Abstract

Oxide semiconductor In2O3 has been extensively used as a gas sensing material for the detection of various toxic gases. However, the pure In2O3 sensor is always suffering from its low sensitivity. In the present study, a dramatic enhancement of sensing characteristic of cubic In2O3 was achieved by deliberately fabricating p-n heterojunction-like NiO/In2O3 composite microparticles as sensor material. The NiO-decorated In2O3 p-n heterojunction-like sensors were prepared through the hydrothermal transformation method. The as-synthesized products were characterized using SEM-EDS, XRD, and FT-IR, and their gas sensing characteristics were investigated by detecting the gas response. The experimental results showed that the response of the NiO/In2O3 sensors to 600 ppm methanal was 85.5 at 260°C, revealing a dramatic enhancement over the pure In2O3 cubes (21.1 at 260°C). Further, a selective detection of methanol with inappreciable cross-response to other gases, like formaldehyde, benzene, methylbenzene, trichloromethane, ethanol, and ammonia, was achieved. The cause for the enhanced gas response was discussed in detailed. In view of the facile method of fabrication of such composite sensors and the superior gas response performance of samples, the cubic p-n heterojunction-like NiO/In2O3 sensors present to be a promising and viable strategy for the detection of indoor air pollution.

1. Introduction

In the past years, indoor air pollution has been realized by many people with respect to public health. Methanal, as typical air pollution indoor, is seriously toxic to human beings and animals even at very low concentration [1]. As a result, the strategies to effectively detect low concentration methanol are of great significance and in huge demand in practical applications. Currently the most commonly used technologies for detecting the low concentration methanal in air include fluorimetry [2], colorimetry [3], high-performance liquid chromatography (HPLC) [4], gas chromatography (GC) [5], and infrared detection [6]. In contrast to such traditionally available approaches for detecting methanal, another alternative monitor using oxide semiconductor as sensor has been recently applied due to its various superiorities, such as low cost, good stability, and short response time [710]. Typically, a series of oxide semiconductor-based gas sensors, like SnO2 [11], TiO2 [12], In2O3 [13], and Fe2O3 [14], have been reported to act as gas sensors for detecting methanal.

Presence of space charge layers in semiconductors is a common phenomenon within a heterojunction structure once two semiconductor materials with different band gaps are intimately contacted [15]. In such formed heterojunction layer, free electrons in one material with higher Fermi energy could migrate into the adjacent material spontaneously due to the demand of band alignment until their Fermi levels allied at the same energy, resulting in the opposite charges at two sides of heterostructure interface. In a word, a unique space charge layer, which contains opposite charges within the different semiconductor component, is apt to be generated [16]. More specifically, in n-type oxide semiconductors, the adsorption of electronegative oxygen generates an electron depletion layer near the semiconductors surface. Thereby, sensor resistance is rest with the close-knit connection between the resistive electron depletion layer and semiconducting cores. In p-type oxide semiconductors, the ionized adsorption of oxygen can create the hole accumulation layer through electrostatic interaction between positively charged holes and negatively charged oxygen. In such interaction circumstances, the sensor resistance strongly depends on the competition of conduction along the cabined cross-sectional area of hole accumulation layer near-surface and that across resistive cores with a wide cross-sectional area [17, 18]. In contrast, the gas response of n-type oxide semiconductors is often relatively higher than that of p-type oxide semiconductors because of the imparity in their gas sensing mechanisms [19]. However, p-type oxide semiconductors are ideally suitable for the designing of novel functionality in high-performance gas sensors, since they own a distinctive catalytic activity with multifarious volatile organic compounds [20]. By this token, the combination of n-type oxide semiconductors with p-type oxide semiconductors to design a novel hybrid or composite semiconductor material is attracting increasing attention as it enables new functional properties, which are not possible in their own starting components. Typically, some admirable candidates have been offered currently by the heterostructured formation of oxide-oxide p-n junctions, improving functional properties in comparison with the pure parent substrate. For instance, p-n heterojunction-like ZnO/TiO2 [21], WO3/TiO2 [22], NiO/TiO2 [23], CuBi2O4/TiO2 [24], and In2O3/ZnO [25] have been synthesized triumphantly.

