Selectively enhanced sensing performance for oxidizing gases based on ZnO nanoparticle-loaded electrospun SnO₂ nanotube heterostructures

Abstract

In this work, we present gas sensors based on ZnO nanoparticle-loaded electrospun SnO₂ nanotube (ZnO/SnO₂) n–n heterostructures (HSs) synthesized by electrospinning combined with facile thermal decomposition. The sensing properties of the pristine SnO₂ nanotubes (NTs) and ZnO/SnO₂ HSs were investigated toward the representative oxidizing (NO₂) and reducing (H₂, CO) gases. Results show that the as-prepared ZnO/SnO₂ HSs exhibit selectively enhanced and diminished sensing performances for oxidizing and reducing gases, respectively. These phenomena are closely associated with the modulation of the local depletion layer on the surface of SnO₂ nanoparticles (NPs) caused by charge transfer at the heterojunctions due to work function difference. A modified grain boundary-controlled sensing mechanism is proposed to describe charge transport in sensing layers based on the contact potential barriers between nanoparticles. Our study indicates that the selection of material system and their synergism are keys to the effective design of gas sensors with semiconducting metal oxide HSs.

---

1. Introduction

Metal oxide semiconductor (MOS) resistive gas sensors have been widely used in the detection of toxic, flammable and explosive gases,^1,2 energy efficiency and emission in combustion processes,³ healthcare,⁴ environmental monitoring,⁵ and food processing.^6,7 With the urgent requirements of sustainable development all over the world, gas sensing techniques are now facing challenges towards high sensitivity and selectivity, and short response and recovery times at low gas concentrations. For instance, the U.S. Environmental Protection Agency (EPA) has set a primary standard of 53 parts per billion (ppb) for NO₂,⁸ above which may cause possible health problems especially for those sensitive populations including children, elderly and people with asthma, just naming a few critical challenges. It is therefore of crucial importance to develop high-performance gas sensors to fulfill those future missions.

Currently, various approaches have been used to approach above-mentioned goals with MOS gas sensors. These include doping or surface decorating with noble metal or metal oxide nanoparticles, compositing with other oxides, or plasma exposure etc.^9–16 Among them, the HSs formed by surface coating with NPs, particularly metal oxide NPs with special functionalities is a widely accepted route to improve the sensing performance.

The metal oxide NPs are considered to be electronically sensitive and can modulate the conduction channel through charge transfer at the interface between the host materials and the attached NPs.¹⁷ Due to the modulation, the electric conduction in the sensing layer of a gas sensor becomes more sensitive to the process of gas adsorption and desorption, remarkably influencing the sensing performance. Generally, two approaches have been employed to realize charge transfer between the host materials and the attached NPs. One is the use of p–n heterojunction.^16,18–20 Due to carrier concentration gradient, charge transfer will occur at the interface until the built-in electric field is formed to prevent carrier flow and finally reaches equilibrium in the p–n heterojunction.¹⁸ The other is the n–n or p–p isotype heterojunction, in which charge transfer occurs due to the work function difference between the host materials and the attached NPs.^19,20 From the engineering's point of view, isotype heterojunctions are preferred since constructing the work function difference is relatively easy to be implemented with a variety of materials. In this work, we then aim to n–n HSs configuration and carefully investigate the contribution of heterojunctions to the sensing performance.

One-dimensional (1D) metal oxides (i.e., nanowires, nanobelts, nanorods, NTs and nanofibers, etc.) have been attracting intensive attentions as potential sensing materials due to their high surface-to-volume ratio and efficiencies for carrier transport.^21,22 Particularly, 1D electrospun nanostructures consisting of many nanograins would result in potential barriers between the adjacent NPs, quite different from other types of 1D nanostructures.^23,24 More importantly, the potential barrier can be easily modulated by the process of gas adsorption and desorption, which ensures that the structures exhibit excellent sensing performance.²⁵ It has been demonstrated that the 1D electrospun nanostructures with NPs coatings can effectively improve sensor performance toward some target gases (NH₃, H₂S, H₂ and NO₂) due to the formation of local heterojunctions.^26–30 Up to date, there still have been some critical challenges left to further exploit the potentials for these structures. (1) Are the NPs-coated 1D electrospun nanostructures efficient for the target gases, regardless of their oxidizing or reducing nature? (2) How does the local heterojunction impact the potential barriers between the adjacent nanograins? (3) What is the possible sensing mechanism for the NPs-coated 1D electrospun nanostructures? To address these issues, here we select two common and typical sensing materials, SnO₂ and ZnO as examples. Although the single 1D electrospun nanostructures of those two materials have exhibited high sensing performance, their synergism is expected to further improve the performance and extend the applications of devices at low concentration and complex atmospheres.

