites are the basic functional material used in a variety of cutting-edge technological applications because it is exceptionally good catalyst and has simple synthesis. Additionally it is very economical and eco-friendly in nature. 26 ZnFe 2 O 4 has been widely used with lithium-ion batteries as the anode materials for the past few years. Nanostructured ZnFe 2 O 4 is a gas sensing material with rapid response and excellent selectivity towards oxidizing and reducing gases. The scientists working on the intrinsic association between shape, structure and gas sensing characteristics have produced essential adaptable synthetic strategies, where these properties of ZnFe 2 O 4 can be tailored with designed functionalities. In this regard, the preparation of nanostructured ZnFe 2 O 4 with exclusive microstructures is escalating its possible gas sensor applications.
In the present paper, a simple sol-gel auto combustion method was used to synthesize nanostructured Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.1, 0.5 and 0.9). 27 As a result, a large specic surface area pore size was exhibited by the prepared ZnFe 2 O 4 materials. A exible device was fabricated by using a simple drop drying technique. Additionally, NO 2 gas sensing characteristics were investigated with various working temperatures. We proved that a nanostructured exible Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.5) based sensor shows a high response at an operating temperature of 90 C, with excellent selectivity, good stability and reproducibility.

Experimental section
2.1 Synthesis of nanostructured Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.1, 0.5 and 0.9) A sol-gel-auto combustion technique was employed to prepare nanostructured Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.1, 0.5 and 0.9). 27 In this method, the exothermic reaction of xerogel, which is an aqueous solution of metal nitrates (zinc nitrate and iron nitrate) and fuel (glycine) was carried out. All the reagents were of analytical grade from Sigma-Aldrich, USA. An appropriate amount of nitrates and fuel were dissolved in distilled water under constant stirring at 80 C according to the stoichiometric composition of the fuel to oxidizer ratio. Aer 25-30 min a brown colored thick gel was formed. The obtained solution was placed on a hot plate at 180 C to initiate the combustion, then it was ignited to form a lightweight powder and annealed at 650 C for 5 h.

Characterization
The structural analysis of the nanostructured Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.1, 0.5 and 0.9) powder was made with a Bruker-D8 X-ray diffractometer (XRD) using Cu Ka 1 radiation. The optical property of absorbance was calculated by a UV-visible double beam spectrophotometer (Systronic-2203). Fourier transform infrared spectroscopy (FT-IR) (PerkinElmer L160000A) in the wavelength range of 500-4000 cm À1 was also used for the structural elucidation. The morphology and elemental composition were observed by a Carl Zeiss (Merlin compact 60-27) eld emission scanning electron microscope (EDX and FESEM). Particle size and morphology were further conrmed by transmission electron microscopy (TEM) (Philips, Holland TEM instrument) operated at an accelerating voltage of 120 kV. Resistance and voltage were measured using the Keithley multimeter (2750).

