Graphene-Modified ZnO Nanostructures for Low-Temperature NO2 Sensing.

This paper develops a novel ultrasonic spray-assisted solvothermal (USS) method to synthesize wrapped ZnO/reduced graphene oxide (rGO) nanocomposites with a Schottky junction for gas-sensing applications. The as-obtained ZnO/rGO-x samples with different graphene oxide (GO) contents (x = 0-1.5 wt %) are characterized by various techniques, and their gas-sensing properties for NO2 and other VOC gases are also evaluated. The results show that the USS-derived ZnO/rGO samples exhibit high NO2-sensing property at low operating temperatures (e.g., 70-130 °C) because of their high specific surface area and porous structures when compared with the ZnO/rGO sample obtained by the traditional precipitation method. The content of rGO shows an obvious effect on their NO2-sensing properties, and the ZnO/rGO-0.5 sample has a high response of 62 operating at 130 °C, three times that of pure ZnO. The detection limit of the ZnO/rGO-0.5 sensor to NO2 is as low as 10 ppb under the present test condition. In addition, the ZnO/rGO-0.5 sensor shows a highly selective response to NO2 gas when compared with organic vapors and other inflammable or toxic gases. The theoretical and experimental analyses indicate that the enhancement in NO2-sensing performance of the ZnO/rGO sensor is attributed to the formation of wrapped ZnO/rGO Schottky junctions.

Introduction

With the expansion of industrial production and family automobiles, the issue of environmental pollution has begun to attract increasing attention because of the big threat to the human health and ecological environment.[1−3] Among the pollution sources, especially gaseous pollutants such as NO2, CO, SO2, H2S, and volatile organic compound (VOC) gases have become a chief culprit.[4,5] Nitric oxides (NO and NO2) generated from vehicle exhausts and chemical plants are the main source of acid rain, and the acidification caused by acid rain seriously damages crops and the ecological environment.[6,7] NO2 is also one of the major inducements of hazes which cause a large number of respiratory diseases and other bad impacts on the human health.[8] In order to effectively monitor and control the discharge of NO2 gases and other harmful gases, gas-sensing materials and devices have attracted much attention worldwide.[9,10] For this ends, semiconductor materials are the focus in the field of gas sensors.[11−13] For detecting NO2, a number of gas sensors based on semiconductors have been developed (see Table S1, Supporting Information).[14,15] Because of the chemical stability and low cost, NO2 sensors based on metal oxides, such as ZnO,[16−18] SnO2,[19−21] WO3,[22−24] Fe2O3,[25,26] and NiO,[27,28] have widely been reported. Because of the high operating temperatures (higher than 200–600 °C), these sensors are limited in the practical applications.

Recent research shows that Schottky junction-based gas sensors enable ultrasensitive and rapid response to gas detection.[29,30] The basis for the Schottky junction to enhance gas-sensing performance is to change the height of the Schottky barrier. The change of the barrier causes the resistance change of the gas sensitive material to finally change the acquired electrical signal.[29] Recently, reduced graphene oxide (rGO) materials with large specific surface areas and excellent electrical conductivity have been widely used in gas sensors,[31,32] and the rGO-related materials possess rich surface functional groups that can effectively improve the gas-sensing performance and reduce the operating temperature.[33,34] The work functions of zero-bandgap rGO and n-type semiconductor ZnO are about 4.75 and 4.1 eV, respectively, with a work function difference of 0.65 eV which is helpful to form a suitable Schottky junction.[35−37] Song et al.[38] reported an rGO/ZnO composite as the photoanodes for solar cells, and its photo-to-current conversion efficiency was 1.3 times higher than that of the pure ZnO because of the existence of the Schottky junction. The ZnO/rGO composite with a Schottky junction can be a promising ultrasensitive gas-sensing material.[39]

The ultrasonic spray is the use of electronic high-frequency oscillation generated cavitation to break up liquid precursors to produce natural and elegant water mist.[40−44] The ultrasonic spray method has been widely used in spray and coating processes for the synthesis of various functional nanocomposites. Chai et al.[40] used the ultrasonic spray method to deposit methylamine iodide on a PbI2 substrate and transformed PbI2 into a more uniform and thicker perovskite film for solar cells. Almuntaser et al.[41] used the ultrasonic spray method to synthesize inverted OSC with an active layer infiltrated in the ZnO nanorod array. Ji et al.[44] prepared highly efficient all-inorganic quantum-dot light-emitting diodes by the ultrasonic atomization process using NiO and ZnO as active materials. These reports show that the ultrasonic spray is an effective technology to synthesize nanoparticles and thin films with high homogeneity and stability.[42] Their sizes can be conveniently controlled by changing the concentration of their precursor solutions and atomizing parameters.[43]

In our work, we have developed various methods to synthesize nanomaterials for gas-sensing applications.[11−13,20,31,45−47] Inspired by the above research, we herein develop an ultrasonic spray-assisted solvothermal (USS) method to synthesize wrapped structured ZnO/rGO nanocomposites with Schottky junctions for gas-sensing applications (Figure ). The as-obtained ZnO/rGO nanocomposites synthesized by the USS method exhibit high-performance NO2-sensing property at relatively low operating temperatures (e.g., 70–130 °C). When compared with the ZnO/rGO sample obtained by the traditional precipitation method, the USS-derived ZnO/rGO sample shows a higher specific surface area and porous structure, which provides the theoretical basis to achieve a good gas-sensing property. This paper has systematically characterized the as-obtained ZnO/rGO composites using various techniques and their gas-sensing behaviors have been well evaluated. The related mechanisms for the synthesis and gas-sensing properties of ZnO/rGO nanocomposites have also been discussed.