In this work, we present the gas sensors based on ZnO nanoparticle-loaded electrospun SnO₂ NT n–n HSs synthesized by the electrospinning and thermal decomposition. The sensing properties of the pristine SnO₂ NTs and ZnO/SnO₂ HSs are investigated toward the representative oxidizing (NO₂) and reducing (H₂, CO) gases. Results show that compared with the pristine SnO₂ NTs, ZnO/SnO₂ HSs exhibit selectively enhanced and diminished sensing performances for oxidizing and reducing gases, respectively. A modified grain boundary-controlled sensing mechanism is proposed to interpret these phenomena based on the modulation of local depletion layers on surface of SnO₂ NPs. Our study indicates that the selection of material system and their synergism are keys to the effective design of gas sensors with semiconducting metal oxide HSs.

2. Experimental

2.1. Materials

Zinc acetate dihydrate (Zn(CH₃COO)₂·2H₂O), and anhydrous ethanol were purchased from Chengdu Kelong Chemical Co. Ltd., China. Polyvinylpyrrolidone (PVP, M_w = 1 300 000 g mol⁻¹) was obtained from Sigma-Aldrich. Tin(II) chloride dihydrate (SnCl₂·2H₂O), and N,N-dimethylformamide (DMF) were supplied by Shanghai Chemical Reagent Company, China. All the chemicals were of analytical reagent grade and used without further purification. Deionized water with a resistivity of 18.25 MΩ cm was obtained using an up-water purification system (Chengdu YouPu Technology Co. Ltd., China) at room temperature.

2.2. Preparation of ZnO/SnO₂ HSs

The SnO₂ NTs were synthesized by electrospinning as reported in our previous work.³¹ In experiments, 1.2 g SnCl₂·2H₂O was dissolved in a mixture of ethanol (5 mL) and of DMF (5 mL) with vigorous stirring at room temperature. Afterwards, 1.2 g PVP was added to the above mixture and stirred overnight to form a uniformly transparent solution. Then, the solution was transferred into a plastic syringe and injected in a rate of 0.25 mL min⁻¹ using a no. 21 stainless steel needle connected to a high voltage DC power supply. Here, a rotating aluminum drum was selected as the collector with a distance of ∼18 cm to the needle tip. When a fixed voltage of 21 kV was applied to the system, the electrospun SnCl₂/PVP nanofibers were produced and deposited on the aluminum drum. Finally, the SnCl₂/PVP nanofibers were peeled off from the collector and calcinated for 3 h at 600 °C in air to remove the polymer and form SnO₂ NTs.

The ZnO/SnO₂ HSs were synthesized by a one-step thermal decomposition according to Zhao's method.³² Briefly, a certain amount of saturated ethanol solution of Zn(CH₃COO)₂·2H₂O were dropped onto the dispersed SnO₂ NTs and dried in air. Then the zinc acetate covered SnO₂ NTs were baked at 450 °C for 20 min in air to obtain the ZnO/SnO₂ HSs.

2.3. Characterization

The morphology of resulting products was investigated using a FEI field emission scanning electron microscopy (FESEM) equipped with an Oxford energy dispersive X-ray spectroscopy (EDS) operated at 15 kV. The detailed microstructure of the ZnO/SnO₂ HSs was studied by a transmission electron microscopy (TEM) and a high-resolution transmission electron microscopy (HRTEM) performed with a FEI TEM under an accelerating voltage of 200 kV. The powder X-ray diffraction (XRD) patterns were recorded on a Bruker X-ray diffractometer with Cu Kα radiation (λ = 0.154 nm) operating at 40 kV in the 2θ range from 20° to 80°.

2.4. Fabrication and measurement of gas sensors

The gas sensors were fabricated by a similar method as reported in our previous work.³³ First, the as-synthesized sensing materials, including SnO₂ NTs and ZnO/SnO₂ HSs, were mixed with deionized water in a weight ratio of 4 : 1 and slightly grinded in an agate mortar to form homogeneous pastes. Then, the pastes were coated on a ceramic substrate of Au interdigitated electrode. After drying at room temperature for 24 h in air, the sensing devices were sintered at 400 °C for 2 h to improve the mechanical strength (Fig. 1 inset).

Fig. 1 Photograph of the intelligent gas sensing analysis system of CGS-4TP. The inset is a top view of Au interdigitated electrode and a sample sensor.

