Introduction

Semiconductor metal oxides (SMOs) have attracted tremendous attention in chemical sensors due to the distinctive resistivity changing features in a certain ambient [1, 2]. As an n-type semiconductor, ZnO has been extensively used as a gas-sensing material due to its wide band gap (3.37 eV) and high exciton binding energy (60 meV) at room temperature [3–9]. Nevertheless, some inherent drawbacks including high operating temperature, slow response/recovery time, and low sensitivity may hinder the further development of ZnO-based gas sensors. So, great efforts have been made in the last decades to overcome the above-mentioned limitations using different methods, including noble metal doping, structure optimization, and heterostructure fabricating [10–12]. However, noble metal doping and heterostructure fabricating methods exist the problems of high cost as well as complex fabricating process. It seems that the morphologically controllable synthesis of ZnO nanomaterials is a feasible way for the improvement of the gas-sensing properties [11], because the gas-sensing is a surface-controlled process and influenced by the specific surface area.

Recently, hierarchical nanostructures including urchin-like, flower-like, and forest-like that derive from low-dimensional nanostructures (e.g., nanowires, nanorods, nanobelts, nanotubes, and nanofibers), have aroused a great deal of attention for the improvement of potential gas-sensing performance [12–20]. For example, Yao et al. [18] obtained spherical ZnO nanobelt-flowers via pyrolysis process, the highly exposed surfaces of nanobelt branches provide high response of 6.94–44.1 ppm benzene gas. Hieu et al. [19] prepared urchin-like ZnO microspheres by sputtering Zn onto the template of a polystyrene sphere, and it shows a sensing of tens of ppb levels of NO concentrations with good response and gas selectivity. Yu et al. [20] synthesized micro-lotus ZnO nanosheets using a facile hydrothermal method, and the hierarchical structure presents excellent response towards ethanol and acetone. Among these hierarchical nanostructures, brush-like nanostructures assemblies of nanowires/rods are very attractive for fabricating efficient gas sensor, because the well-aligned morphology can augment the carrier transport and can be easily affected by the gas molecules [21–23]. Although considerable efforts have been devoted to improving gas-sensing performances using hierarchical nanostructures, there is seldom research work on synthesis brush-like ZnO homostructures for enhanced gas-sensing applications.

In this work, brush-like ZnO hierarchical nanostructures have been synthesized combining electrospinning and hydrothermal method. The brush-like ZnO nanowires and ZnO nanofibers gas sensors have been fabricated, and the gas-sensing properties were investigated for toluene and CO under the optimum working temperature. The measured results show that brush-like ZnO nanowires sensor exhibits enhanced gas-sensing performances comparing with ZnO nanofibers, which suggests that the brush-like ZnO nanowires are very promising materials for the application of gas sensors.

Experimental

Synthesis

All the reagents were of analytical grade and used without further purification. In a typical procedure, 0.50 g of zinc acetate dehydrate (Zn(AC)2·2H2O) was dissolved into 2 mL N,N-dimethylformamide (DMF) to form transparent solution under vigorous stirring. Then, 0.25 g poly (vinyl pyrrolidone) (PVP, Mw = 1,300,000) was added into the above solution. The mixture was aged at room temperature for 4 h to obtain a homogeneous viscous sol solution as the precursor for electrospinning. The precursor solution was transferred into a 1-mL syringe with a hypodermic needle (inner diameter = 0.51 mm, 21G) in a home-made electrospinning equipment, and electrospun by applying 8 kV at an electrode distance of 12 cm at room temperature with a relative humidity (RH) of 37 %. The composite nanofibers were collected on an aluminum frame and peeled off from the frame. Finally, the crystal ZnO nanofibers were obtained after calcined at 600 °C for 2 h. In order to get brush-like ZnO nanowires, the as-prepared ZnO nanofibers were grown as our previous work [24]. Figure 1 illustrates the procedures for fabricating brush-like ZnO nanowires.

Fig. 1
figure 1

Processing steps for the fabrication of brush-like ZnO nanowires combining electrospinning and hydrothermal method

Full size image

Characterization

The thermal behaviors of the samples were characterized using a thermal gravimetric analyzer (TGA, Melttler Toledo 825). Fourier transform infrared spectroscopy (FTIR, Bruker TENSOR 27) was introduced to characterize the samples using KBr pellets. The crystallographic information of the samples was investigated by X-ray powder diffraction (XRD, DRIGC-Y 2000A) using Cu Kα1 radiation (λ = 1.5406 Å) as the radiation source. Morphology and structural investigations were observed using field emission scanning electron microscopy (FE-SEM, JEM-7100F) and transmission electron microscope (TEM, JEM-2100F). The specific surface area was measured using a surface area analyzer (Tristar 3020).