Figure 1

Schematic description of the process for the synthesis of ZnO/rGO nanocomposites as active materials for gas-sensing applications.

Results and Discussion

Formation of ZnO/rGO Nanocomposites

There are three possible chemical processes involved in the formation of ZnO/rGO nanocomposites. The first reaction is the precipitation of Zn2+ ions, that is, Zn2+ + 2OH– → Zn(OH)2; then the Zn(OH)2 species transform to ZnO nanocrystals via a nucleation and growth process under the solvothermal condition, that is, Zn(OH)2 → ZnO.[48] Meanwhile, the alkaline environment with pH > 7 facilitates the reduction of graphene oxide (GO) to partly form rGO, that is, GO → rGO.[49] In the first step, Zn2+ ions are adsorbed on GO nanosheets, forming GO/Zn(OH)2 species in an alkaline solution during ultrasonic spray and stirring. In the second step, GO species are in situ reduced to rGO, and the ZnO/rGO nanocomposites are formed during the solvothermal treatment at 180 °C. The cavitation forces generated by the ultrasonic spray process during the addition of the Zn2+/GO suspension ensure that the GO completely and uniformly disperses and coated with Zn(OH)2 particles, providing regular growth conditions during the hydrothermal process. Then, the in situ growth of ZnO nanocrystals on rGO nanosheets make the ZnO particles tightly attach to the rGO surface, and the strong interaction between ZnO nanocrystals and rGO nanosheets is favorable for efficient electron transport at their interface. To investigate the effect of rGO on the NO2-sensing performance, we synthesized a series of ZnO/rGO nanocomposites with various mass ratios (0.5–1.5%) of rGO to ZnO using the ultrasonic spray solvothermal process. For comparisons, the traditional precipitation method was used to synthesize p-ZnO/GO-0.5. Pure ZnO nanocrystals without GO was also synthesized.

Characterization of ZnO/rGO Nanocomposites

The phases and chemical compositions of the ZnO/rGO nanocomposites were first characterized using the techniques of X-ray diffraction (XRD), Raman, and X-ray photoelectron spectroscopy. Figure shows the typical XRD patterns of GO, ZnO, and ZnO/rGO samples. One can see that ZnO and ZnO/rGO nanocomposites with various rGO contents have peaks at Bragg angles (2θ) of 31.7°, 34.4°, and 36.2°, assignable to the (110), (002), and (101) planes of the hexagonal wurtzite ZnO (JCPDS no. 79-206), respectively.[48] The XRD peaks of the ZnO/rGO-0.5 nanocomposite are sharper and stronger than those of the ZnO/rGO-1.5 but weaker than those of the pure ZnO sample. It seems to suggest that the addition of GO influences the crystallinity of the ZnO/rGO nanocomposites. For the ZnO growth, rGO nanosheets act as the nuclei through a heterogeneous nucleation and growth process. For GO (Figure a), the broadened XRD peak at ∼23° is the characteristic of the (002) plane of GO. The strong peak at ∼10° belongs to the layered structure of GO restacking. Thermogravimetric analysis can prove the existence of carbon (Figure S1). The weak XRD peak at 43° in GO disappears in the ZnO/rGO samples because of the small content and high dispersion of GO in the ZnO/rGO nanocomposites.

Figure 2

(a) XRD patterns and (b) Raman spectra of the ZnO/rGO nanocomposites: (A) GO, (B) ZnO, (C) ZnO/rGO-0.5, (D) ZnO/rGO-1, and (E) ZnO/rGO-1.5; (c) detailed Raman spectrum of ZnO nanocrystals.