The gas-sensing properties of the sensors were investigated with a CGS-4TP intelligent gas sensing analysis system (Beijing Elite Tech Co. Ltd., China) as shown in Fig. 1. This multifunctional system consists of a heating system, a probe adjustment system, a measurement and data acquisition system, as well as a set of built-in measurement control software. An external temperature control provided by the system can conductively adjust the sensor temperature from 20 °C to about 500 °C with a precision of 1 °C. Two probes were pressed on sensor substrates to export electrical signals. All sensors were pre-heated at different operating temperatures for about 30 min. When the sensor resistances were stable, the target gas with a certain concentration was injected into the test chamber (1.8 L in volume) in 1000 standard cubic centimeters per minute (sccm) by a mass flow controller (MFC). After the sensor resistances reached new constant values, the dry air was injected into the test chamber in 1000 sccm by MFC. The operating temperature, the sensor resistance and response values of these sensors were automatically reported by the analysis system. The response value (R) was defined as R = R_g/R_a or R = R_a/R_g for the oxidizing and inducing gases, respectively, where R_a was the sensor resistance in the ambient air and R_g was the sensor resistance in the detected gas. The response time and recovery time were defined as the time taken for the sensor to reach 90% of its maximum response and decrease to 10% of its minimum response, respectively.

3. Results and discussion

3.1. Morphology and structure

The morphologies of the as-prepared electrospun SnO₂ NTs and ZnO/SnO₂ HSs were observed using FESEM. Fig. 2a shows a hollow tubular structure of the SnO₂ NTs with an outer diameter ranging from ∼470 nm to ∼580 nm. The surface is extremely rough, covered with nanosized SnO₂ grains with an average diameter of ∼90 nm. The thickness of the tube wall is about 130 nm. Fig. 2b shows a low-magnification SEM image of the ZnO/SnO₂ HSs after thermal decomposition process. It can be seen that, the tubular structure is maintained, and the surface becomes much rougher. At higher magnification (Fig. 2c), it can be observed that many tiny ZnO NPs with a mean size of ∼40 nm are loaded on the surface of SnO₂ NTs. The ZnO/SnO₂ HSs comprising of two different components are further confirmed by the EDS mapping analysis (Fig. 2d–g), confirming the existence of Zn (3.37 wt%) element besides Sn (75.45 wt%) and O (21.18 wt%) elements, where the SnO₂ NTs are uniformly covered with ZnO NPs. According to the EDS data, the percentage of ZnO in ZnO/SnO₂ HSs is 4.19 wt%. The tubular structure and rough surface of ZnO/SnO₂ HSs are expected to significantly increase the active sites for gas adsorption, and provide an effective way to improve gas–solid interaction.

Fig. 2 FESEM images of (a) the electrospun SnO₂ NTs and (b and c) ZnO/SnO₂ HSs. (d–g) shows the EDS elemental mapping images of the ZnO/SnO₂ HSs.

To further analyze the detailed microstructure and morphology of ZnO/SnO₂ HSs, TEM and HRTEM observations were carried out. As shown in Fig. 3a and b, the TEM images confirm that the ZnO/SnO₂ HSs present rough tubular structure, and the tube walls are uniformly coated with tiny ZnO NPs, being consistent with the observations in FESEM images. In the HRTEM images of ZnO/SnO₂ HSs (Fig. 3c and d), uniform lattice fringe are observed, and the lattice fringes of d = 0.34 nm and 0.26 nm match well with the crystallographic planes of rutile SnO₂ (110) and hexagonal wurtzite ZnO (002), respectively.

Fig. 3 Bright-field (a and b) TEM and (c and d) HRTEM images of the ZnO/SnO₂ HSs.

In order to analyze the composition and crystalline structure, XRD measurement was then performed. As shown in Fig. 4a, high intensity diffraction peaks are indexed to pure SnO₂ crystals with a rutile structure, well agreeing with the reported values from the Joint Committee on Powder Diffraction Standards card (JCPDS no. 41-1445). No characteristic peaks from other impurities are observed in the XRD pattern, and the sharp diffraction peaks confirm the good crystallinity of SnO₂ NTs. Fig. 4b shows the XRD pattern taken from ZnO/SnO₂ HSs. Compared to the pure SnO₂ NTs, two new but relatively weak diffraction peaks at 2θ = 31.8° and 36.3° are detected, which can be respectively assigned as the (100) and (101) planes of the hexagonal wurtzite ZnO (JCPDS no. 41-1451), further indicating the formation of ZnO NPs coated on the surface of SnO₂ NTs. According to Scherrer formula, the crystallite size is about 51.3 nm, 22.5 nm for SnO₂ and ZnO nanoparticles, respectively. Note that in the XRD patterns, the peak at 2θ = 33.9° indexing to the (101) plane of SnO₂ is rather intense and slightly broaden, the diffraction peak at 2θ = 34.4° assigning to the (002) plane of ZnO is overlapped and cannot be observed. Nevertheless, the (002) plane of ZnO is clearly observed in HRTEM images (Fig. 3c and d).

Fig. 4 XRD patterns of (a) the electrospun SnO₂ NTs and (b) the ZnO/SnO₂ HSs.

