Nanometer-Thick ZnO/SnO2 Heterostructures Grown on Alumina for H2S Sensing.

Designing heterostructure materials at the nanoscale is a well-known method to enhance gas sensing performance. In this study, a mixed solution of zinc chloride and tin (II) chloride dihydrate, dissolved in ethanol solvent, was used as the initial precursor for depositing the sensing layer on alumina substrates using the ultrasonic spray pyrolysis (USP) method. Several ZnO/SnO2 heterostructures were grown by applying different ratios in the initial precursors. These heterostructures were used as active materials for the sensing of H2S gas molecules. The results revealed that an increase in the zinc chloride in the USP precursor alters the H2S sensitivity of the sensor. The optimal working temperature was found to be 450 °C. The sensor, containing 5:1 (ZnCl2: SnCl2·2H2O) ratio in the USP precursor, demonstrates a higher response than the pure SnO2 (∼95 times) sample and other heterostructures. Later, the selectivity of the ZnO/SnO2 heterostructures toward 5 ppm NO2, 200 ppm methanol, and 100 ppm of CH4, acetone, and ethanol was also examined. The gas sensing mechanism of the ZnO/SnO2 was analyzed and the remarkably enhanced gas-sensing performance was mainly attributed to the heterostructure formation between ZnO and SnO2. The synthesized materials were also analyzed by X-ray diffraction, scanning electron microscopy, energy-dispersive X-ray, transmission electron microscopy, and X-ray photoelectron spectra to investigate the material distribution, grain size, and material quality of ZnO/SnO2 heterostructures.

Introduction

It is well-known that metal oxide semiconductors are an excellent gas sensing material simply because of their high sensitivity to small concentrations of gases, long-term stability, low fabrication costs, and high potential to provide a small and low power-consuming sensor.[1−4] In order to improve the gas sensing performance of sensors containing a single metal oxide, many research works have been performed using the homojunction between grains to study the grain size effect and morphological influence on gas sensing.[5] However, due to the fact that the single metal oxides give an equal response to a wide range of gases, the selectivity of the sensor still remains a challenging issue.[6] Many efforts have been focused on improving metal oxides’ selectivity by (a) doping with noble catalytic metals to promote the reaction to a specific gas, (b) surface modification, (c) using machine learning, and (d) use of multicompositional sensing films.[7,8] In the case of multicompositional films, recent reports have indicated that by applying heterojunctions of dissimilar oxides in the form of surface decoration and/or as composites, the gas sensing behavior of resistive sensors is remarkably improved.[9] The main point with heterostructures made by dissimilar oxides is the formation of a depletion layer in the junction. The width of this depletion layer can be changed by the exposed target gas, as a consequence of which conductivity is altered.[10] Moreover, the oxides with different crystal structures as well as nanoparticle size may also lead to an increase in gas adsorption sites on the oxide surface and improve overall performance.[11] Therefore, heterostructures can greatly improve the sensitivity of sensors.

For gas sensing, many studies have also demonstrated that nanomaterials, for example, Si nanowires, are also good sensors.[4,12] They can be easily integrated with CMOS technology, especially in the form of diodes or transistors.[13,14] However, sensors in the form of Si nanowire transistors may show low sensitivity, and the technology used for this purpose is costly and complicated for mass production.[4,15] Therefore, resistive metal oxides are more common sensors because of their simple operation and fabrication technology.[16]

For gas sensing, there are also many applications for medical and safety purposes as well as environmental monitoring requiring robust, compact, and high-performance H2S gas sensors.[17,18] It is well-known that polluted air contains dangerous gases, for example, O3, CO2, CHa, and CO, where H2S gas is corrosive and flammable with an infamous rotten egg smell.[11,19] Exposure to even low concentrations of H2S over a long period of time can cause negative effects on human health such as irritation of the nose, eyes, throat, and respiratory system.[18,20] In addition, identifying H2S in exhaled breath contributes to the detection of symptoms of a variety of oral diseases such as halitosis.[21]