The presence of rGO in the ZnO/rGO nanocomposites was further confirmed by Raman spectra. Figure b shows the typical Raman spectra of GO, ZnO, and ZnO/rGO nanocomposites, and their detailed Raman data are summarized in Table . The pristine GO and ZnO/rGO nanocomposites exhibit two characteristic peaks at ∼1345 and 1591–1598 cm–1, belonging to the D band and the G band of graphene, respectively; the D band corresponds to the breathing mode of k-point phonons of the A1g symmetry near the edges of graphitic structures (local defects or disorder)[50] and the G band belongs to the E2g mode of sp2 domains of the rGO species.[51] The ratio (ID/IG) of the D (at 1346 cm–1) and G (at 1598 cm–1) bands of GO is about 1.01; the ID/IG ratios of the ZnO/rGO samples increase to ≥1.07 (Table ). The ZnO/rGO-0.5 sample shows the D and G bands at 1346 and 1593 cm–1, respectively. The G band increases from 1593 to 1596 cm–1 as the rGO amount increases from 0.5 to 1.5%. In the hydrothermal process, numerous functional groups on GO, like carboxyl groups and hydroxyl groups, act as reactive sites to immobilize Zn2+. The Raman peaks (D and G bands) verify the existence of rGO in the ZnO/rGO samples.[49,52] In the ZnO/rGO-x samples (x = 0.5, 1 and 1.5), the second-order D (2D) bands at ∼2667 cm–1 and D + G bands at ∼2918 cm–1, belonging to rGO, reconfirm the formation of rGO in the in situ solvothermal process. The 2D band is related to the number of layers in the rGO, and an intense and sharp band means monolayer rGO species.[53] The 2D/G intensity ratios of single-, double-, triple-, and multi-layer graphene sheets are >1.6, ≈0.8, ≈0.30, and ≈0.07, respectively.[54] In this work, the 2D/G intensity ratio is found to be 0.11 for the ZnO/rGO-0.5 nanocomposite, indicating that the number of rGO layers is between three and four layers. The wide bands at 439–1200 cm–1 belonging to ZnO can be found in the ZnO/rGO nanocomposites.

Table 1

Raman Data, Morphology, and NO2-Sensing Properties of the ZnO and ZnO/rGO Samples

sample	position of D band/cm–1	position of G band/cm–1	ID/IG	morphology	optimal operating temperature/°C	response at NO2 50 ppm	response/recovery time/s
pristine GO	1346	1598	1.01	sheet
ZnO				sheet	∼160	30	10/23
p-ZnO/rGO-0.5	1345	1591	1.06	rod	∼130	18	5/22
ZnO/rGO-0.5	1346	1593	1.05	small rod	∼130	62	3/10
ZnO/rGO-1	1345	1594	1.07	small rod	∼130	31	17/46
ZnO/rGO-1.5	1344	1596	1.08	small rod	∼130	29	15/66

Figure c shows the Raman spectrum of pure ZnO. The single-crystal wurtzite structure of ZnO belongs to the C64 (P63mc) symmetry group, there are two identical formula units in each primitive cell, and all the atoms in the primitive cell occupy the 2b position with the C3 symmetry.[55] The phonon dispersion relationship of ZnO consists of 12 branches in the Brillouin zone (q = 0), and one can deduce that Γ = 2 × (A1 + B1 + E1 + E2) according to the group theory.[56] In these modes, the Γ points of the Brillouin zone are divided into eight groups, which are A1(TO), A1(LO), B1L, B1H, E1(TO), E1(LO), E2H, and E2L, where the central optical mode is Γopt = A1 + 2B1 + E1 + 2E2 and the phonon mode is Γaco = A1 + E1.[57] In these optical phonon modes, A1 + E1 + 2E2 is Raman-activated; A1 and E1 are both Raman and infrared-activated; for low-frequency and high-frequency B1 is Raman inactive mode.[58] Because A1 and E1 are polar, they can be divided into transverse optical (TO) and longitudinal optical (LO) phonons. Raman-active E2 has high (E2high) and low-frequency (E2low) phonon modes, which are respectively related to the vibration of oxygen atoms and Zn sublattices.[59,60]

XPS is an important tool to get a better understanding of the chemical state and composition of the ZnO/rGO nanocomposite. Figure shows the XPS spectra of ZnO/rGO-0.5. The survey spectrum in Figure a indicates that Zn, O, and C are the major elements. Figure b shows the Zn 2p spectrum, with two bands at 1045.0 and 1021.8 eV, corresponding to Zn 2p1/2 and Zn 2p3/2 in the form of ZnO, respectively.[61] As Figure c shows that the C 1s spectrum can be subdivided into four peaks at 289.6, 287.8, 285.8, and 284.5 eV, corresponding to O–C=C, C–O–C/C=O, C–O, and C–C/C=C (attributed to the sp2/sp3 carbon atom) functional groups of rGO, respectively.[62−65] The O 1s XPS peak shown in Figure d can be decomposed into three Gaussian components at ∼532.6, 531.9, and 531.0 eV; the bands at 531.9 and 531.0 eV can be indexed to chemisorbed oxygen (OC) species and lattice oxygen (OL) in the wurtzite structure of hexagonal ZnO, respectively;[66] the band at 532.6 eV can be indexed to the C=O bond in the ZnO/rGO nanocomposite.[67]

Figure 3

XPS spectra of ZnO/rGO-0.5: (a) survey scan, (b) Zn 2p, (c) C 1s, and (d) O 1s.