In2O3, a classical n-type III–V semiconductor with a wide band gap (3.6 eV), displays a predominant sensing performance towards a broad category of gases, such as H2S [26], CO [27], C3H6O [28], and C2H5OH [29]. Hence, In2O3-based sensors show an efficient sensing performance in different conditions. For example, Sun et al. pointed out that In2O3 was a highly sensing material for the detection of various gases, and the adding of palladium and platinum was proved to be a simple and efficient route to enhance the sensing properties [30]. Wang and coworkers reported the Er-doped In2O3 nanotube, and the response of the Er-doped In2O3 nanotube sensors to 20 ppm methanal was 12 at 260°C in the subsequent gas sensing investigation, which was very promising for the detection of dilute methanal gas in practice [31]. NiO has been widely used in sensors [32], antiferromagnetic devices [33], fuel cells [34], and dye-sensitized photocathodes [35] due to its simple structure, electrochemical stability, high durability, and low cost. As a typical p-type semiconductor, NiO is also a benign candidate to design the new functionality in high-performance gas sensors. For instance, Ju et al. reported NiO/SnO2 p–n junction which was formed by depositing NiO nanoparticles onto the surface of SnO2 hollow sphere sensors, and the response of NiO/SnO2 sensor was much higher than that of pristine SnO2 hollow spheres [36]. Li and coworkers have synthesized the p-NiO/n-ZnO nanowire heterojunction structure for the applications in UV sensors [37]. In comparison with the great success above-mentioned, the design of high-performance NiO/In2O3 oxide-oxide semiconductor gas sensors with p-n heterojunction-like structure is still very much in the early stages of investigation. Particularly, no one has reported the synthesis and gas sensing performance of cubic NiO/In2O3 composites with NiO as loaded material.

Based on the above background, in the present study p-n heterojunction-like NiO/In2O3 microparticles were prepared through a hydrothermal synthesis method. The as-synthesized product was characterized using SEM-EDS, XRD, and FT-IR. Then, a comparative gas sensing study was performed to illuminate the conspicuous sensing properties of the NiO/In2O3 sensors. The experimental results demonstrated that a dramatic enhancement of sensing characteristic of cubic In2O3 has been achieved in comparison with the single In2O3 materials, and the cause for the improvement of gas sensing property to methanal can be assigned to the deliberate surface ornament of cubic In2O3 substrate with petal-shaped NiO nanoparticles. In view of the facile method of fabrication of the composite sensors and the superior gas response performance, we believed that the p-n heterojunction-like NiO/In2O3 sensors present to be a promising and viable strategy for the detection of indoor methanal pollution.

2. Experimental

2.1. Synthesis of NiO/In2O3

Originally, the In(OH)3 and Ni(OH)2 were prepared by a hydrothermal reaction process. In detail, a spot of InCl3 was firstly dissolved in 100 mL distilled water under vigorous stirring at room temperature. Then, the moderate sodium dodecyl benzene sulfonate (SDBS) was added to the solution. Next, the mixture was transferred to a teflon-lined stainless steel autoclave, sealed, and maintained at 473.15 K for 24.0 h. The autoclave was allowed to cool down to room temperature naturally. The resulting white precipitate (In(OH)3) was filtered and washed with distilled water. For the Ni(OH)2, solid NiCl2 was firstly dissolved in 160 mL distilled water, and the urea was added to the solution under vigorous stirring. The mixture was transferred to the stainless steel autoclave, sealed, and maintained at 473.15 K for 24.0 h. The green precipitate (Ni(OH)2) was obtained. Secondly, the preparation of Ni(OH)2/In(OH)3 compound was conducted in the stainless steel autoclave. The as-prepared In(OH)3 was added to the saturated NiCl2 solution, followed by the 5.0 h hydrothermal reaction at 473.15 K. The laurel-green compound Ni(OH)2/In(OH)3 was filtered and washed with distilled water. Ultimately, Ni(OH)2/In(OH)3 composites were placed in crucible and heat-treated for 1.0 h at 873.15 K to yield yellow-green solid NiO/In2O3 products.