Fabrication and measurement of gas sensors

In a typical produce, the as-prepared ZnO nanostructures were mixed with moderate terpineol and ethyl cellulose with a weight ratio of 2:8:1 to form a homogeneous paste, which was coated onto a ceramic tube with a pair of Au electrodes and Pt wires for electrical contacts (Fig. 2a). Then, the elements were dried in air and annealed at 600 °C for 2 h. Finally, a Ni–Cr heating wire was inserted in the tube for controlling the working temperature by tuning the voltage. The alumina tube was welded onto a pedestal with six probes to give the final sensor unit, as shown in Fig. 2b. The gas-sensing properties of the sensors were measured by CGS-1TP intelligent analysis system (Elite, Beijing, China) using a static test system. The target vapor was injected into the test chamber (about 1 L in volume) by a syringe through a rubber plug. After fully mixed with air (RH was about 25 %), the sensor was put into the test chamber manually. When the response reached a constant value, the sensor was taken out of the chamber to recover in air. The sensor sensitivity (response) was defined as R a/R g, where R a and R g were the resistance of the sensor upon exposure to air and to the target gas. The time taken by the sensor to reach 90 % of the total resistance change was defined as the response and recovery time in the case of gas adsorption and desorption, respectively.

Fig. 2
figure 2

a Schematic of ZnO sensing element, b photograph of the integrated sensor coated with the ZnO sensing materials

Full size image

Results and discussion

Morphology and structure analysis

To obtain the ZnO nanofibers, Zn(Ac)2/PVP composite nanofibers were calcined using muffle furnace. The thermal behaviors experiments were carried out on pure PVP nanofibers and Zn(Ac)2/PVP composite nanofibers. Figure 3 shows the thermal events combustion process for Zn(Ac)2/PVP composite nanofibers. The first process could be attributed to the loss of residual solvent and water remaining in the composite nanofibers from 100 to 270 °C [25, 26]. The second mass loss was ascribed to the complete degradation of PVP and the decomposition of acetates between 270 and 500 °C. Over 500 °C, the weights of the samples remain constant, which proves that all the volatiles and organic components are removed completely. The weight loss process of pure PVP nanofibers also contains three steps, which begins to occur at about 350 °C and completes at about 600 °C. Comparing with PVP nanofibers, the Zn(Ac)2/PVP composite nanofibers has lower thermal stability. It can be attributed to the interaction between the PVP and ZnO materials. Based on the TGA curves, the temperature of 600 °C was selected as the calcination temperature in the following procedure.

Fig. 3
figure 3

TGA curves of a PVP nanofibers and b Zn(Ac)2/PVP composite nanofibers

Full size image

Figure 4 depicts the FTIR spectra of PVP nanofibers, Zn(Ac)2/PVP composite nanofibers, ZnO nanofibers, and brush-like ZnO nanowires. It can be clearly seen that the FTIR spectra of PVP nanofibers and Zn(Ac)2/PVP composite nanofibers possess strong peaks at around 1285, 1438, and 1661 cm−1 which are assigned to the C–C, C–H, and C=O stretching vibration of PVP [27]. The broad peak observed at 3280–3500 cm−1 comes from the water absorption [25]. However, for the spectra of ZnO nanofibers and brush-like ZnO nanowires, it is found that the characteristic peaks of PVP have disappeared because of the removal of the organic molecules. Simultaneously, a new peak is observed at around 475 cm−1 in the spectra of ZnO nanofibers and brush-like ZnO nanowires, which can be attributed to the vibration of Zn–O bonds in ZnO crystals [13, 25].

Fig. 4
figure 4

FTIR spectra of various samples: a PVP nanofibers, b Zn(Ac)2/PVP composite nanofibers, c ZnO nanofibers, and d brush-like ZnO nanowires

Full size image

Figure 5 shows the XRD diffraction patterns of ZnO nanofibers and brush-like ZnO nanowires. Both of the diffraction peaks are corresponding to hexagonal wurtzite structure of ZnO, which can be indexed with standard JCPDS data card no. 36-1451. There is no remarkable shift in diffraction peak, which indicates that no intermediate products are produced during the reaction. From the curve b, we can clearly find that the (002) peak is much stronger and sharper for brush-like ZnO nanowires, which indicates the high-quality crystallinity and c axis orientation.

