Hydrothermally Synthesized ZnSnO3 Nanoflakes Based Low-Cost Sensing Device for High Performance CO2 Monitoring

This work reports a room temperature operative ZnSnO3 nanoflakes-based CO2 gas sensor. The perovskite ZnSnO3 nanoflakes are synthesized by a one-pot hydrothermal technique. The prepared material was characterized via XRD, SEM, UV-visible spectroscopy, and DLS measurement for confirming the crystal structure, surface morphology, optical properties, and size distribution. The X-ray diffraction pattern revealed that ZnSnO3 was in the orthorhombic phase and average crystallite size examined by the Scherrer formula was 8.05 nm. Optical studies were done by the UV–vis spectroscopy and a direct optical band gap was found to be 3.27 eV. The surface morphology of ZnSnO3 was found to nanoflakes are almost uniform dimensions. The fabricated sensor device of ZnSnO3 detected the CO2 gas at room temperature (RT) for different concentrations. The best sensor response was found to be 4.93 for 1000 ppm of CO2 whereas at 200 ppm the response and recovery times were found to be 5.92 s and 7.23 s respectively. HOMO-LUMO gap energy of ZnSnO3 without and with interaction from CO2 molecule was found 1.165 eV and 1.577 eV, respectively. DFT studies are used for a better understanding of sensing mechanisms.

Clean air is the most necessary thing after water (H 2 O) for human beings to be healthy but human activities associated with socioeconomic developments are the key points for the sources of pollution. 1 Carbon dioxide (CO 2 ) is a necessary component of the photosynthesis process in plants. It is transformed into O 2 and glucose molecules as a consequence the life of the human becomes possible on Earth. 2,3 CO 2 is a greenhouse gas, which is colorless and odorless. The acceptable concentration of CO 2 is beyond a special limit be below 600 ppm. Whenever the concentration level is above a specific limit, it may affect human health. The threshold limit of CO 2 concentration proposed by the American Society of Heating, Refrigerating, and Air Conditioning Engineers (ASHRAE) in occupied buildings is not higher than 1000 ppm. 4 The concentration of CO 2 is regularly increasing in the ambient atmosphere due to deforestation and the burning of hydrocarbons. As on 17 September 2021, the CO 2 gas concentration in our atmosphere was 413.05 ppm which became a worldwide concern. 5-9 So, the monitoring of CO 2 is necessary for the safety of human health and environmental protection.
Over the years, a variety of sensors have been created based on the principles of work function, capacitive, optical, acoustical, electrochemical, 10 gas chromatography (GC), mass spectrometry (MS), and multiple electrodes. 2,11,12 However, these devices have a number of drawbacks in comparison to the chemi-resistive sensing device, including high maintenance costs (GC and MS), complex structures, and limited device lifetimes (optical and electrochemical). 13,14 The chemi-resistive sensors are developed by numerous nanostructured materials including metal oxide semiconductor (MOS), graphene, conducting polymers, carbon nanotubes, and metal-organic frameworks. [15][16][17][18][19] In comparison to MOS, other nanostructures such as carbon nanotubes are relatively expensive and have poor selectivity. 20,21 Graphene-based CO 2 gas sensors are prone to conglomeration between layers, have a low yield, and have high prices. Also, conducting polymers-based CO 2 gas sensors possess low stability, long recovery time, and poor mechanical strength. 22 Thus, MOS is the most identified group for monitoring CO 2 due to their exceptional properties like easy fabrication, costeffectiveness, high stability, and miniaturized versatile morphologies, etc. 23,24 The MOS devices are used successfully in many areas like climate change, security processes, and medical equipment such as breath analyzers. 25,26 Various studies have been reported on the MOS chemiresistive type CO 2 gas sensors, for example, Kannan et al. fabricated the ZnO thin films and checked the CO 2 gas sensing properties for 1000 ppm at 300°C. Here, 101% sensor response was found with 20 s of response time. 27 29 Jeong et al. studied La@ZnO nanomaterial-based CO 2 gas sensor for 5000 ppm concentration at 400°C temperature. 