Scanning electron microscopy (SEM) observations were conducted to investigate the morphology and microstructure of the GO, ZnO, and ZnO/rGO nanocomposites, especially in comparing the ultrasonic spray method with the traditional precipitation one. Figure shows the typical results. The typical SEM image of the pristine GO in Figure a indicates that the GO consists of restacking layers with a large area. The SEM image (Figure b) of pure ZnO formed via the ultrasonic spray process shows that the ZnO sample consists of apparent sheetlike particles of 100–800 nm in size. Figure c,d shows the SEM images of the p-ZnO/rGO-0.5 and ZnO/rGO-0.5 nanocomposites, respectively. One can see that the morphology of ZnO nanocrystals changes from a sheet-like shape to a rod one when GO is added. The rodlike petals of the ZnO/rGO-0.5 nanocomposite have an average length of about 500 nm, which is about two-thirds of the size of the p-ZnO/rGO-0.5 nanocomposite (about 800 nm). Figure e is a high-magnification SEM image of a ZnO/rGO-0.5 material. We can identify that the ultrasonic spray method effectively overcomes the agglomeration of ZnO nanoparticles and reduces their particle sizes. From Figure c,e, one can see that the graphene sheets cover the dispersed ZnO nanorods, marked by arrows. This morphology allows the formation of a Schottky junction between rGO and ZnO, enabling electrons to move freely to different ZnO nanocrystals, and then improving the gas-sensing properties of the ZnO/rGO nanocomposites. Figure f shows the SEM image of the ZnO/rGO-1.5 sample. With more rGO added, flexible rGO sheets cover on the surface of ZnO or mixed among ZnO flowers with visible edges. Comparing Figure d,f, there is no obvious change in the grain size of the ZnO rods; however, more rGO results in agglomerative ZnO petals. When comparing the ultrasonic spray and traditional precipitation methods, one can find that the ultrasonic spray method is helpful to make GO sheets well dispersing and wrapping in ZnO nanorods.

Figure 4

(a–f) SEM images of the ZnO/rGO nanocomposites: (a) GO, (b) ZnO, (c) ZnO/rGO-0.5 by the precipitation process, (d,e) ZnO/rGO-0.5 by the ultrasonic spray process, and (f) ZnO/rGO-1.5 by the ultrasonic spray process; (g–j) TEM images of ZnO/rGO-0.5 obtained by the ultrasonic spray process: (g,h) Low-magnification TEM images, and (i,j) HRTEM images.

Transmission electron microscopy (TEM) was further used to observe the microstructure of the ZnO/rGO nanocomposite. Figure g–j shows the typical TEM and high resolution TEM (HRTEM) images of the ZnO/rGO-0.5 sample. In Figure g, a low-magnification TEM image shows that ZnO nanocrystals grow between rGO nanosheets, and the edge outlines of rGO (marked by arrows) can be seen as the convincing evidence. The high-magnification TEM image in Figure h suggests that the ZnO/rGO-0.5 sample consists of ZnO nanorods tightly wrapped by rGO to form a core/shell-like configuration. The HRTEM image in Figure i shows a clear view of the interface between ZnO and graphene. The local enlargement of the interface is shown in Figure j, and one can easily find two types of lattice fringes with different arrangement directions. The lattice fringes with a d-spacing of ∼0.202 and ∼0.34 nm correspond to the (101) plane of wurtzite hexagonal structure ZnO and the (002) plane of rGO, respectively.

Figure shows the N2 adsorption–desorption isotherms and pore-size distribution curves of the ZnO/rGO-0.5 and p-ZnO/rGO-0.5 samples synthesized by the traditional precipitation and ultrasonic spray methods to compare their specific surface areas and pore-size distributions. The adsorption–desorption isotherms of p-ZnO/rGO-0.5 are close to a type II shape (B in Figure a).[68,69] As the interaction between adsorbate molecules is stronger than that between the adsorbate and the adsorbent, it is difficult to adsorb the adsorbate during the initial adsorption; the isotherms show a concave shape, and the amount of adsorbed gas increases with the increase in the partial pressure (P/P0). The ZnO/rGO-0.5 sample synthesized by the ultrasonic spray method exhibits a type IV pattern with an H3 hysteresis loop (A in Figure a), which indicates typical physical adsorption on nonporous or macroporous adsorbents.[68,70] There is a strong interaction on the adsorbate surface, so the isotherm presents a convex shape. The H3 hysteresis loop isotherm in the type IV pattern isotherm and no obvious saturated adsorption platform indicate that the pore structure of ZnO/rGO-0.5 is very irregular, corresponding to sheetlike, rod-shaped materials, consistent with the observed microscopic morphology (Figure ). Because of the presence of fissure pores in the ZnO/rGO nanocomposites, no adsorption saturation is observed as the relative pressure increased.[68] According to their N2 physisorption measurement, the Brunauer–Emmett–Teller (BET) surface areas of the ZnO/rGO and p-ZnO/rGO-0.5 samples are 32.4 and 4.6 m2/g, respectively. Figure b shows the pore-size distribution curves of the ZnO/rGO and p-ZnO/rGO-0.5 samples. One can see that the ZnO/rGO sample obtained by the ultrasonic spray has a wide pore range of 2–200 nm and a Barrett–Joyner–Halenda (BJH) desorption cumulative pore volume of 0.112 cm3/g, whereas the p-ZnO/rGO-0.5 sample obtained via the traditional precipitation method shows less pores and a small BJH desorption cumulative pore volume of 0.013 cm3/g. According to the BJH model and the desorption result, the average pore diameters (i.e., 4V/A) of the samples of p-ZnO/rGO-0.5 and ZnO/rGO-0.5 are 12.7 and 15.3 nm, respectively.[71] The nitrogen adsorption–desorption isotherms reveal that the ultrasonic spray method is efficient to form porous and high-surface area nanocomposites.