Fig. 5
figure 5

XRD patterns of the obtained nanostructures: a ZnO nanofibers, b brush-like ZnO nanowires

Full size image

Figure 6a reveals the SEM images of the ZnO nanofibers after calcination, which displays a randomly distributed and interweaved to form a network. The amplified inset image shows that the diameter is about 160 nm for the single nanofiber, and the BET surface area of the ZnO nanofibers can reach to 2.892 m2g−1. At the same time, the detail morphology characterization of ZnO nanofibers is performed using TEM. Figure 6b shows the low-resolution TEM image of ZnO nanofibers, it is found that the nanofibers were composed of plenty of nanoparticles with sizes around 25–35 nm. Figure 6c illustrates the typical high-resolution TEM (HRTEM) image, and the measured lattice spacing values of 0.26 nm is corresponding to the ZnO (0001) plane [10, 28]. Figure 6d exhibits the selected area electron diffraction (SAED) pattern, which can be attributable to the fact that the ZnO nanofibers are polycrystalline.

Fig. 6
figure 6

The ZnO nanofibers, a SEM image, b TEM image, c HRTEM image, and d corresponding SAED pattern

Full size image

The brush-like ZnO nanostructures were obtained by hydrothermal growth, which is built up from compactly aggregated, irregular-shaped ZnO nanowires (Fig. 7a). From the high-magnification SEM image (inset), we can clearly observe that the nanowires attach vertically to the sidewall of nanofibers with a length of 3 μm. Figure 7b presents the brush-like ZnO nanowires with a low magnification, and ZnO nanowires are uniformly distributed on the surface of nanofibers. Figure 7c exhibits the typical HRTEM image of ZnO nanowires. The lattice fringes between two adjacent planes are about 0.52 nm which is equal to the lattice constant of the ZnO. The measured results demonstrate that the ZnO nanowires grow along the [0001] direction [9, 29], and the corresponding SAED pattern proves that the ZnO nanowires are single crystals (Fig. 7d). Moreover, the measured BET surface area of the brush-like ZnO nanowires can reach to 5.982 m2g−1, which is twice as big as ZnO nanofibers.

Fig. 7
figure 7

The brush-like ZnO nanowires, a SEM image, b TEM image, c HRTEM image, and d corresponding SAED pattern

Full size image

Sensing performance and mechanism

Due to its unique hierarchical structures, brush-like ZnO nanowires can provide an advantage for gas-sensing applications. Therefore, we fabricated chemical gas sensors based on ZnO nanofibers and brush-like ZnO nanowires, and then compared their gas-sensing performance for toluene and CO detection.

In order to optimize the working temperature, gas-sensing experiments were conducted on the as-prepared gas sensors. Figure 8a shows the response of ZnO nanofibers and brush-like ZnO nanowires sensors to 100 ppm toluene under the working temperature between 120 and 280 °C. The responses of both ZnO nanofibers and brush-like ZnO nanowires gas sensors all increase slowly and reach their maximum at 240 °C, and then decreases with increasing the temperature further. The decrease of response can be attributed to that some adsorbed test gas molecules easily escape from the surface of ZnO before their reaction since the working temperature is too high [11, 30]. At the optimum temperature of 240 °C, brush-like ZnO nanowires exhibit the highest response of about 12.7 to toluene, which is 3.3 times larger than that of ZnO nanofibers. Moreover, the similar tendency can be also observed for the CO under different working temperatures as Fig. 8b. Under the optimum working temperature (220 °C), the response of brush-like ZnO nanowires to CO can reach to 5.9, which is about 1.25-folds higher than that of ZnO nanofibers.

Fig. 8
figure 8

Relationship between working temperature and response of the as-prepared sensors under 100 ppm, a toluene, b CO

Full size image

Figure 9 shows the sensing transient of as-fabricated sensors to toluene and CO with different concentrations under their optimum temperatures. From Fig. 9a, b, we can clearly observe that both ZnO nanofibers and brush-like ZnO nanowires gas sensors show a clear and fast response increases with increasing the concentration of toluene and CO. After many cycles between exposure to detection gas and fresh air, the resistances of the sensors can recover to their initial values, which suggest that all the fabricated sensors have good reversibility and distinguished reproducibility. In addition, the brush-like ZnO nanowires sensor exhibits a higher response under the same gas concentration compared with ZnO nanofibers sensor, which is about twofold to sixfold enhancement. However, there is no remarkable enhancement for CO, when the gas concentrations are under 100 ppm. At the same time, in order to confirm the relationships between response and gas concentrations, the sensitivities of fabricated sensors are plotted as a function of the gas concentration as Fig. 9c, d. The measured results exhibit that the fabricated gas sensors tend to reach saturation states at high gas concentration for toluene and CO. Moreover, the brush-like ZnO nanowires sensor displays improvement response for toluene, and the detection limit for brush-like ZnO nanowires can down to 1 ppm for toluene and 5 ppm for CO with the responses of 3.8 and 3.2, respectively.