3.2. Gas-sensing performance

The operating temperature of metal oxide gas sensors plays a vital role on their sensing performance.^34,35 Generally, by elevating operating temperature, one could achieve an optimal response with improved gas–solid interaction and relatively large surface adsorption. In our work, to determine the optimal operating temperature, the temperature dependent sensing behaviors of the SnO₂ NTs and ZnO/SnO₂ HSs sensors to the target gases NO₂ (10 ppm), H₂ (50 ppm) and CO (50 ppm) are investigated at temperatures ranging from 150 °C to 400 °C. As shown in Fig. 5, the response of all sensors exhibits a trend of “increase-maximum-decay” with the operating temperature, and the largest response is found at 300 °C for all sensors. We can find that, in the initial stage, the gas chemisorption is higher than the gas desorption on the oxide surfaces at low temperature range, thus the adsorption quantity increases with the elevated temperature. As the temperature rises, the gas chemisorption will gradually decrease and the gas desorption will increase, until a balance of these two processes is reached. At this balance point, the adsorption quantity reaches a maximum, or in other words, the sensor exhibits the best response. After this critical point, the exothermic gas chemisorption becomes difficult and the gas desorption dominates, leading to a reduced response.³⁶ Thus 300 °C is selected as the optimal operating temperature and all the following measurements were performed at this temperature.

Fig. 5 Sensor responses of the electrospun SnO₂ NTs and ZnO/SnO₂ HSs at different temperatures (150–400 °C) toward (a) NO₂ (10 ppm), (b) H₂ (50 ppm) and (c) CO (50 ppm).

The dynamic resistance behaviors of the SnO₂ NTs and ZnO/SnO₂ HSs were investigated towards a typical oxidizing gas NO₂. Fig. 6a shows the real-time response–recovery curves of both sensors to NO₂ with the concentration of 0.5–100 ppm. Once the sensors are exposed to a certain concentration of NO₂ in each dynamic cycle, the resistance of both sensors will increase; when the gas supply is stopped, the sensor resistance decreases and finally returns to its baseline, confirming that both sensors are with desired reversibility and repeatability. This is the characteristics of n-type semiconducting oxides to oxidizing gases, and has been well interpreted.³⁷ Note that from the inset of Fig. 6a, the ZnO/SnO₂ HSs exhibit a smaller baseline (∼2.4 kΩ) than that of the SnO₂ NTs (∼4.9 kΩ), being closely related to the modulation of ZnO NPs on the depletion layer of SnO₂ NPs at the local heterojunctions.^9,38 We will discuss this in next section. When focus on a certain dynamic cycle, the ZnO/SnO₂ HSs present a larger resistance change compared with the case of SnO₂ NTs, implying that a larger sensor response is achieved, especially at high gas concentrations. Fig. 6b shows the response of both sensors as a function of gas concentration from 0.5 ppm to 100 ppm. The responses of both sensors increase with gas concentration, roughly obeying linear relationship. Taking the 10 ppm NO₂ as an example, the ZnO/SnO₂ HSs has a response of 65.9, being six times higher than that of the SnO₂ NTs (10.6). Clearly, surface functionalization would assist to improve the response of SnO₂ NTs based gas sensors. Nevertheless, the response and the recovery times of the ZnO/SnO₂ HSs are prolonged (Fig. 6c and d). Based on the above results, we can see that the ZnO/SnO₂ HSs can remarkably enhance the sensor response toward the representative oxidizing gas (NO₂). We then look back to the first issue: is this structure also efficient for the reducing gases?

Fig. 6 Gas sensing performance of the electrospun SnO₂ NTs and ZnO/SnO₂ HSs toward NO₂ in concentration range of 0.5–100 ppm: (a) typical dynamic curves; (b) sensor response as a function of gas concentration; (c) response time; (d) recovery time. Inset of panel (a) shows dynamic curves in concentration range of 0.5–3 ppm. Inset of panel (b) shows sensor responses as a function of gas concentration (0.5–10 ppm).

Accordingly, the sensing performance of the SnO₂ NTs and ZnO/SnO₂ HSs were investigated to the typical reducing gases H₂ and CO in the range of 10–50 ppm. From Fig. 7a and b, it is observed that the sensor resistances decrease when H₂ or CO are injected, and then reach steady values; once the gas supply is stopped, the resistances will increase and finally return to their baselines. Although the dynamic behavior of the sensors to reducing gases is completely different from the case of NO₂, the phenomenon can also be well explained by the sensing mechanism of n-type semiconducting oxides to reducing gases.³⁷ The baseline resistance of the ZnO/SnO₂ HSs is lower than that of SnO₂ NTs, being consistent with the results in the case of NO₂. More interestingly, the ZnO/SnO₂ HSs exhibit much smaller resistance change than that of SnO₂ NTs, being opposite to the case of NO₂. The resistance change of ZnO/SnO₂ HSs could hardly be detected at the concentration of 10 ppm for H₂, even for CO at 25 ppm. The change of sensor resistance is expressed as sensor response and summarized as a function of gas concentration shown in Fig. 7c and d. There exist two features: first, the sensor response roughly increases with the gas concentration; the ZnO/SnO₂ HSs exhibit smaller sensor response compared with the SnO₂ NTs. Even at high concentration of 100 ppm, the responses of ZnO/SnO₂ HSs are only 1.9 and 1.1 for H₂ and CO, respectively, indicating that the ZnO/SnO₂ HSs can greatly decline the sensor response to reducing gases.