Nanostructured metal oxides with different morphologies have been utilized for a wide range of applications.[22−25] Among oxide materials, both SnO2 and ZnO can be easily grown at low temperatures. These oxide materials are commonly used in several sensors for bio and thermal applications. Because low operational temperatures are expected, SnO2 and ZnO recently attracted widespread attention for use as low-power gas-sensing materials.[26−28] Among metal oxides, heterostructures such as SnO2/CuO, CuO/ZnO, WO3/NiO, ZnO/NiO, ZnO/SnO2, and so on. have been synthesized with different methods as an active material for H2S gas sensing.[6,29−33] ZnO/SnO2 sensors show remarkably higher sensitivity than sensors made only from SnO2 or ZnO under identical experimental conditions.[6] Therefore, this study has chosen to carry out the gas sensing performance of ZnO–SnO2 materials. In the case of ZnO/SnO2, for example, Zhu et al. reported well-designed hierarchical and highly-ordered nanobowl ZnO/SnO2 gas sensors which are sensitive and selective in detecting as low a concentration as 1 ppm H2S gas with long-term stability and repeatability.[34] The hierarchical sensing materials were synthesized via a sequential process combining hard template processing, atomic-layer deposition, and hydrothermal processing.[34] Furthermore, Guo et al. published a study about the hydrothermal synthesis of ZnO/SnO2 for H2S detection.[35] The results revealed that this type of heterostructure has a better H2S gas response and selectivity among some interfering gases, for example, NO, SO2, CO, CH4, and C2H5OH.[35] The most important works concerning gas sensing of ZnO/SnO2 heterostructure are summarized in Table .

Table 1

Comparative Results of ZnO/SnO2 Sensors for Gas Sensing

material	concentration (ppm)	response (Ra/Rg)	T (°C)	target gas	ref.
SnO2/ZnO	0.5	11.5	100	H2S	(6)
ZnO/SnO2	0.5	30	450	H2S	this paper
SnO2 promoted with ZnO	0.5	4.5	350	H2S	(41)
ZnO/SnO2 heterogeneous nanospheres	0.5	3.94	300	H2S	(35)
SnO2 promoted with ZnO	0.5	0.71	350	H2S	(42)
Au-doped ZnO/SnO2 nanofibers	1	73.3	350	H2S	(11)
ZnO/SnO2 heterostructure	1	317	350	H2S	(10)
SnO2 nanobowls branched ZnO NWs	1	6.24	250	H2S	(34)
CuO functionalized SnO2–ZnO core–shell NWs	10	1.69	RT	H2S	(43)
SnO2–ZnO core–shell NWs	25	3.08	400	ethanol	(44)
SnO2/ZnO hierarchical nanostructures	25	3	400	ethanol	(45)
ZnO/SnO2 nanofibers	50	63.3	250	H2S	(46)
SnO2–ZnO core–shell NWs	200	280	400	ethanol	(47)
SnO2 doped ZnO	200	40	450–500	ethanol	(48)

For metal oxides, the optimum surface temperature for proper interaction with airborne species and sensing is above 300 °C.[36] In order to fabricate such devices, it is required that the metal oxides with a designed texture are deposited on dielectric substrates which are chemically stable at high temperatures.[26] Alumina substrates are widely used for depositing polycrystalline functional metal oxides due to their long list of positive features, including high-temperature stabilities and excellent dielectric properties.[26] Different techniques like molecular beam epitaxy, vapor deposition, pulsed laser ablation, sputtering, ultrasonic spray pyrolysis (USP), electrophoretic deposition, and various sol–gel methods have been utilized for depositing ZnO and SnO2 on alumina substrates.[37−39] Among these methods, USP provides some degree of control on the microstructure and morphology of the grown layers with a growth rate as high as 100 nm/min.[40] The main advantage of this method is the ability to grow continuous layers that can perfectly follow the shape of the substrate. Besides all these positive features, USP consumes less energy compared to in-vacuum deposition techniques, which makes it more interesting for gas sensing applications.[40]

To the best of our knowledge, this is the first study to demonstrate ZnO/SnO2 heterostructure grown on alumina substrates using USP deposition for H2S gas sensing purposes. In this design, the sensing layer is integrated with the heater on the rear of the sensor. The material characterization and gas sensing properties of the grown layers with different ratios of zinc chloride and tin (II) chloride dihydrate in the initial precursor are performed and discussed. The sensing layer’s thickness effect on the performance of the sensors is undeniable; however, this research mostly focuses on the heterojunction effect of the sensors. Therefore, the selectivity and gas sensing mechanism of the ZnO/SnO2 has been systematically investigated.