Fig. 9
figure 9

Dynamic sensing transients of the ZnO nanofibers and brush-like ZnO nanowires sensors; a toluene, b CO; The dependence of the sensors response, c toluene, d CO

Full size image

Response and recovery characteristics are the main parameters in design of gas sensors for desired applications. Figure 10 shows the magnified response/recovery plots for toluene and CO with the concentration of 100 ppm. The measured response/recovery times of brush-like ZnO nanowires sensor are just 9/4 s and 6/2 s for toluene and CO, respectively, which are much shorter than that of ZnO nanofibers sensor (11/5 s for toluene and 32/12 s for CO). From the measured results, we can find that the brush-like ZnO nanowires sensor has faster response/recovery speed comparing with ZnO nanowires sensor.

Fig. 10
figure 10

Response and recovery behavior of sensors to100 ppm. a toluene, b CO

Full size image

For a gas sensor, gas selectivity is another important parameter in practical applications. Therefore, the responses of brush-like ZnO nanowires sensor to 100 ppm methanol, carbon monoxide, acetone, ammonia, methane, and hydrogen are also investigated under the optimized working temperature of 240 °C (Fig. 11). It is obvious that the brush-like ZnO nanowires sensor exhibits higher response to toluene than to other tested gases, and there are no apparent response to methane and hydrogen. Nevertheless, further investigation may still be required to improve the selectivity of the material via surface modification and bulk doping with heteroatom for methanol and carbon monoxide. At the same time, to compare the sensing performances of the fabricated brush-like ZnO nanowires gas sensor, we summarized the previous reported ZnO-based toluene gas sensor in Table 1 [31–36]. From the measured results, we can find that the fabricated brush-like ZnO nanowires gas sensor has higher sensitivity, faster response/recovery time, and lower working temperature.

Fig. 11
figure 11

Response of brush-like ZnO nanowires sensor to different kinds of gases for 100 ppm at 240 °C

Full size image
Table 1 Gas-sensing properties comparison of toluene gas sensors based on various structures of ZnO in the literatures and in this study
Full size table

The excellent sensing capabilities observed in the brush-like ZnO nanowires sensor are likely to be associated to the hierarchical structures, which exhibits unique loose structures and high surface area [37]. The brush-like ZnO nanowires possess much higher specific surface area than that of the ZnO nanofibers. Compared with ZnO nanofibers, when exposed to the target gas, the hierarchical brush-like structures with a number of branch structures can increase the surface area to absorb more target gas and provide more active sites for the interaction with the gas as in Fig. 12a, and the adsorbed oxygen ions will interact with the target gas to produce CO2 and H2O (Fig. 12b). At the same time, larger amount of electrons will release into the conduction band of ZnO, leading to the lower resistance and higher response of the sensor. Furthermore, the gas diffusion speed and length are significant factors that affect the response and recovery time of ZnO gas sensors [30]. ZnO nanowires are generally assembled in highly periodic, thus there is no impediment in gas diffusion towards the entire sensing surface, resulting in fast response and recovery time. As a consequence, the gas-sensing performance of brush-like ZnO nanowires is greatly improved.

Fig. 12
figure 12

a Schematic illustrations of hierarchical structures sensing mechanism, b schematic diagram of catalytic reactions

Full size image

Conclusions

In summary, the brush-like ZnO nanowires was synthesized by two-step method combining electrospinning and hydrothermal. The unique brush-like ZnO hierarchical structures are composed of the ZnO nanowires branches which are uniformly deposited on the surface of ZnO nanofibers. In order to investigate the gas-sensing characteristics, two types of sensors were prepared for the detection of toluene and CO. The experiments results show that the response of brush-like ZnO nanowires sensor can reach to 12.7 and 5.9 for toluene and CO under the optimum temperature, which is much higher than that of ZnO nanofibers sensor. Moreover, brush-like ZnO nanowires sensor also exhibits fast response/recovery times for toluene (9/4 s) and CO (6/2 s), which are much shorter than that of ZnO nanofibers sensor (11/5 s for toluene and 32/12 s for CO). These measured results suggest that brush-like ZnO nanowires can be a promising material structure for the application high-performance gas sensors.