Fig. 7 Gas sensing performances of the electrospun SnO₂ NTs and ZnO/SnO₂ HSs toward H₂ and CO in the range of 10–100 ppm: typical dynamic curves toward (a) H₂ and (b) CO; sensor responses as a function of gas concentration toward (c) H₂ and (d) CO.

In order to intuitively compare the sensing performances of the SnO₂ NTs and the ZnO/SnO₂ HSs to oxidizing and reducing gases, the sensor responses of both sensors to CO (50 ppm), H₂ (50 ppm) and NO₂ (5 ppm) are displayed in Fig. 8. For the reducing gases, the ZnO/SnO₂ HSs exhibit declined device response. Particularly, the response of ZnO/SnO₂ HSs (1.44) for H₂ becomes nearly one-fifth of the SnO₂ NTs (6.89). While for the oxidizing gas (NO₂), the ZnO/SnO₂ HSs present remarkably enhanced effect, showing six times higher response (30.84) than that of the SnO₂ NTs (5.09). Thanks to the opposite trend, the ZnO/SnO₂ HSs exhibit a excellent selectivity to NO₂, although the concentration of NO₂ is only one-tenth of the reducing gases. On the basis of this effect, we believe that our devices can be used to selectively detect oxidizing gases under complex atmospheres, in which reducing gases and oxidizing gases coexist, for instance, in chemical industries, laboratories, mines etc.

Fig. 8 Comparison of the sensor responses based on the electrospun SnO₂ NTs and ZnO/SnO₂ HSs toward CO (50 ppm), H₂ (50 ppm) and NO₂ (5 ppm).

3.3. Gas-sensing mechanism

It is well known that the gas-sensing property is heavily dependent on surface area and porosity.³⁹ To examine these parameters, nitrogen adsorption and desorption measurements are performed on the sensor materials (Fig. S1†). The isotherms roughly belong to the type III isotherm in the IUPAC classification, and exhibit type H3 loop, indicating that all samples have similar pore size distribution, i.e., average pore size of ∼2–50 nm in diameter.⁴⁰ The BET surface areas of ZnO/SnO₂ HSs is ∼5.65 m² g⁻¹, only a litter higher than that of the electrospun SnO₂ NTs (∼5.54 m² g⁻¹). Since the difference of the BET surface area of both the samples may be negligible, we reasonably believe that the main factors leading to sensing difference should be other reasons. In addition, after surface loading of ZnO NPs on SnO₂ NTs, the original tubular structure is still maintained and the mean size of ZnO nanoparticles is ∼40 nm, so the morphology of the ZnO/SnO₂ HSs is almost the same as SnO₂ NTs. Accordingly, the true factor resulting in the sensing difference may be the modulation of depletion layer on surface of SnO₂ NPs caused by charge transfer at the local heterojunctions.

Compared with 1D single-crystal nanowires, the tube walls of both SnO₂ NTs and ZnO/SnO₂ HSs are constructed with nanosized SnO₂ grains, resulting in a large amount of homojunctions at the boundaries between the adjacent nanograins. Due to the structure characteristic of tube walls, there would be some interspaces between nanograins in sensing layers.^41–43 As shown in Fig. 9a₁, when the SnO₂ NTs sensor is exposed to the air, the oxygen molecules will diffuse into the tube walls through interspaces and surround the nanograins. Then the oxygen molecules will be chemisorbed on the surface of nanograins and become oxygen species (O₂⁻, O⁻ or O²⁻) by extracting electrons from the conduction band of SnO₂ grains, resulting in an electron depletion layer on the surface of oxides.³⁷ At the same time, the energy band upward bends at the grain boundaries. The generated potential barriers between homojunctions will obstruct the transport of electrons under the applied voltage. Under this circumstance, the sensor will maintain the produced initial resistance, i.e. the baseline in air. If the sensor is exposed to the atmosphere including reducing gases (Fig. 9a₂), the chemisorbed oxygen species will interact with the reducing gases and being removed from the surface of the oxides. The captured electrons will return to the conduction band, and the potential barriers will decrease, resulting in a depressed sensor resistance. When the sensor is exposed to an oxidizing gas (Fig. 9a₃), the oxidizing gas (NO₂) will diffuse into the tube walls through interspaces and be chemisorbed on the surface of the nanograins, further trapping electrons from the conduction band. As a consequence, the potential barrier will increase, resulting in a larger sensor resistance.^25,44,45 In the sensing process, the modulation of potential barriers between the adjacent nanograins is key to determine the change of sensor resistance, i.e. the sensor response. The potential barrier is induced by the electron depletion whose width can be denoted by the Debye length (λ_D). The Debye length of an oxide material can be expressed by the following equation:⁴⁶