30 In most cases, MOS-based CO 2 gas sensors were operated at high temperatures. However, further improvement in sensor technology has resulted in today's RT CO 2 sensor. Sonkar et al. fabricated titanium dioxide and poly aniline nanocomposite thin film and detected CO 2 at 30°C. In this work maximum sensor response was found to be 53, whereas response and recovery time were found to be 552 s and 342 s respectively. 31 Similarly, M. Amarnath et al. synthesized NiO-In 2 O 3 coated reduced graphene oxide by wet chemical method and reported 40% sensor response for CO 2 gas with 6 s response time and 18 s recovery time. 5 From the literature, it can be observed that most sensing materials show low sensor response with high response and recovery times. So, it is necessary to develop highly sensitive and selective CO 2 gas sensors operated at low temperatures. In the present work, we synthesized the ZnSnO 3 by one-pot hydrothermal method and detected CO 2 gas at RT (30°C and 56%RH) between 200-1000 ppm concentration. ZnSnO 3 is a perovskite compound having a wide bandgap of 3.72 eV and it has n-type semiconducting nature and high thermal stability up to 1073 K. ZnO-SnO 2 is a binary system known to have two compounds: ZnSnO 3 and Zn 2 SnO 4 between which ZnSnO 3 is better because of its perovskite nature and unique physical, electrical, and chemical characteristics like high mobility, high electrical conductivity, good optical property, and good stability. 32 ZnSnO 3 has been synthesized by different techniques like sol-gel, hydrothermal, co-precipitation, laser ablation, electrospinning etc. by different researchers across the globe. 33,34 Hydrothermal method is the easiest, low-temperature, versatile and economical route. It is widely used for the preparation of nanostructure with metal doping, functionalization, heterostructures, and phase-controlled with high purity. Also, for the permeation of a particular morphology and composition z E-mail: ajeet.bbau.2018@gmail.com; balchandra_yadav@rediffmail.com hydrothermal is the best approach for the preparation of nanomaterials. 35 Hydrothermal synthesis of ZnSnO 3 nanoflakes.-The pure perovskite ZnSnO 3 nanoflakes synthesized by a one-step hydrothermal process are shown in Fig. 1. In this process, 0.02 M CTAB and 0.1 M zinc acetate dihydrate were mixed in 25 ml distilled water and magnetically stirred at RT for 30 min. Again, in a separate beaker 0.02 M CTAB and 0.1 M tin (IV) chloride pentahydrate were mixed in 25 ml distilled water via the same process. Here CTAB worked as a capping agent. Further, both solutions were mixed with continuous stirring for 30 min and formed a uniform solution. Later, NaOH aqueous solution was dropped to maintain the pH (8)(9). After that, the solution was transferred into an autoclave (100 ml) and maintained at 180°C temperature for 20 h in the oven. The autoclave was naturally cooled at RT. The obtained precipitate was centrifuged at 3500 rpm and washed with distilled water and ethanol 3 times. Further, the precipitate was dried at 70°C for 5 h and then annealed at 450°C for 2 h in a furnace.
Sensor fabrication and measurement.-The sensing film of ZnSnO 3 was fabricated via the spin coating technique on the glass substrate (2 cm × 1 cm). The synthesized ZnSnO 3 powder was dissolved in isopropyl alcohol via ultrasonication. The solution of ZnSnO 3 was spin coated on a glass substrate at 2000 rpm for 60 s and dried at 60°C for 5 min on a hot plate. This process was repeated three times at the same condition for accomplishing the required thickness. Further, the film was annealed at 300°C for 1 h. Finally, silver contacts were made on the film and kept for drying at 80°C.
To obtain CO 2 sensing characteristics, a sensing set-up was selfdesigned which is already described in our previous work depicted in Fig. 2. 20,38 The CO 2 gas sensing measurement was carried out at RT (56 %RH and 30°C) for various concentrations (200-1000 ppm). Before the measurement of CO 2 sensing, the Ag electrodes were deposited on the sensing film and placed inside the chamber for 24 h to acquire saturation of air adsorption. The sensing film was placed inside the chamber and connected to the electrometer (Keithley 6517B). Also, the ends of the Keithley electrometer were connected to the computer system via GPIB (KUSB 488).
This prepared film was used for monitoring CO 2 gas at RT. The concentration of CO 2 gas inside the chamber is calculated by following relation Eq. 1.
Here, C is concentration (ppm), V is volume of sensing chamber (ml), V x is volume inside the chamber (ml), T c is operated temperature, and T r is RT.