Device fabrication and construction of in-house sensor testing unit
The samples were coated with a drop drying method on exible electrodes prior to testing, which can be described as follows. Initially, an approximate amount of the as-prepared Zn (x) Fe (1Àx) 2 O 4 (x ¼ 0.1, 0.5 and 0.9) nanostructured powder was mixed with dimethylformamide to prepare the homogeneous paste and coated onto exible pre-patterned interdigitated electrodes (IDEs). This was then allowed to dry at room temperature, and the device was calcinated at 150 C for 3 h to enhance its stability. Finally, the fabricated device was connected to a Keithley multimeter (2750) in an in-house dynamic gas sensing setup. The sensing examination of the developed gas sensor was examined by a sensor testing unit, which was explained comprehensively in our earlier paper. 28 The construction of the in-house sensing setup is schematically demonstrated in Fig. 1. The sensor response was calculated as S ¼ (R a À R g )/R a for oxidizing gases or (R g À R a )/R g for reducing gases, where R a is the resistance value in absence of air and R g is the resistance in the presence of the analyte gas. The sensor was analyzed in both at and bending position to demonstrate the exibility of the sensor. A bending angle of 60 was used to determine the exibility.
agreement with the standard JCPDS no. 89-1012, which shows the product is highly pure and has no other impurities. This indicates that metal nitrates were fully transformed into ZnFe 2 O 4 at 650 C. Whereas in the case of the Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.1 and 0.9) materials, due to their compositional variations, a slight peak shi was noticed (Fig. 2a). Nanostructured Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.5) deection peaks were comparatively broadened, indicating its small crystallite size. The average crystal sizes of nanostructured Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.1, 0.5 and 0.9) are about 21.5 nm, 16.8 nm and 18.7 nm, respectively. The crystalline size was estimated by the Debye-Scherrer formula D ¼ 0.89l/b cos q (where l ¼ 1.54060Å, q is the Bragg angle and b is the peak full width at half maximum).  The UV-vis spectra of the Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.1, 0.5 and 0.9) are shown in Fig. 2b and a strong absorption was observed for all the prepared samples. Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.5) exhibits peak shi from UV to visible light region and reveals the nanostructured Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.5) having a higher efficiency to absorb visible light than Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.1 and 0.9). Moreover, the absorption spectrum of the Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.5) material contained both regions (UV and visible) of Zn (x) -Fe (1Àx)2 O 4 (x ¼ 0.1 and 0.9). These results show that an equal composition of precursor material causes expansion and enhancement of the photoresponse to the visible region. Hence, we observe that the photocatalytic activity of the prepared Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.5) was superior to that of Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.1 and 0.9) in visible light.
A Fourier transform infrared (FTIR) spectrophotometer was used to analyze the Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.1, 0.5 and 0.9) nanostructured materials in the range from 500 to 4000 cm À1 and the results are shown in Fig. 2c. The FTIR spectra of all three samples show the peak at 1274 cm À1 represented ]C-H inplane stretching. The peak at 923 cm À1 is from C-C out-ofplane stretching vibrations 29 and the broad band observed at 3430 cm À1 is related to O-H vibrations. Peaks at vibrations around 500 cm À1 are attributed to Fe-O and Zn-O vibrations. 30 The nanostructured Zn    Fig. 3d. The size and structure of the material were in harmony with the FESEM results. These results conrm a massive amount of particles are of a nano size and assembled to form a spherical structure. Fig. 4 represents the elemental analysis of the nanostructured Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.1, 0.5 and 0.9) materials and evidences the presence of Zn, Fe and O.