where ε is the static dielectric constant, k is the Boltzmann's constant, T is the absolute temperature, q is the electrical charge and n_c is the carrier concentration. For the SnO₂ NTs, the Debye length is calculated to be ∼3.2 nm with ε = 13.5 × 8.85 × 10⁻¹² F m⁻¹, T = 573 K, n_c = 3.6 × 10¹⁸ cm⁻³.⁴⁷ In principle, for the pure SnO₂ NTs, the width of the depletion layer could be expanded or compressed due to the change of electron concentration caused by oxidizing and reducing gases, which eventually determine the potential barriers between the adjacent nanograins.

Fig. 9 Schematically showing the modulation of potential barriers between SnO₂ NPs to reducing and oxidizing gases based on grain boundary-controlled sensing mechanism: (a) the electrospun SnO₂ NTs; (b) the ZnO/SnO₂ HSs. (c) Energy-band diagram of ZnO/SnO₂ heterojunction.

When ZnO NPs are attached on the surface of SnO₂ NTs (Fig. 9b), there exist two physical phenomena. One occurs on the surface of ZnO NPs due to gas adsorption, and the other is at the interface between the ZnO NPs and SnO₂ NPs. When the ZnO/SnO₂ HSs is exposed in air (Fig. 9b₁), oxygen molecules are chemisorbed on the surface of ZnO NPs, then the gas molecules will capture electrons from the conduction band of ZnO NPs, resulting in electron depletion.³⁷ The width of electron depletion layer can also be indicated by the Debye length. For ZnO NPs, ε = 8.75 × 8.85 × 10⁻¹² F m⁻¹,⁴⁸ T = 573 K, the carrier concentration varies in the range of 10¹⁶ to 10¹⁹ cm⁻³,⁴⁹ thus the Debye length changes from ∼22 nm to ∼0.7 nm, smaller than the mean size (40 nm) of ZnO NPs. This indicates that the region of ZnO NPs close to SnO₂ NPs is not affected by the gas chemisorption. On the other hand, since the work function of ZnO NPs (5.2 eV) is larger than that of SnO₂ NPs (4.9 eV), electrons will transfer from ZnO NPs to SnO₂ NPs until a same Fermi level is reached (Fig. 9c). Thus, the width of depletion layer at the local heterojunctions reduces on the side of SnO₂. Particularly, the ZnO NPs on or close to the boundaries between the adjacent SnO₂ NPs can heavily decrease the potential barriers (Fig. 9b₁). The relatively weak potential barriers cause a small sensor baseline for the ZnO/SnO₂ HSs. When the sensor is exposed to the oxidizing gas (Fig. 9b₃), the electron depletion layer of the SnO₂ NPs expands to a width as wide as the pure SnO₂ NTs. The sensor experiences from a less resistive state to a more resistive one, resulting in an increased resistance change, which is responsible for the enhanced sensing performance for oxidizing gases with the ZnO/SnO₂ HSs. On the contrary, when reducing gas is supplied (Fig. 9b₂), the captured electrons are released, and the depletion layer becomes narrow. However, since the depletion layer has already been shortened by the electron transfer due to the n–n heterojunction with different work functions, the change of sensor resistance is decreased, leading to a smaller sensor response.²⁰ This is the reason why the ZnO/SnO₂ HSs exhibit increased and decreased sensor responses respectively for oxidizing and reducing gases.

Interestingly, as previously reported,^20,38 gas sensors developed by SnO₂–ZnO core–shell (C–S) nanofibers or nanowires exhibit opposite sensing performance, showing increased response for reducing gases and decreased response for oxidizing gases. In C–S structures, the model suggests that as the shell thickness increases, the dominant electrical transport pathway is localized from the core (SnO₂) to the shell layer (ZnO). When the shell thickness increases to be comparable to the Debye length of the shell material, the shell layer is completely depleted in air due to the chemisorption of oxygen species, resulting in a relatively high initial resistance. If the reducing gas is supplied, a large amount of electrons will return to the shell layer, leading to a low sensor resistance. The process from a more resistive state to a less one contributes to a large response. While for oxidizing gases, the depleted shell layer can hardly release free electrons to further adsorb gas molecules, only causing a small resistance change. In our case, the ZnO NPs are dispersively attached on the surface of SnO₂ NTs, not forming a continuous conducting shell, and the transport pathway is still confined to the SnO₂ NTs. Duo to electron transfer from ZnO NPs to SnO₂ NTs at the interface, the SnO₂ NTs become electron accumulated, beneficial to adsorb oxidizing gases, leading to an enhanced sensor response. The opposite sensing performances caused by the SnO₂–ZnO C–S structures and the ZnO/SnO₂ HSs demonstrate that the structure of heterojunctions can tremendously impact the sensing properties of sensor devices.