2.2. Characterization of the Samples

The p-n heterojunction-like NiO/In2O3 microparticles were characterized by scanning electron microscopy (SEM, Hitachi S-4800) and X-ray diffraction (XRD, Rigaku D/MAX-3C diffract meter). Fourier-transform infrared (FT-IR) spectra of materials were performed on a Bio-Rad FTS135 spectrometer using a KBr wafer technique.

2.3. Gas Sensor Fabrication and Response Test

Firstly, In2O3 and NiO/In2O3 slurries were formed by dispersing them in deionized water. Then, the as-prepared In2O3 and NiO/In2O3 slurries were directly drop-coated on the outer surface of alumina substrates with two Au electrodes. After drying for approximately 3.0 h at 60°C, the sensors were annealed subsequently at 600°C for another 3.0 h. To guarantee the long-term stability, aging test for 2 days was carried out to the sensors.

For the measurement system (WS-30A), the gas response of the sensor can be calculated by the equation: (where is resistance in gas and is resistance in air). The gas responses to formaldehyde (HCHO), benzene (C6H6), methylbenzene (C6H5CH3), trichloromethane (CHCl3), ethanol (C2H5OH), and ammonia (NH3) were measured. In addition, the gas concentrations were also studied. The substrate temperature was adjusted between 220°C and 320°C by controlling the heater powers.

3. Results and Discussion

3.1. Structure and Morphology

Figure 1 shows the SEM pictures of the (a) In(OH)3, (b) Ni(OH)2, (c) Ni(OH)2/In(OH)3, and (d) p-NiO/n-In2O3. As can be seen in Figure 1(a), the In(OH)3 precursors have an ordered cubic structure with size of ~10 μm, and the In(OH)3 microparticles also have a relative smooth surface. In Figure 1(b), the parallel samples of Ni(OH)2 nanoparticles possessed a uniform petal-shaped morphology with very good dispersibility. After decorating the cubic In(OH)3 precursors with Ni(OH)2 nanoparticles, it can be seen in Figure 1(c) that almost each of cubic In(OH)3 microparticles was surrounded by petal-shaped Ni(OH)2 nanoparticles, forming the precast complex structure. After the calcination of Ni(OH)2/In(OH)3 for 24.0 h, the p-n heterojunction-like NiO/In2O3 microparticles were prepared. As seen in Figure 1(d), calcining process did not influence seriously the morphology of Ni(OH)2/In(OH)3; therefore, the cubic In2O3 and petal-shaped NiO were observed. Compared to the smooth surface of original In(OH)3 precursors, the synthetic p-NiO/n-In2O3 have been altered to the rough surface structure (inset in Figure 1(d)). In structure, the gas sensing properties always strongly relied on their surface properties and shapes of the composites, and the p-NiO/n-In2O3 oxide semiconductor microparticles met the requirement of great surface-to-volume ratios. Moreover, in favor of the gas sensing characteristics, NiO maintained the excellent dispersibility, other than agglomerating together.

XRD spectra of the as-prepared (a) In(OH)3, (b) Ni(OH)2, (c) Ni(OH)2/In(OH)3, and (d) p-NiO/n-In2O3 composite particles are exhibited in Figure 2. In Figure 2(a), the In(OH)3 cubes were well crystallized, which could be corroborated by the JCPDS 16-161. Namely, all the peaks of Figure 2(a) could be indexed to a pure body centered cubic phase of In(OH)3. In Figure 2(b), almost all the reflections of the Ni(OH)2 could be readily indexed to a hexagonal phase (JCPDS number 14-0117), and no other characteristic peaks were detected for impurities. After the loading of Ni(OH)2, the precursor compounds Ni(OH)2/In(OH)3 demonstrated the XRD patterns both of Ni(OH)2 and In(OH)3, which in turn offered the evidence of the resultant compounds Ni(OH)2/In(OH)3. The resultant Ni(OH)2/In(OH)3 compounds contain the typical peaks of Ni(OH)2 and In(OH)3, implying the presence of both Ni(OH)2 and In(OH)3. The final products p-NiO/n-In2O3 are displayed in Figure 2(d). It should be noted that not only the patterns of NiO (JCPDS number 44-1159) but also the peaks of In2O3 (JCPDS number 06-0416) appeared, suggesting the successful synthesis of the final NiO/In2O3 products.