Gas sensing properties
The present study deals with the advantages of the presynthesized nanostructured materials Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.1, 0.5 and 0.9) as NO 2 sensing substances and their characteristics. In a chemiresistive gas sensor, the sensitivity mainly depends upon the operating temperature. Thus, the responses of Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.1, 0.5 and 0.9) based sensors were measured by changing the temperature from room temperature to 300 C at a constant gas concentration of 30 ppm and the related examined results are illustrated in Fig. 5a. In the gure, it is clearly shown that cone shape curves represent that the  initial NO 2 response was enhanced with operating temperature and reaches its highest value for about 90 C, and aerwards reduces slowly. The threshold temperature for the sensor was observed at 90 C. Among all three prepared samples, Zn (x) -Fe (1Àx)2 O 4 (x ¼ 0.1, 0.5 and 0.9), the maximum response was observed for the equal amount of precursor material i.e., Zn (x) -Fe (1Àx)2 O 4 (x ¼ 0.5). The obtained increase-decrease result responses can be described as follows: at room temperature, NO 2 gas molecules partially interact with the surface absorbed atmospheric oxygen molecules, which gives a lower response. Whereas, with increasing the operating temperature the rate of reaction becomes higher and there is an increase in the oxygen ions on the sensor surface from the absorbed atmospheric oxygen species (O 2(gas) / O 2(ads) / O 2(ads) À / 2O (ads) À ) responsible for the maximum response. From the result, 90 C is the optimum working temperature for the selected nanostructured Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.5). In the case of Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.1 and 0.9), the sensor response observed was 0.74 and 0.54%, respectively, for 90 C at 30 ppm and comparatively these sensitivity results were a much smaller response than the sensor response (1.41%) of the equal ratio precursor (Zn (0.5) Fe (0.5)2 O 4 ). The reason behind this kind of behavior might be due to the Zn (0.5) Fe (0.5)2 O 4 sensor having enough porosity so that NO 2 gas can easily penetrate inside the sensing material throughout the surface. The nanostructured Zn (x) -Fe (1Àx)2 O 4 (x ¼ 0.1 and 0.9) sensor material has a low porosity compared to the Zn (0.5) Fe (0.5)2 O 4 materials.
The response of the Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.1, 0.5 and 0.9) sensor variations along with the NO 2 gas dilutions was examined at an optimum working temperature (Fig. 5b). It is observed that as the gas concentration value increases slowly from 1 ppm, the sensor response increased up to 30 ppm and then saturated. This kind of sensing behavior can be estimated as follows: a very low surface interaction at a minute gas concentration leads to a low sensor response. The sensitivity increased stepwise when the gas concentration increased to create an enhanced sensor surface interaction, leading to a rising response at a critical concentration, i.e. 30 ppm. Followed by a increased NO 2 gas concentration, the sensor surface was fully covered by and therefore had no possibility to react with new molecules due to the saturation level.
The tremendous fabricated current device is solely transparent and mechanically exible. To investigate the nanostructured Zn (0.5) Fe (0.5)2 O 4 sensor, exibility properties were measured in at (red), bending (black) and aer bending (blue) conditions with a bending radius of 60 for 30 ppm NO 2 at 90 C. Fig. 6a illustrates the sensor response based on exibility and temperature variations. The response of the sensor device decreased moderately while bending due to the reduction of the sensitive surface area. Fascinatingly, the fabricated device did not deteriorate considerably aer these mechanical transformations. Moreover, the few dynamic cycles of the responserecovery of the exible sensor shown in Fig. 6b-  The nanostructured Zn (0.5) Fe (0.5)2 O 4 sensor selectivity was studied by exposing different analytes, such as benzene, carbon monoxide, acetone, toluene, LPG, isopropanol and ethyl acetate, as shown in Fig. 7a. The concentration of all the analytes was maintained constant, i.e. 30 ppm at 90 C. The sensor response against NO 2 was remarkably higher than against other analytes. This analysis provides the information regarding the high selectivity of the sensor towards NO 2 . This is due to the chemisorbed atmospheric oxygen present on the sensor surface effectively initiating a high response towards NO 2 . In the practical application of gas sensors, long-term stability is one of the essential characteristic parameters. Therefore, the response of the prepared exible Zn (0.5) Fe (0.5)2 O 4 sensor to 30 ppm of NO 2 at a temperature of 90 C was examined for 30 days as illustrated in Fig. 7b. In that period of time, the sensor response was recorded with only a little uctuation. Therefore, fabrication of attractive and promising sensor with an excellent durability is possible with this technique. These sensors are very useful in a direct industrial application. Bending tests were implemented on the sensor during the test to understand the exibility of the sensor. Aer two hundred repeats of bending of the prepared exible sensor, the device did not have a large deviation of response (Fig. 7c). To examine the repeatability of the exible sensor, the Zn (0.5) Fe (0.5)2 O 4 device were exposed to seven cycles of 30 ppm NO 2 and the dynamic resistance responses are shown in Fig. 7d. The test revealed that the sensor response is constantly maintained aer several exposure cycles. From the obtained results, the sensor shows good reproducibility for a long time, which conrms the stability of the prepared gas sensor towards NO 2 .

Gas sensing mechanism
A well-known p-type semiconductor oxide 31 such as ZnFe 2 O 4 , works as a sensor and the functioning mechanism involved is based on the change in resistance of the atmospheric oxygen molecule chemisorption on the surface of the sensing material. 32,33 ZnFe 2 O 4 contains holes as the major charge carrier. Initially, when the exible ZnFe 2 O 4 sensor is exposed to zero air, the atmospheric oxygen molecules in the form of O 2(ads) À , O (ads) À and O 2À (ads) adsorb on the surface of the sensor. The sensor resistance increases due to the formation of a thick electron space charge layer on the surface (Fig. 8a). Followed by the exposure to an oxidizing gas, i.e. NO 2 (electron accepting), molecules interact with adsorbed oxygen molecules, which leads to a decrease in resistance (Fig. 8b). The exible ZnFe 2 O 4 gas sensor performance towards NO 2 gas is compared with previously reported ZnFe 2 O 4 gas sensor for various analytes as listed in Table 1.

Conclusion
In summary, a sol-gel auto combustion method was used for the preparation of nanostructured exible Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.1, 0.5 and 0.9), which were coated on a pre-patterned exible electrode by a simple drop drying process and heated  aerwards. The synthesized material was examined as a sensing material for the possible chemiresistive gas sensing application. It was found that the equal concentration of precursor material (Zn (x) Fe (1Àx)2 O 4 (x ¼ 0.5)) used in device exhibited an ultra-high sensing performance and excellent long-term stability, selectivity and reproducibility towards 5 ppm of NO 2 at a working temperature of 90 C.

Conflicts of interest
There are no conicts to declare.