Experimental Section

Alumina plates (3 mm × 3 mm × 0.5 mm) with 99% purity were utilized as substrates. The USP system was utilized to form both a sensing and a heating layer. The details about the USP deposition system have already been presented in our previous work.[26] The substrate temperature of 325 °C was used for all depositions with a precursor spray rate of 4 mL/min. In the first step, the prepared chips were directly placed on the hot stage of the USP system to form microheaters. In the case of the microheater, the precursor solution was prepared by dissolving 0.2 mol/L of SnCl2·2H2O (23742.260, AnalaR NORMAPUR, VWR) in pure ethanol (99.90% v/v). The microheater was formed by USP deposition for 75 min on the rear of the sample followed by 1000 °C furnace annealing for an hour (80% N2 and 20% O2). For deposition of the sensing layer (on the front of the sample), separate precursor solutions which consist of ZnCl2 (29156.231, AnalaR NORMAPUR, VWR) and SnCl2·2H2O with a molarity of 0.2 mol/L were prepared using ethanol as solvent. Then different precursor solutions are prepared by mixing a proportional volume of ZnCl2 and SnCl2·2H2O solutions with the volume ratios as presented in Table . Seven samples (S0 to S6) were prepared, among which S6 contained a pure SnO2 layer and S0 contained the highest investigated amount of ZnCl2. By choosing the appropriate deposition time, the thickness of the sensing layer for all samples was kept the same (130 ± 5 nm). The details of the deposition time for the samples are also presented in Table .

Table 2

Details of USP Depositions of Sensing Layers

	S0	S1	S2	S3	S4	S5	S6
ZnCl2:SnCl2·2H2O precursor solution ratios (V:V)	6:1	5:1	4:2	3:3	2:4	1:5	0:6
deposition time (min)	18.00	13.12	9.00	5.70	5.78	4.80	4.00
measured Zn/Sn ratio from EDX	0.2210	0.1852	0.0998	0.0485	0.0298	0.0258	0

The scanning electron microscopy (SEM) (TESCAN MAIA3) tool was utilized to observe the morphology of the grown layers.

The X-ray diffraction (XRD) data were collected in θ–2θ and grazing incidence modes by using Cu Kα radiation in a nonfocusing geometry using a Philips X’Pert MRD tool. In order to obtain decent beam intensity and high acquisition, an X-ray mirror was applied in primary optics and a parallel-plate collimator was placed prior to the detector in the secondary optics.

To study the elemental composition and chemical state of the ZnO/SnO2 heterostructure materials, X-ray photoelectron spectra (XPS) (ESCALAB 250Xi, Thermo Fisher, USA) analysis was carried out in a spectrometer with a monochromatic Al-Kα X-rays source and spot size of 500 μm. All the binding energies were charge corrected by using the C 1s peak at 284.8 eV of the surface contamination carbon. The fitted peaks of O 1s XPS spectra were deconvoluted using an XPS Peak 4.1 software.

The cross-sectional view of ZnO/SnO2 samples was characterized by transmission electron microscopy (TEM) (Talos F200X, FEI/Thermo Fisher, USA) with an operating voltage of 200 kV. Energy-dispersive X-ray spectroscopy (EDX) was also applied in the TEM tool to determine the element content of ZnO/SnO2 heterostructure samples. Furthermore, nanobeam electron diffraction and selected area electron diffraction (SAD) were also employed to analyze the phase and crystalline quality in the grain.

Before gas sensing measurements, in order to stabilize the structure of the sensors, 600 °C was chosen to anneal the samples in an air atmosphere for an hour. Figure a displays how the prepared gas sensors were mounted on the ceramic support and electrically interconnected to a chip (25 × 35 mm2) with feedthroughs. Diffusion bonding of 75 μm gold wire was used to make electrical contacts within both heater and sensor structures. The sensors with their chip were mounted inside the closed chamber equipped with mass-flow controllers (see Figure b for more details). The cross-sectional SEM micrograph of the SnO2 heater deposited by the USP technique at 325 °C for 75 min is presented in Figure c.