4. Conclusions

In summary, the ZnO nanoparticle-loaded electrospun SnO₂ nanotube n–n HSs were synthesized by electrospinning combined with facile thermal decomposition. The sensing properties of the pristine SnO₂ NTs and ZnO/SnO₂ HSs were investigated toward the representative oxidizing (NO₂) and reducing (H₂, CO) gases. The response of ZnO/SnO₂ HSs is 30.84 for 5 ppm NO₂, six times higher than that of SnO₂ NTs (5.09). While for 50 ppm H₂, the response of ZnO/SnO₂ HSs (1.44) becomes nearly one-fifth of the SnO₂ NTs (6.89). Results indicate that ZnO/SnO₂ HSs exhibit selectively enhanced and diminished sensing performances respectively for oxidizing and reducing gases. These phenomena are closely associated with the modulation of the local depletion layer on surface of SnO₂ NPs caused by charge transfer at the heterojunctions due to work function difference. A modified grain boundary-controlled sensing mechanism is proposed to describe charge transport in sensing layers based on the contact potential barriers between nanoparticles. Our study indicates that the selection of material system and their synergism are keys to the effective design of gas sensors with semiconducting metal oxide HSs.

Acknowledgements

This work was supported by the fund from the Science and Technology on Plasma Physics Laboratory at Research Center of Laser Fusion, CAEP and the Science and Technology Development Fund of CAEP (2014A0302014).