Characterizations
The structural analysis of as-synthesized perovskite ZnSnO 3 was analyzed by X-ray diffractometer (XRD; Bruker Advance D8 instrument). The surface morphology of ZnSnO 3 film was analyzed through field-emission electron microscopy (FESEM; JEOL JSM 7610 f). Optical analysis was done by UV-visible spectrophotometer (Thermo scientific Evolution 201). The presence of various functional groups existing in ZnSnO 3 was confirmed by Fourier transorm infrared spectrometer (FTIR; Thermonicolate 6700). Particle size and the zeta potential of ZnSnO 3 was measured by Nanozetasizer (Thermo Scientific NSZS90). flakes can be more easily visualized. From the surface morphological investigation, we observed that ZnSnO 3 possess a highly porous nanoflakes sheets which enhanced the adsorption/desorption of analytes to the sensing surface because of the availability of more active sites. These properties enhance the sensing performance and various studies have been reported on the improvement of sensing performance due to perovskite nanostructure. The elemental analysis was carried out by EDX and the corresponding spectrum is shown in Fig. 3c. EDX confirmed the presence of Zn, Sn, and O elements in the ZnSnO 3 nanoflakes. Also, the elemental composition in weight% and atomic% is shown in the inset of Fig. 3c. The thickness of the sensing filmwas calculated by Scanning electron microscopy (SEM),  Structural analysis.-The structural analysis has been carried out by using X-ray diffraction in the range of 20°to 70°. The XRD pattern of ZnSnO 3 powder is depicted in Fig. 4 which is well matched withJCPDS card no. 28-1486 and revealed the orthorhombic phase (space group Pnma). 39 The peaks of the ZnSnO 3 appear at 26.44°, 31.61°, 34.06°, 36.29°, 51.66°, 57.61°, and 65.54°c orrespond to diffraction planes as (110), (012), (104), (015), (116), (214), and (036) respectively. The crystallite size of the ZnSnO 3 was calculated by Scherrer formula, 40 and the average crystallite size was found to be 8.05 nm. When the crystalline dimension of the material goes below 20 nm, the nanostructure dimension allows all atoms to be within the Debye length of the material. Materials may be highly reactive for the adsorption of analytes and sensor response increased significantly.

Results and Discussion
Optical analysis.-The UV-visible absorption spectrum of ZnSnO 3 is depicted in Fig. 5a, which was used for the calculation of absorption coefficient (α) and optical band gap (E g ). Here, α can be calculated by using absorbance of photon (A) by Beer-Lambert law given in Eq. 2.
where t is thickness of cuvette and A is absorbance of material. Further, optical bandgap energy (E g ) by was determined by the extra linear plot between (αhv) 2 vs hv. 41,42 The optical energy band gap of ZnSnO 3 was found to be 3.27 eV.
Dynamic light scattering (DLS) analysis.-The DLS method is usually used to find out nanoparticle size distribution in the range of nano tomicrometer in colloidal suspension. In the DLS, the motion of the nanoparticles is considered Brownian motion. Brownian motion is the random motion of the nanoparticle which is normally observed under a high-power ultra-microscope. DLS measures the scattered light as the function of time following Stokes-Einstein assumptions, which is used to determine the nanoparticle hydrodynamic diameter. Nanoparticle hydrodynamic diameter is the diameter of nanoparticle and solvent molecules that are diffused at the same rate as the colloidal solution. There are many advantages of DLS which includes its quick, easy, and accurate operation for the analysis of colloidal suspensions. 43,44 The radius of a hypothetical sphere that diffuses at the same rate as the particle under study is known as the hydrodynamic radius (R h  Here k B = Boltzmann constant, T = absolute temperature, D τ = diffusion coefficient and η = viscosity of the medium. The particle size distribution of ZnSnO 3 is shown in Fig. 5b in which the particles are distributed in the range of 80-105 nm whereas, average particle size was estimated as 91.35 nm. The Zeta potential of the ZnSnO 3 colloidal suspension was found as 5.28 mV (−50 mV to +50 mV) whereas the conductivity value was 0.0149 mS cm −1 which is shown in Fig. 5b.