Figure 5

(a) Nitrogen adsorption–desorption isotherms and (b) pore-size distribution curves of (A) ultrasonic spray-derived ZnO/rGO-0.5 and (B) precipitation-derived ZnO/rGO-0.5.

Taking SEM/TEM observations and N2 adsorption–desorption isotherms into account, we can conclude that GO nanosheets and the ultrasonic spray process highly influence the microstructure of the ZnO species in the ZnO/rGO-x nanocomposites, which are of a porous, high-surface area structure with ZnO nanocrystals (400–600 nm). The pure ZnO sample consists of sheetlike particles (0.8–1 μm), and the p-ZnO/rGO-0.5 sample obtained by traditional precipitation exhibits a nonporous and low-surface area structure.

Gas-Sensing Properties of ZnO/rGO Nanocomposites

The ZnO/rGO-x samples (x = 0, 0.5, 1, 1.5) obtained via the ultrasonic spray and precipitation methods are used to prepare sensors.

The NO2-sensing properties operating at various temperatures (70–200 °C) upon exposure to NO2 with various concentrations (0.5–50 ppm) were first tested. Figure shows the R–t curves and the (Ra/Rg)–C response curves. When the operating temperature is 70 °C, as shown in Figure a, the ZnO/rGO-0.5 sample exhibits an obvious response to the NO2 gas with a low concentration of ≤10 ppm, but the pure ZnO shows a weak response even to a 100 ppm NO2 gas. As shown in Figure b, the ZnO/rGO-x (x = 0.25, 0.5, 1) sensors exhibit a sharp NO2-sensing response, when the operating temperature increases to 100 °C. As the operating temperature increases from 130 to 160 °C, the ZnO/rGO-0.5 sensor exhibits the highest response, and the ZnO/rGO-1 nanocomposite also shows an increasing NO2-sensing response (Figure c,d). When the operating temperature increases to 200 °C, as shown in Figure e, the pure ZnO, ZnO/rGO-1, and ZnO/rGO-1.5 samples show a NO2-sensing response approaching to the ZnO/rGO-0.5 sample. From Figure , we can see that a small amount of rGO highly improves the NO2-sensing property of the ZnO/rGO-0.5 sensor at a low operating temperature (i.e., 70 °C), whereas the pure ZnO sensor shows a low NO2-sensing response testing at 70–200 °C. However, excessive rGO can reduce the NO2-sensing response of ZnO/rGO samples (Figure S6). We compared the NO2-sensing response (Ra/Rg) of the ZnO/rGO-x samples (x = 0, 0.5, 1, 1.5) as a function of operating temperature (Figure S2). The NO2-sensing of the pure ZnO sample and ZnO/rGO-x showed a trend of rising first and then decreasing with the increase of operating temperature. The response of the pure ZnO sample is much lower than that of the GO-doped ZnO samples, indicating that the addition of GO nanosheets improved the gas-sensing properties especially at a low operating temperature.

Figure 6

NO2-sensing curves of the ZnO/rGO nanocomposites (ZnO/rGO-x, x = 0, 0.5, 1, 1.5) at different operating temperatures: (a) 70, (b) 100, (c) 130, (d) 160, and (e) 200 °C.

Because the ZnO/rGO-0.5 sample shows the best NO2-sensing response, we compare their R–t curves under various operating temperatures (Figure S3). The optimum operating temperature is 130–160 °C, and the NO2-sensing response operating at 130 °C reaches a summit of 62. We investigated the cycling stability of the ZnO/rGO-0.5 nanocomposite upon exposure to NO2 with various concentrations (0.1–50 ppm) operating at 130 °C. Figure shows the typical R–t and (Ra/Rg)–CNO curves of the ZnO/rGO-0.5 samples. The ZnO/rGO sensors show an obvious response to NO2 with a low concentration of 0.1 ppm. The response (Ra/Rg) of the ZnO/rGO-0.5 sensor shows a quadratic function relation with the concentration of NO2 (CNO/ppm): Ra/Rg = 5 + 5.38CNO0.6 (R2 = 0.998).

Figure 7

(a) R–t curve of the ZnO/rGO-0.5 sensor exposure to NO2 gases with various concentrations (0.01–50 ppm) operating at 130 °C; (b) experimental and fitting plots of response (Ra/Rg) vs concentration of NO2 gas for the ZnO/rGO-0.5 sensor operating at 130 °C.