Figure 1

(a) Schematic image of the sample holder, (b) schematic image of the testing set-up, and (c) cross-sectional SEM micrograph of the SnO2 heater deposited by USP on alumina substrates.

Gas sensing measurements were done using a mass flowmeter and apparatus described in our previous work.[49] The sensor temperature and electrical signal were measured by an s-type thermocouple (Pt and 90%Pt10%Rh wires with 25 μm diameter) and Keithley 2410 electrometer, respectively.

The response calculated according towhere Ig and Ia are the sensor currents in target gas and atmospheric air, respectively.

Results and Discussion

Figure displays the SEM cross-section micrographs (130 ± 5 nm thickness) of ZnO/SnO2 heterostructure and the corresponding top view plane. The inset images in Figure at higher magnification show the morphology of the grown crystallites. According to our previous work,[26] ZnO crystallites do not grow continuously on alumina substrates, while SnO2 crystallites perfectly grow on alumina substrates.[26] This behavior also applies to the heterostructure growth of ZnO/SnO2. This means that for the sensors with a lower amount of SnCl2·2H2O in the initial precursor, it is necessary to increase the deposition time in order to keep the thickness the same in all sensors. Previously, the effect of thickness on the gas sensing properties of the sensors has already been investigated in detail;[50] however, in this study, the heterojunction effects on gas sensing properties are targeted. Therefore, by changing the deposition time, the thickness of the all sensors was adjusted to be around 130 nm to obtain decent gas sensing measurements.

Figure 2

The cross-sectional and plan view SEM micrographs of the sensors prepared by USP at 325 °C on alumina substrates: (a) pure SnO2 (S6), (b) S5, (c) S4, (d) S3, (e) S2, (f) S1, and (g) S0.

As observed from the SEM images in Figure a–g, with an increase in the ZnCl2 ratio in the precursor the morphology of the grown layers changes slightly. However, thanks to the benefits of USP deposition, the grown crystallites in each sensor have a uniform distribution all over the substrates. In the case of pure SnO2 (Figure a) the nanocrystallites around 57 nm were found on the surface (line averaging was utilized to determine grain size). This number decreased to 47, 50, and 50 nm for the other three sensors (S5, S4, and S3 respectively) as shown in Figure b–d. The smallest crystallites (around 35 and 39 nm) were found on the surface of S2 and S1, respectively. Then it increased to 49 nm for S0. This result shows that by starting with pure SnO2 in sample S6 and continuously increasing the ZnCl2 ratio, the size of the crystallites decreases (up to sample S1), and then, it increases (for sample S0). These results are in good agreement with previous reports stating that by introducing Zn atoms to the SnO2 initial precursor solution, the size of the grown crystallites is affected.[51]

In this part, the crystal structure of the grown layers was studied by applying XRD. Figure a illustrates the XRD patterns of our samples. The results reveal that all samples have almost similar XRD characteristics. Apart from the substrate peaks (alumina), there are three distinguishable peaks at 26.6, 33.9, and 51.8° corresponding to the sensor layer. These peaks are indexed to the (110), (101), and (211) of the tetragonal rutile SnO2 phase.[51] The absence of diffraction peaks of hexagonal ZnO can be due to the relatively small amount of ZnO grown on alumina or to the crystallites being too small to be detected by XRD. Due to the lattice mismatch of the ZnO and SnO2, it is less likely to dope ZnO crystallites into the SnO2. Therefore, ZnO crystallites might have been grown in isolation all over the surface, forming an external heterostructure with SnO2 (more information will be presented later in the XPS part). In spite of the high proportion of ZnCl2 in the initial USP precursor, the highest intensity for (110) reflection of SnO2 is obtained only for S1. This may also mean that the inclusion of ZnCl2 in the precursor influences the morphology of the film. This may happen because of the slower deposition rate for this sample compared to the other samples. Although the thickness of the layers is the same, deposition time may affect the growth rate of some planes, such as (110). It is important to mention here that there was no detection of any crystalline ternary phases in these layers.