CO 2 gas sensing of ZnSnO 3 .-The prepared ZnSnO 3 film was exposed to CO 2 gas, and electrical resistance was measured with time for various CO 2 concentrations at RT (56%RH and 30°C) using an electrometer. The sensing characteristics of ZnSnO 3 film were examined by sensor response, response time, and recovery time. Generally, the sensor response of n-type materials for oxidizing gas is the ratio of resistance in presence of gas to air (S = R g /R a ). 46,47 To investigate the gas sensing characteristics of ZnSnO 3 film, different concentrations of CO 2 gas were exposed to the sensing film and the corresponding sensing curves were plotted as shown in Fig. 6a. The sensing characteristics curves of ZnSnO 3 film recorded at 200, 400, 600, 800, and 1000 ppm concentration of CO 2 . When the sensing film was exposed to the CO 2 gas, the electrical resistance suddenly increased. At lower concentration variation in resistance was low whereas at higher concentration variation in resistance wass high as shown in Fig. 6a. As shown in Fig. 6b for 200 ppm, the sensor response, response time, and recovery time was evaluated. The sensor response at 200 ppm was found to be 1.36, whereas the maximum sensor response was found to be 4.93 under 1000 ppm. The sensor response curve with concentrations shown in Fig. 6c, indicates that the sensor response linearly increases as concentration increases. The sensitivity of the ZnSnO 3 sensor was determined by the slope of linear fitted curve between sensor response vs concentration and it was found as 0.0044 sensor response/concentration (ppm) as shown in Fig. 6c. The linearity coefficient was found 0.993 which indicated the sensing film increases the sensor response linearly which is respectable for development of effective gas sensor.
The response and recovery times of the sensor are crucial parameters and they can be defined as follows; the 90% time of the sensor to reach a maximum resistance of the film on the exposure of CO 2 is called response time. Likewise, 90% time taken by the sensor when it comes back to the initial stage is called recovery time. 48 The response and recovery time towards 200 ppm CO 2 concentration was observed as 5.92 and 7.32 s, respectively as shown in Fig. 5d. The response and recovery time curve indicates that as the concentrations of CO 2 increases, response and recovery times linearly increase. For higher concentrations, the interaction between gas and sensing film may be increased and take a longer time to acquire saturation in comparison to lower concentrations. The calculated sensing parameters towards each concentration are presented in Table I. The ZnSnO 3 sensing film was found to highly sensitive and fast response and recovery time were observed from the previously reported work as shown in Table II.
In a gas sensor selectivity of a sensor is the significant parameter to predict that the sensor becomes highly sensitive toward a specific gas. 49 In this study, the selectivity was examined inside a chamber with 200 ppm concentrations of LPG, ethanol, ammonia, acetone, and other target gases. Figure 7a displays the selectivity curves of the ZnSnO 3 sensing film. The sensor response was determined to be 1. 04, 1.07, 1.11, 1.17, and 1.36 for ethanol, ammonia, acetone, LPG,  and CO 2 , respectively. The perovskite ZnSnO 3 is an n-type semiconducting material and it was found to be highly sensitive to CO 2 in comparison to ethanol (1.30 times), acetone (1.27 times), ammonia (1.22 times), and LPG (1.16 times).