To confirm the interaction between p-NiO and n-In2O3 during the formation of gas sensors composites, FT-IR spectroscopy was employed to monitor the chemical bond transformation of parallel precursor and p-n heterojunction-like NiO/In2O3 materials. Figure 3(a) depicts the FT-IR spectrum of In(OH)3. The strong and broad bands at 3425 cm−1 (O-H stretching vibrations), 827 cm−1 (O-H bending vibration), 1161 cm−1, and 1107 cm−1 and 505 cm−1 (In-OH absorption bands) were the characteristic absorption peaks of the native In(OH)3. In Figure 3(b) of Ni(OH)2, the peak at 3642-3424 (O-H stretching vibrations), a broad band at 1637 cm−1 which belongs to the bending vibration of absorbed water, and the peak at 530 cm−1 (Ni-OH bending vibration) were presented [3840]. In the spectrum of Ni(OH)2/In(OH)3 in Figure 3(c), the bands at 530-505 cm−1 (Ni-OH) and 1161 cm−1 and 1107 cm−1 (In-OH) appeared simultaneously, implying that Ni(OH)2 has been attached to the surface of In(OH)3. After the thermal transition from Ni(OH)2/In(OH)3 to p-NiO/n-In2O3, p-NiO/n-In2O3 spectrum is shown in Figure 3(d). As we can see, the characteristic peaks at 482 cm−1 and 545, 488 cm−1 belong to the stretching vibrations of Ni-O and In-O-In, respectively. In combination with the analysis of above-mentioned SEM and XRD, we can confirm that NiO was firmly assembled onto the surface of the In2O3.

On the basis of SEM, XRD, and FT-IR, a possible formation mechanism of NiO/In2O3 microparticles was proposed: the In(OH)3 solid powder was put in the saturated NiCl2 solution in advance. Then, the OH- from In(OH)3 react with Ni2+ in the solution, generating the insoluble Ni(OH)2 and replacing segmental indium ion. During this process, the Oswald ripening in the initial formation of Ni(OH)2 occurred due to the rapid spread of nickel ion. As to a certain size of Ni(OH)2, the Oswald ripening was ceased because the atomic exchange of Ni+ cannot reduce the total interfacial energy. Therefore, Ni(OH)2 could germinate on the surface of In(OH)3, forming the Ni(OH)2/In(OH)3 composites. Along with the increase of reaction time, the loading of Ni(OH)2 increased on the surface of In(OH)3. Obviously, the rich surface hydroxyl of In(OH)3 has served as active site for the surface modification process through a condensation reaction with the hydroxyl of Ni(OH)2. In the meantime, the thus-treated particles were also coated by the SDBS molecule, which has an appropriate alkyl chains to stabilize these tiny nanocrystals and keep the particles from agglomerating. In another word, the spontaneous agglomerations in the reactor have been avoided by the use of SDBS molecule to reduce the surface activity of nanoparticles through the surface modification process. Thus, Ni(OH)2/In(OH)3 composite particles with a comparatively smaller diameter and a narrower size distribution could be generated. After calcination, the final product of NiO/In2O3 compound was obtained.