References

1. N. Barsan, D. Koziej and U. Weimar, Sens. Actuators, B, 2007, 121, 18–35.
2. A. Tricoli, M. Righettoni and A. Teleki, Angew. Chem., Int. Ed., 2010, 49, 7632–7659.
3. D. J. Wales, J. Grand, V. P. Ting, R. D. Burke, K. J. Edler, C. R. Bowen, S. Mintova and A. D. Burrows, Chem. Soc. Rev., 2015, 44, 4290–4321.
4. N. Shehada, G. Brönstrup, K. Funka, S. Christiansen, M. Leja and H. Haick, Nano Lett., 2014, 15, 1288–1295.
5. G. F. Fine, L. M. Cavanagh, A. Afonja and R. Binions, Sensors, 2010, 10, 5469–5502.
6. A. Berna, Sensors, 2010, 10, 3882–3910.
7. A. Ponzoni, E. Comini, I. Concina, M. Ferroni, M. Falasconi, E. Gobbi, V. Sberveglieri and G. Sberveglieri, Sensors, 2012, 12, 17023–17045.
8. http://www3.epa.gov/ttn/naaqs/criteria.html.
9. S.-W. Choi, A. Katoch, J.-H. Kim and S. S. Kim, ACS Appl. Mater. Interfaces, 2014, 6, 17723–17729.
10. S.-W. Choi and S. S. Kim, Sens. Actuators, B, 2012, 168, 8–13.
11. P. Rai, Y.-S. Kim, H.-M. Song, M.-K. Song and Y.-T. Yu, Sens. Actuators, B, 2012, 165, 133–142.
12. X.-j. Wang, W. Wang and Y.-L. Liu, Sens. Actuators, B, 2012, 168, 39–45.
13. D. J. Yang, I. Kamienchick, D. Y. Youn, A. Rothschild and I. D. Kim, Adv. Funct. Mater., 2010, 20, 4258–4264.
14. H. W. Kim, S.-W. Choi, A. Katoch and S. S. Kim, Sens. Actuators, B, 2013, 177, 654–658.
15. G. Lu, L. E. Ocola and J. Chen, Adv. Mater., 2009, 21, 2487–2491.
16. M. Mashock, K. Yu, S. Cui, S. Mao, G. Lu and J. Chen, ACS Appl. Mater. Interfaces, 2012, 4, 4192–4199.
17. D. R. Miller, S. A. Akbar and P. A. Morris, Sens. Actuators, B, 2014, 204, 250–272.
18. D. Neamen, Semiconductor physics and devices, McGraw-Hill, Inc., 2002.
19. S. Sen, P. Kanitkar, A. Sharma, K. Muthe, A. Rath, S. Deshpande, M. Kaur, R. Aiyer, S. Gupta and J. Yakhmi, Sens. Actuators, B, 2010, 147, 453–460.
20. S.-W. Choi, A. Katoch, G.-J. Sun, J.-H. Kim, S.-H. Kim and S. S. Kim, ACS Appl. Mater. Interfaces, 2014, 6, 8281–8287.
21. J. Huang and Q. Wan, Sensors, 2009, 9, 9903–9924.
22. M. J. Spencer, Prog. Mater. Sci., 2012, 57, 437–486.
23. R. Ab Kadir, Z. Li, A. Z. Sadek, R. Abdul Rani, A. S. Zoolfakar, M. R. Field, J. Z. Ou, A. F. Chrimes and K. Kalantar-zadeh, J. Phys. Chem. C, 2014, 118, 3129–3139.
24. S. Bai, S. Chen, Y. Zhao, T. Guo, R. Luo, D. Li and A. Chen, J. Mater. Chem. A, 2014, 2, 16697–16706.
25. A. Katoch, S.-W. Choi and S. S. Kim, Nanotechnology, 2014, 25, 455504.
26. Q. Qi, P.-P. Wang, J. Zhao, L.-L. Feng, L.-J. Zhou, R.-F. Xuan, Y.-P. Liu and G.-D. Li, Sens. Actuators, B, 2014, 194, 440–446.
27. S.-W. Choi, J. Zhang, K. Akash and S. S. Kim, Sens. Actuators, B, 2012, 169, 54–60.
28. Z. Wang, Z. Li, T. Jiang, X. Xu and C. Wang, ACS Appl. Mater. Interfaces, 2013, 5, 2013–2021.
29. L. Xu, R. Zheng, S. Liu, J. Song, J. Chen, B. Dong and H. Song, Inorg. Chem., 2012, 51, 7733–7740.
30. B.-H. Jang, O. Landau, S.-J. Choi, J. Shin, A. Rothschild and I.-D. Kim, Sens. Actuators, B, 2013, 188, 156–168.
31. Y. Huang, Y.-E. Miao, L. Zhang, W. W. Tjiu, J. Pan and T. Liu, Nanoscale, 2014, 6, 10673–10679.
32. X. Zhao, P. Wang, Y. Gao, X. Xu, Z. Yan and N. Ren, Mater. Lett., 2014, 132, 409–412.
33. K. Diao, M. Zhou, J. Zhang, Y. Tang, S. Wang and X. Cui, Sens. Actuators, B, 2015, 219, 30–37.
34. L. Wang, Z. Lou, J. Deng, R. Zhang and T. Zhang, ACS Appl. Mater. Interfaces, 2015, 7, 13098–13104.
35. T. Kida, S. Fujiyama, K. Suematsu, M. Yuasa and K. Shimanoe, J. Phys. Chem. C, 2013, 117, 17574–17582.
36. Y.-F. Sun, S.-B. Liu, F.-L. Meng, J.-Y. Liu, Z. Jin, L.-T. Kong and J.-H. Liu, Sensors, 2012, 12, 2610–2631.
37. C. Wang, L. Yin, L. Zhang, D. Xiang and R. Gao, Sensors, 2010, 10, 2088–2106.
38. A. Katoch, S.-W. Choi, G.-J. Sun and S. S. Kim, J. Mater. Chem. A, 2013, 1, 13588–13596.
39. K. Sahner and H. Tuller, J. Electroceram., 2010, 24, 177–199.
40. S. J. Gregg and K. S. W. Sing, Adsorption, Surface Area, and Porosity, Academic Press, London, 1991.
41. S.-W. Choi, J. Y. Park and S. S. Kim, Chem.–Eng. J., 2011, 172, 550–556.
42. S.-W. Choi, J. Y. Park and S. S. Kim, J. Mater. Res., 2011, 26, 1662–1665.
43. C. Jin, S. Park, H. Kim and C. Lee, Sens. Actuators, B, 2012, 161, 223–228.
44. C. Xu, J. Tamaki, N. Miura and N. Yamazoe, Sens. Actuators, B, 1991, 3, 147–155.
45. N. Barsan and U. Weimar, J. Electroceram., 2001, 7, 143–167.
46. S. M. Sze and K. K. Ng, Physics of semiconductor devices, John Wiley & Sons, 2006.
47. H. Ogawa, M. Nishikawa and A. Abe, J. Appl. Phys., 1982, 53, 4448–4455.
48. D. C. Look, J. W. Hemsky and J. Sizelove, Phys. Rev. Lett., 1999, 82, 2552.
49. Ü. Özgür, Y. I. Alivov, C. Liu, A. Teke, M. Reshchikov, S. Doğan, V. Avrutin, S.-J. Cho and H. Morkoc, J. Appl. Phys., 2005, 98, 041301.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra03061k

‡ K. D. and Y. H. contribute equally to this work.

---