We calculated the response and recovery times according to their R–t response curves, as shown in Figure S4. Figure S4a shows the response and recovery times of the ZnO/rGO sensors upon exposure to 50 ppm NO2 at 130 °C, according to their R–t curves in Figure c. Their response times are less than 17 s, and their recovery times are 16–66 s. It is notable that the ZnO/rGO-0.5 sample exhibits the shortest response/recovery time (less than 20 s at 130 °C). The response and recovery times (Figure S4b,c) of the pure ZnO and ZnO/rGO-0.5 sensors operating at 130 °C are calculated according to the R–t curves in Figure a. We see that the response times of pure ZnO nanocrystals up to 93 s at a low NO2 gas concentration, and the response time is shortened as the NO2 gas concentration increases. The response time of the ZnO/rGO-0.5 sample is much less than the pure ZnO sample, and the effect of the NO2 concentration on the response time (approximately 20 s) is weaker when the NO2 gas concentration is above 1 ppm. The recovery time of the pure ZnO sample exceeded 40 s upon exposure to NO2 gas because of its low desorption rate. Compared with the pure ZnO sensor, the ZnO/rGO-0.5 sensor shows a relatively short recovery time (Figure S4b,c).

The response of the ZnO/rGO-0.5 sensor to VOCs, CO, H2, H2S, and SO2 was also checked. Figure a shows the typical R–t curves with a 50 ppm gas or vapor at 130 °C. This sensor shows a higher and more rapid response to NO2 than the other gases. Figure b compares the response upon exposure to these gases. The histogram clearly shows that the ZnO/rGO-0.5 sensor has an excellent selective response to NO2 gas.

Figure 8

Comparisons of (a) the R–t curves and (b) their corresponding responses of the ZnO/rGO-0.5 sensor upon exposure to various types of vapors/gases (50 ppm) operating at 130 °C.

Low gas concentration detection is a good performance of gas-sensing materials. The experimental conditions allow the detection gas concentration to be 10 ppb. At an operating temperature of 130 °C, we tested the response of ZnO/rGO-0.5, p-ZnO/rGO-0.5, and pure ZnO to NO2 gas at a concentration of 10 ppb, as shown in Figure a. The pure ZnO curve is straight, which does not respond to NO2 gas at low concentrations; when rGO is doped to ZnO, the composite responds to low-concentration NO2 gas; the ZnO/rGO composites synthesized by the ultrasonic spray method have better response to low concentration NO2 gas, and the response value is about four times that of the ZnO/rGO composite synthesized by the precipitation method. The doping of rGO is the key factor for the gas-sensitive material’s response to NO2 gas. The ultrasonic spray method changes the morphology of the material, forming a large specific surface area and a large number of active sites, thereby increasing the response of ZnO/rGO composites for NO2 gas.

Figure 9

(a) R–t curve of ZnO/rGO-0.5 (A), p-ZnO/rGO-0.5 (B), and ZnO (C) for NO2 gas response at a concentration of 10 ppb; (b) comparisons of the responses of the ZnO/rGO-0.5 sensor obtained by different synthetic methods: ultrasonic spray and traditional precipitation.

In order to demonstrate the advantage of the ultrasonic spray method in the synthesis of gas-sensing active materials of ZnO/rGO, Figures b and S5 compare the NO2-sensing performance of the sensors based on ZnO/rGO-0.5 (by ultrasonic spray) and p-ZnO/rGO-0.5 (by traditional precipitation) from a low concentration of 0.5 ppm to a high concentration of 50 ppm operating at 130 °C. The responses of the ZnO/rGO-0.5 and p-ZnO/rGO-0.5 sensors increase with the increase in gas concentration. However, the p-ZnO/rGO-0.5 sensor shows a much weaker response than the ZnO/rGO-0.5 sensor. The enhancement in the NO2-sensing performance of the ZnO/rGO-0.5 nanocomposite synthesized by the ultrasonic spray method should be attributed to its higher surface area and a more porous microstructure.

Possible Mechanism of ZnO/rGO Sensors to NO2 Gas

The ZnO/rGO-0.5 sample with a small amount of rGO (i.e., 0.5%) exhibits a much higher and quicker response to NO2 gas than the pure ZnO nanocrystals. We try to understand the possible reason on the basis of the synergetic effect of ZnO and rGO. The ZnO nanorods behave as the key sensing material, and rGO nanosheets serve as highly conductive channels.

The ZnO/rGO nanocomposite obtained here consists of a majority of ZnO nanorods covered by a small number of rGO nanosheets. A typical Schottky junction between ZnO and rGO can be formed because of their suitable difference of work function, as shown in Figure a. As an n-type semiconductor, the gas-sensing mechanism of ZnO-based sensors involves the adsorption of oxygen species in air,[72] and then the resistance changes before and after being exposed to a target gas.[73−75] GO can enhance the specific surface area of the ZnO/rGO active materials (see Figure ) and promote the adsorption of oxygen species and target gases. The resistance of the ZnO/rGO sensor is influenced by the adsorption of oxygen species. In the air, the oxygen species adsorbed on the ZnO/rGO surface experiences a series of reactions (see eq ).[76] At an elevated operating temperature, the equilibria shift to the right. The resistance of a ZnO/rGO sensor exhibits a platform value because the oxygen captures electrons from ZnO/rGO to form a space-charge layer (Figure b). When the sensors are exposed to an oxidizing target gas such as nitrogen dioxide, the interaction between the oxidizing gas and the electron on the surface of the sensors generates adsorbed nitrogen dioxide gas molecules that can be represented as eq .[77] Because NO2 has a higher electron affinity (2.28 eV) than the pre-adsorbed oxygen (0.43 eV), the adsorbed NO2 molecules can capture the remaining electrons of O– and O2– ions absorbed on the surface of the ZnO/rGO nanocomposites to produce O–, O2– ions and nitric monoxide, as shown in eqs and 4.[78] During the above process, the carriers (i.e., electrons) of the ZnO/rGO nanocomposite participate in the reactions, and then result in a lack of carriers to further enlarge the thickness of the depletion layer, forming a higher resistance state, as shown in Figure c.