Figure 3

(a) θ–2θ diffractograms from sensors S1 to S6 as well as alumina substrate, and (b) TEM analysis of S1, S3, and S6 samples. The BF, SAD, and EDX images have been illustrated.

Further study was performed by TEM and EDX to analyze the crystallites of the sensing materials in smaller dimensions and to determine their distribution. Figure b shows the bright field (BF), SAD, and EDX images of S1, S3, and S6. In this figure, the SAD image of S1 is only shown because the other samples had a similar feature. The BF images reveal that the material contains mainly crystalline grains but there is a disordered area at the initial stage of the film growth (marked by dashed rings). These phases appear in SAD as concentric rings; meanwhile, there is a hollow ring, which is representative of the disordered region. The rings are identified as planes (110), (101), (111), and (211), which are also consistent with XRD results. There is also a Z-contrast in the BF images where the major distribution of SnO2 appear brightly and lighter elements such as ZnO are dark regions. There are stacking fault defects in the crystalline grains (marked by arrows). The EDX analysis has been focused on detecting Sn, O, and Zn elements in the samples. The presence of Zn is successively decreased from S1 to S3 until Zn vanishes completely in S6. On closer observation of sample S1, the EDX of Zn appears like bright regions containing small spots, but there are also dark regions without Zn in the layer. The Zn amount in sample S4 is obviously smaller and also nonuniform, like for sample S1.

In order to validate the chemical states and surface elemental composition, XPS measurements of S1 and S6 were carried out. As shown in Figure a, the peaks of Zn, Sn, and O, together with C, can be clearly observed in the survey spectrum for S1, whereas peaks of Sn, O, and C are detected for S6, and no other impurities could be observed. The C signal originates mostly from surface hydrocarbons because of the exposure of samples to air or from adventitious hydrocarbons.

Figure 4

(a) XPS survey spectra of S1(pure SnO2) and S6 (ZnO/SnO2). (b) Zn 2p, (c) Sn 3d spectra for S1 and S6. (d) O 1s spectra for S6 and (e) S1.

As shown in Figure b, sample S1 has two strong peaks around 1021.65 and 1044.7 eV in the Zn 2p spectrum. These peaks correspond to Zn 2p 3/2 and Zn 2p 1/2, respectively, indicating that the Zn is in the +2 valence state. In addition, the difference in binding energies between them is 23.05 eV; hence, the substance exists in the form of ZnO.[52] The high-resolution spectrum of Sn 3d for sample S6 shows doublet binding energies at 486.70 and 495.15 eV, corresponding to Sn 3d 3/2 and Sn 3d 5/2, respectively, as shown in Figure c. The energy difference between doublet binding energies is estimated at 8.45 eV, which is in excellent agreement with the reported values.[51,53] Moreover, the Sn peaks of the S1 composite shift 0.35 eV to lower energies, indicating that the interaction and electron transformation may exist between ZnO and SnO2 until equilibrium is reached (the same level of Femi energy).[54,55] The Sn 3d XPS peak positions confirm that the oxidation state of the Sn ions is +4 and the difference in binding energy between the two peaks (6.45 eV) indicates that the substance is SnO2, which is in good agreement with XRD analysis.

Oxygen species play an important role in the gas sensing ability of the sensors by changing the thickness of the depletion layer. Due to the asymmetric shape of the O 1s spectrum in both S1 and S6, it can be deconvoluted into different symmetric peaks to verify the status of oxygen species. For S6 (SnO2), according to Figure d, O 1s can be deconvoluted into nearly two peaks at binding energies of 530.56 and 531.4 eV. The lower binding energy corresponds to the oxygen atoms located in the lattice in the form of Sn–O. The peak at 531.4 eV could be ascribed to the oxygen vacancies in the structure of regular rutile SnO2. In the case of S1 (ZnO/SnO2) O 1s can be deconvoluted into three peaks as shown in Figure e. The lower binging energy at 530.16 eV is attributed to the lattice oxygens in the form of Zn–O and Sn–O. The second binding energy at 531.45 eV is ascribed to the oxygen vacancies in the ZnO/SnO2 structure. The highest binding energy located at 531.99 eV is related to the oxygen atoms chemisorbed at the surface of synthesized materials. This indicates that the ability of chemisorbing of oxygen on the surface of S1 (ZnO/SnO2) is remarkably stronger than single component SnO2. Therefore, the surface oxygen adsorption ability of the sensor greatly contributes to the capability of reacting with target gas molecules and it has the potential to perform well as a gas sensing material.[56]