The effect of humidity on the sensing film is also an important characteristic; we executed the experiment at 200 ppm CO 2 in presence of humidity. The saturated solution of K 2 SO 4 was used for the increasing humidity which was measured by a hygrometer.  When humidity is low (10%-30%RH) sensor response is high, whereas in the mid humidity range (30%-60%RH) the sensor response is almost same and minute change was observed in the sensor response as shown in Fig. 7b. When humidity was high (60%-90%), the sensor response decreased because the moisture weaken the interaction between CO 2 molecules and sensing surface and the sensor becomes inefficient. So, the ZnSnO 3 sensing film is highly efficient and stable in mid-humidity regions. Reproducibility is also an important parameter to define the capability of a sensor to repeat a similar measurement. The reproducibility graph for 200 ppm is shown in Fig. 7c through five cycles. From Fig. 7c, it can be observed that the almost same measurement in every cycle was obtained. To check the long-term stability of ZnSnO 3 sensing film, sensing characteristics for 200 ppm after a time interval were observed. Fig. 7d shows the long-term stability curve with an error bar and it can be observed that minute changes were found in the sensor response. When the number of days increases, the error in sensor response increases because the moisture layer experienced on sensing film weakens the power of adsorption of CO 2 gas molecules. CO 2 gas sensing mechanism.-Many theories and phenomena, like the Electronic Depletion Layer (EDL), the Hall Accumulation Layer (HAL), the Bulk Resistance Control Theory, etc., have been used to describe the gas sensing mechanism of a gas sensor. But mostly, the MOS gas sensing mechanism is explained by the phenomenon of adsorption/desorption. Electrical resistance typically increases when n-type semiconductors are exposed to oxidizing gases like CO 2 and NO 2 , whereas resistance decreases when exposed to reducing gases (ammonia, acetone, ethanol, etc.), and reverses for p-type semiconductors. The gas sensing mechanism of ZnSnO 3 perovskite is shown in Fig. 8. Prior to exposure to CO 2  Because of the formation of oxygen species on the surface, an electronic depletion layer is created, establishing a depletion region that acts as a potential barrier. Due to oxidizing nature of CO 2, it interacts with oxygen species, transports electrons to CO 2 molecules, and forms metastable CO 3 2− complexes on the surface of the sensing film as following Eq. 8. 6,8 The barrier height and width increase due to the formation of CO 3 2− complex, so the resistance of sensing film abruptly increases and is dependent on the CO 2 concentration.
As the flow of CO 2 is turned off, the CO 3 2− complexes get convert into CO 2 and the trapped electrons are released back to the conduction band of the sensing film and acquire its initial state.
DFT studies.-Density-functional theory (DFT) is a computational quantum mechanical modeling method used in physics, chemistry, and materials science to examine the electronic structure (or nuclear structure) of many-body systems, particularly atoms, molecules, and condensed phases. 55 We used Gaussian 09 program for the optimization and Gauss View to design the molecular structures. 56 For these calculations, we used the B3LYP method  with the basis set of LanL2DZ. Different electronic parameters such as HOMO-LUMO energy gap, electron affinity, ionization potential, electronegativity, and chemical potential were evaluated for the ZnSnO 3 and ZnSnO 3 interacted with CO 2 gas as given in Eqs. 9-13 and these parameters are tabulated in Table III. 57 Optimized structures of ZnSnO 3 and ZnSnO 3 -CO 3 are depicted in Figs. 9a and 9b and HOMO-LUMO orbitals are shown in Fig. 10. In the sensing measurement after the interaction of CO 2 gas resistance of the sensing device increase, whereas in the DFT measurements, it can also be seen that after the interaction of the CO 2 with the ZnSnO 3 the HOMO-LUMO energy gap increases from 1.165 eV to 1.577 eV as shown in Fig. 10. The materials reactivity with analytes can be defined by the adsorption energy. In the present work, DFT analysis has been carried out for optimization of ZnSnO 3 without interaction and with interaction from CO 2 , ethanol, and ammonia. For the calculation of adsorption energy of analytes Eq. 14 was used. 58 The adsorption energies of CO 2 , ethanol, and ammonia were found to −2.41, −1.54, and −0.61, respectively. The adsorption energy of CO 2 is much more negative in comparison to ethanol and NH 3 , implying the strong interaction between ZnSnO 3 and CO 2  molecules. The DFT analysis and sensing performance are showing the same nature of material.

Conclusion
The perovskite ZnSnO 3 nano-flakes were synthesized by a onepot hydrothermal route and thin film was fabricated by spin coater on the glass substrate. XRD analysis confirmed the orthorhombic phase of ZnSnO 3 and the average crystallite size was found as 8.05 nm. The direct optical bandgap of ZnSnO 3 was found as 3.75 eV. The surface morphology of ZnSnO 3 nanoflakes was confirmed by FESEM and elemental analysis by EDAX. The ZnSnO 3 nanoflakes film detects the CO 2 gas below threshold concentration (200-1000 ppm) and the maximum sensor response was found 4.93 for 1000 ppm CO 2 , whereas at 200 ppm, the response and recovery time were found to be 4.92 s and 7.32 s, respectively. Hence the ZnSnO 3 sensor is moderately appropriate for developing a commercial CO 2 gas sensor operating at room temperature. Also, DFT studies reveal that after the interaction of the CO 2 with the ZnSnO 3, the HOMO-LUMO energy of gas increases from 1.165 eV to 1.577 eV.