Figure 10

Possible mechanism of the ZnO/rGO nanocomposite for NO2 sensing. (a) Typical Schottky junction between ZnO and rGO can be formed because of their suitable difference of work function. (b) Oxygen captures electrons from ZnO/rGO and forms a space-charge layer. (c) ZnO/rGO is exposed to NO2 gas, enlarging the depletion layer and forming a higher resistivity state.

The work function of ZnO is 4.1 eV, close to its conduction band,[79,80] and the work function of rGO is about 4.75 eV.[35,36] This energy band structure exhibits some similarity to the semiconductor/metal heterojunction (Figure a).[37] In the contacted ZnO/rGO junction (Figure b), electrons will go from ZnO to rGO to form an equilibrium at the Fermi level, forming a Schottky junction in the ZnO/rGO interface. As NO2 molecules react with ZnO/rGO, electrons transfer from ZnO/rGO to NO2, forming an increase of the Schottky barrier (Figure c). As a result, a sharp rise in the resistance of the ZnO/rGO nanocomposite occurs because of the difficulty for the electrons to flow and pass through the Schottky barrier.

The selective gas-sensing property of the ZnO/rGO nanocomposite to NO2 at a low operating temperature can be associated with the higher NO2 adsorption and oxidation ability.[81,82] At a low operating temperature, the adsorbed oxygen exists as O2– and O–,[76] and the reducing gases (e.g., H2, H2S, and CH4) cannot react with O2– and O– because of their low reactivity. However for the case of electrophilic NO2, it directly captures free electrons from ZnO/rGO via the reactions (eqs and 4). This point may account for the selective response of the ZnO/rGO sensor to NO2 at a low temperature.

In the ZnO/rGO nanocomposite, the conduction path is mainly controlled by ZnO nanorods because of the small amount of rGO, but the addition of rGO obviously influences the electrical transport properties and the microstructure of ZnO/rGO nanocomposites. All in all, we can understand the possible reasons why and how the rGO nanosheets influence the gas-sensing performance of the ZnO/rGO nanocomposites from several aspects. First, the addition of rGO changes the morphology of ZnO nanocrystals from a sheetlike shape to rodlike one (Figure ), forming a wrapped ZnO/rGO structure which improves the NO2-sensing response. Second, some micro-area conductive networks and rGO/ZnO Schottky junctions accelerate electric transfer to promote the low-temperature NO2 detection. Third, the method of ultrasonic spray is the key to synthesize the ZnO/rGO nanocomposites with a high specific surface area and unique porous structure (Figures and 5). The novel ultrasonic spray method can make rGO nanosheets and ZnO nanorods fully contact and combine one another to form efficient rGO/ZnO Schottky junctions, which provide more active sites for the detection of NO2 and other gases (or vapors).

Conclusions

In this work, a novel USS method has been developed to synthesize ZnO/rGO nanocomposites with unique core/shell Schottky junctions. The gas ZnO/rGO sensors exhibit a highly selective response to nitrogen dioxide with a low concentration (ppb-/ppm-levels) operating a low temperature of 70–130 °C. Comparing the gas-sensing performance of the ZnO/rGO nanocomposites obtained by ultrasonic spray and traditional precipitation, the ultrasonic spray method has highly promoted the gas-sensing performance of the ZnO/rGO nanocomposites. The possible reasons for this promotion have been attributed to the formation of ZnO/rGO Schottky junctions and the effective regulation in their microstructures (i.e., specific surface area, porous structure, and interface). The ultrasonic spray method has been proved to be an efficient strategy to synthesize multiphase and multicomponent nanocomposites with controllable micro-area interfaces and morphologies for the high-performance applications in gas-sensing and other functional devices.

Experimental Section

Chemical Reagents and Setups

All chemical reagents were of analytical grade and used as received without further purification. Zinc nitrate hexahydrate (Zn(NO3)2·6H2O) and potassium hydroxide (KOH) were obtained from Wind Boat Chemical Reagent Technology Co. Ltd (Tianjin, China). GO was purchased from Huicheng Graphene Technology Application Co. Ltd (Wuxi, China). Deionized water was used in the experiments. An ultrasonic sprayer (XO-1800W, Nanjing Sino Instrument Manufacturing Co. Ltd, China) and a freeze dryer (FD8-3, SIM International Group Co. Ltd, USA) were used in the synthesis.