The response of a gas sensor is remarkably affected by the operating temperature. Hence, the relationship between the gas sensing response and the operating temperature was first investigated in Figure , where all sensors are exposed to 5 ppm of H2S at different operating temperatures ranging from 200 to 600 °C. For sensors S3 to S6, the response is measured to be less than six according to eq at all operating temperatures. For sensors S0, S1, and S2, their gas sensing responses continued to grow as the operating temperature increased from 200 to 450 °C. The maximum amount of response was obtained at 450 °C (113, 172, and 45 for S0, S1, and S2, respectively), indicating that the optimal operating temperature for this designed material combination could be chosen as 450 °C and the most sensitive sensor is S1. For this operating temperature the response and recovery time are 37, and 57 s, respectively. However, all the responses began to decrease as the operating temperature further increased above 450 °C (up to 600 °C). Basically, certain activation energy is required to have a reaction between the target gas and adsorbed oxygen species. At low operating temperatures, gas molecules are not activated enough to overcome the activation energy barrier to react with oxygen species on the surface of ZnO/SnO2 which leads to low response and longer response and recovery time. By increasing the temperature, the conversion of surface-adsorbed oxygen species and higher reaction activity contribute to the higher response. By increasing the temperature above the operating temperature, H2S gas adsorption is too difficult to be adequately compensated for the increased surface reactivity which causes a low utilization rate of the sensing material.[57] In addition, at higher operating temperatures the response and recovery time are shorter, but the response is lower, and because of the higher operating temperature, more energy has to be consumed. In this work, sensor S1 displayed the highest response at 450 °C, which may be closely related to both the establishment of heterostructure and the effective increase in the specific surface area for oxygen adsorption.

Figure 5

H2S gas sensing response of the sensors (S1 to S6) as a function of operating temperature.

Figure a shows the corresponding dynamic gas sensing response curves of the S1 (ZnO/SnO2) sensor at the optimal operating temperature of 450 °C, for various H2S concentrations ranging from 0.5 to 20 ppm. The obtained sensing response values are about 37, 66, 99, 172, 206, and 218 for 0.5, 1, 2, 5, 10, and 20 ppm H2S, respectively (see Figure b). There is a small drop in the signal for 5 ppm and this drop can be attributed to the saturation levels of different materials that we have in the heterostructure (SnO2 and ZnO). This saturation of the sensor starts at 5 ppm as can be observed from Figure b (blue line). This sensor could recover to its initial value when exposed to air again, implying satisfactory stability and reproducibility of the proposed H2S gas sensor. All the response times and recovery times for different concentrations of target gas are presented in Figure b. The repeatability of the fabricated ZnO/SnO2 (S1) sensor was measured to 0.5 ppm H2S gas at 450 °C 10 times as shown at the beginning of the dynamic response in Figure a (response: 36.25 ± 1.25). The measured results indicate that the response has little change. This confirms the good stability of the fabricated ZnO/SnO2 (S1) sensor. In addition, the long-term stability of the sensor over 30 days is presented in Figure S2. To estimate the limit of the detection (LoD) of S1 the calibration curve was prepared for the concentration range of 0.5–20 ppm (Figure d—blue line and Figure c in the logarithmic scale). The data were fitted using the logarithmic fitting (least squares method where the concentration values were logarithm), the fitted curve and its equation are presented in Figure c. Then the fitted line was used to extrapolate its intersection with the doubled noise level (response value of 1), this LoD is lower than 0.3 ppm of H2S.