Synthesis of ZnO2/rGO Samples

ZnO/rGO nanocomposites were prepared via an ultrasonic spray-assisted hydrothermal process (Figure ). The mass ratio of GO to ZnO was theoretically designed to be 0, 0.5, 1, and 1.5%, and the resulting samples were named ZnO, ZnO/rGO-0.5, ZnO/rGO-1, and ZnO/rGO-1.5, respectively. Typically, 1 g/L of GO aq solution was prepared by dispersing 100 mg of GO in 100 mL of H2O with ultrasonic treatment for 6 h; meanwhile, 0.15 mol/L of Zn(NO3)2 aq solution was prepared by dissolving 44.62 g of Zn(NO3)2·6H2O in 1000 mL of H2O. After the above GO (6.15 mL) and Zn2+ solution (100 mL) were mixed by magnetic stirring for about 3 h, the GO/Zn2+ mixture obtained was transferred into a metering pump (20 rpm) which flowed to the ultrasonic spray nozzle. A three-necked flask with a KOH solution (pH = 13.5, 50 mL) was used as the reactor to which the GO/Zn2+ mixture was injected via the ultrasonic-spray manner. The three-necked flask was kept in an ice bath and magnetically stirred for 2 h. The resulting GO/Zn(OH)2 suspension was finally transferred into a Teflon-lined steel autoclave (100 mL), which was heated at 180 °C for 12 h. The solids were collected by filtration and washing four times using water and ethanol. The as-obtained solid was then freeze-dried and the ZnO/GO-0.5 nanocomposite was synthesized. The other ZnO/GO samples with various GO contents were prepared by a similar process.

For purposes of comparison, ZnO with 0.5% of GO was also synthesized via a traditional precipitation method, and the resulting sample was named p-ZnO/GO-0.5. The precipitation method just did not involve the ultrasonic spray process and the other steps were similar to the ultrasonic spray-assisted hydrothermal process.

Characterization of ZnO/rGO Samples

Raman spectra were recorded on a LabRAM HR Evolution spectrometer (HORIBA Jobin Yvon, France) using the 633 nm line as the excitation source at room temperature. XRD patterns were recorded on a D/MAX 2500 XRD diffractometer (Rigaku) with a Cu Kα radiation (λ = 0.15406 nm). XPS spectra were performed on an ESCALAB 250xi spectroscope (Thermo Scientific Ltd., England) using an Al Kα (1486.6 eV) radiation source. The SEM and TEM images were recorded by SEM (JSM-7001F JEOL, Japan) and TEM (Tecnai G2 F20 S-TWIN, FEI, USA). The specific surface areas were tested on an ASAP 2460 (Micromeritics, USA) machine at a liquid nitrogen temperature of −196 °C.

Gas-Sensing Test of ZnO/rGO Nanocomposites

The gas sensors using the ZnO/rGO nanocomposites as active materials were fabricated according to the above work. The ZnO/rGO nanocomposites were first mixed with a small amount of ethanol to form the paste in an agate mortar. Then, the as-obtained ZnO/rGO pastes were brush-coated onto the surface of Al2O3 microtubes with Au electrodes. After the ZnO/rGO coatings were air-dried, the coating process was repeated until a continuous coating was formed. Each ZnO/rGO-coated Al2O3 microtube was then fixed to the special pedestal with six poles by welding the four Pt electrodes to four poles of the pedestal. A heating coil (Ni–Cr) was then inserted through the Al2O3 microtube and its two ends were welded with the other two poles of the pedestal. A schematic diagram of the ZnO/rGO sensor was shown in Figure . The sensors with pure ZnO nanocrystals were fabricated used a similar process. The gas-sensing tests were conducted on a WS-30A system from Zhengzhou Winsen Electronics Technology Co., Ltd (Zhengzhou, China). NO2 was used as the target gas to evaluate the gas-sensing performance of the ZnO/rGO sensors. To investigate the selectivity of the ZnO/rGO sensors, some organic vapors (e.g., methanol, ethanol, and acetone), H2, SO2, H2S, and CO were also used as the target-sensing substances. The target gases were sampled using a syringe-like sampler with a volume range of 0.1–900 μL. According to the densities of liquid substances and the chamber volume, the concentrations (0.01–50 ppm) of the target substances were calculated. The relative humidity (RH) was 35% and the working temperatures were 70–200 °C. The ZnO/rGO sensor (R) was connected in series with the load resistor (R0) with a known resistance, and a source voltage (U0) of 5 V was loaded on the circuit. The system recorded the voltage (U) loaded on R0, and the resistance (R) of the ZnO/rGO sensor can be therefore calculated according to eq . For a reducing gas (or vapor) and an n-type semiconductor sensor, the sensor response (Sr) is defined as Ra/Rg, where Ra is the baseline resistance in air and Rg is the resistance in target gas, as shown in eq . The sensor response (Sr) is defined as Rg/Ra, as shown in eq , for an oxidizing gas and an n-type semiconductor. The response time (τres) is defined as the time required to reach 90% of the final response upon target gas exposure. Similarly, the recovery time (τrec) is defined as the time interval over which the sensor response drops to 10% of the stabilized response in the target gas when placed in clean air.