Figure 6

(a) Dynamic response curve of the S1 sensor for H2S gas with concentrations ranging from 0.5 to 5 ppm at 450 °C operating temperature. (b) Response and recovery time versus concentration, and (c) Prediction for the low detection limit of the sensor. (d) Response of S1 toward 5 ppm of H2S and some other interfering gases with higher concentrations.

The response actions of ZnO/SnO2 of the S1 sensor for six different gases are shown in Figure d. It can be seen that the highest response of the S1 sensor is observed for 5 ppm H2S sensing, which is almost 14 times higher than methanol, ethanol, and some other gases (with higher concentration levels, see the dynamic responses for these gases at Supporting Information, Figure S1). The reason behind the selectivity of ZnO/SnO2 can be attributed to the relatively small molecular size of H2S among most of the gases, which leads it to possess a larger adsorption capacity on the same surface adsorption area.[34] In another work from Fu. et al. demonstrated that ZnO will react with H2S and transfer to ZnS leading to a larger response because the conductivity of ZnS is higher than that of ZnO.[6] The abovementioned results demonstrate that the ZnO/SnO2 sensor S1 can be used for H2S detection. However, humidity has a negative effect on the sensor’s response. In order to have more information, sensor S1 was exposed to 5 ppm of H2S within relative humidity levels of 50, 60, and 80%. As presented in Figure d, the sensor’s responses are 36.3, 17.5, and 8.2 respectively. This means that future works are required to improve the performance of the sensor in humid conditions.

The gas-sensing mechanism of ZnO/SnO2 heterostructures could be described by the surface depletion and grain boundary mechanism. The gas response of the sensor in pure SnO2 can be originated from the homojunction barrier of the intergranular barrier in SnO2.[20] In the air atmosphere, the adsorption of oxygen molecules in the form of O2–, O–, and O2– generates an electron depletion layer at the surface, resulting in a decrease in conductivity.[4] Upon exposure to the target gas, the reaction between H2S molecules and oxygen preadsorbed species on the surface of SnO2 can return electrons to the conduction band.[10] This decreases the width of the depletion layer and increases conductivity. In the case of ZnO/SnO2 heterostructure, the sensitivity is higher than the pure SnO2. This behavior is mainly attributed to the heterojunctions between ZnO and SnO2.[35] When dissimilar semiconducting materials, for example, SnO2 and ZnO, which have different Fermi levels, are brought into contact, the electrons from SnO2 with lower work function (Φ = 4.9 eV) transfer across the interface to ZnO with higher work function (Φ = 5.2 eV) until the Fermi levels are equilibrated.[45] As a result, band bending and charge distribution happen at the interface of the junction. Consequently, the accumulation layer is formed on the ZnO side and a depletion layer is created on the SnO2 side. In order to have any conduction in the sensing material, the charge carriers have to overcome this potential barrier.[45] Upon exposure to the air atmosphere, electrons are extracted from the conduction band of ZnO and SnO2, and consequently, the potential barrier between ZnO and SnO2 increases. When this junction is exposed to the H2S gas, the potential barrier decreases.[34] Herein, this junction plays a critical role in enhancing the gas response of the sensor.

Conclusions

Novel gas sensors made of ZnO/SnO2 materials on alumina substrates with high sensitivity for sensing the H2S gas within an operating temperature range of 200–650 °C have been presented. A heater has been also integrated into the rear of the sensor for raising the operating temperature. The performance of the sensing material was improved by optimizing the solution ratios made of zinc chloride and tin (II) chloride dihydrate. The sensor with 5:1 (ZnCl2: SnCl2·2H2O) ratio in the USP precursor exhibited ∼95 times better response than the pure SnO2 sensor toward 5 ppm H2S at the operating temperature of 450 °C. The enhanced gas-sensing performance was due to the heterostructure formation between ZnO and SnO2. The main advantage of the sensor is its inexpensive one-step USP-based fabrication process on alumina substrate and its relatively high sensitivity in the range of single ppm and sub-ppm concentrations of H2S without any additional functionalization. Other advantages are the integrated heating element on the rear of the sensor made with a similar process (USP) and possible miniaturization of the sensing structure for future improvement of sensor power consumption. Based on our results, the simple ZnO/SnO2 heterostructures made by USP are good candidates for improved selective and sensitive H2S detection.