3.2. Gas Sensing Properties of the p-n Heterojunction-Like NiO/In2O3 Microparticles

The dynamic sensing transients of sensors (pure In2O3, NiO/In2O3 by hydrothermal reaction for 3.0 h and NiO/In2O3 by hydrothermal reaction for 24.0 h) to 600 ppm methanal gas were measured at 260°C (in Figure 4). As we can see in Figure 4, a pure In2O3 sensor displayed the lowest response. However, such low response could be slightly increased by NiO/In2O3 for hydrothermal reaction of 3.0 h. Further, the gas response value of the NiO/In2O3 sensors by hydrothermal reaction for 24.0 h was 85.55, which was almost four times more than pure In2O3. The gas response of the 24.0 h-NiO/In2O3 sensors was significantly higher than the other two sensors, implying that the gas response of In2O3 sensors was tightly dependent upon the configuration of the NiO component. Longer reaction time of hydrothermal reaction generated more NiO content. The enrichment of NiO nanoparticles on the surface of In2O3 displayed a better response.

There were several probable reasons which might be responsible for the preferable gas sensing performance of the p-n heterojunction-like NiO/In2O3. (1) The NiO/In2O3 with higher specific surface area than In2O3 substrate could afford more active sites to react with methanal gas molecules [41]. (2) The p-n heterojunction-like NiO/In2O3 structure was favorable for preventing the undesirable aggregation and ensured the stability of the sensors. (3) It is deemed to be a hetero-p-n-junction between NiO and In2O3 that increased the sensor resistance. Therefore, NiO/In2O3 samples showed superior methanal sensor performance.

For the NiO/In2O3 sensors, working temperature is also an important factor. Figure 5 presents the relationship between the working temperature and the sensor response. In the range of 220–320°C, the sensor response to methanol was enhanced with increasing working temperature and up to 85.5 at 260°C. Later, the sensor response of NiO/In2O3 was decreased at higher working temperature. The optimum working temperature of NiO/In2O3 sensor device was determined at 260°C. As is well known, the increase of temperature would facilitate the chemical reaction and then promote the gas response. But further increase in temperature leads to the decrease of response. The reason is put down to the low utilization rate of the sensing layer because the gas is consumed at the surface of the sensing layer, leading to reduction of the penetration depth of the target gas [42]. Lower temperature might not excite the adsorbed methanal and free electrons in NiO/In2O3 effectively, while, at higher temperature, methanal gas molecules were hard to be adsorbed, and oxygen molecules might be also desorbed, causing less influence on the resistance of the film [43]. In the same measurement system, the response of pure In2O3 to 600 ppm methanal was 28.5 at 280°C, which was much lower than that of the p-n heterojunction-like NiO/In2O3. These findings showed that the p-n heterojunction-like gas sensor had achieved more superior sensing properties than that of bare In2O3 materials.

The response and recovery time, which were also important parameters for a gas sensor, were generally defined as the times to reach 90% variation in resistance upon exposure to methanal and air were defined as the 90% response time and 90% recovery time, respectively. And the faster the gas diffusion speed is, the less the response/recovery time is needed [44, 45]. In Figure 6, for 600 ppm methanal gas, NiO/In2O3 sensors exhibited rapid gas sensing behavior when the target gases were injected or released. These results could be attributed to the nanogaps between NiO and In2O3. Benefiting from the nanogaps between NiO and In2O3, which could improve the gas diffusion speed, the p-n heterojunction-like NiO/In2O3 sensors perform higher rates of gas adsorption and desorption than solo In2O3. For the response time, the NiO/In2O3 sensors exhibited a shorter response time for all the heating temperatures. After injection of methanal gas, a higher concentration of the adsorbed oxygen species on the surface of the composite microparticle sensor reacted with methanal, resulting in a short response time. Even though the recovery time of the composite microparticle sensor was shorter than that of the pristine In2O3 sensor, no significant optimizing effect was observed for the composites. The recovery reaction consists of the following serial steps: (1) gas-phase diffusion of oxygen to the sensor material surface, (2) adsorption of the oxygen molecules at the surface, (3) dissociation of oxygen molecules into oxygen atoms, and (4) ionization of the atomic oxygen. The slower recovery time might be due to the sluggish surface reactions, including the adsorption, dissociation, and ionization of oxygen, compared with the fast gas-phase diffusion in the present sensor [46].

Figure 7 shows the sensor responses of the sensor device upon exposure to various concentrations of methanal at a working temperature of 260°C. The sensitivities of pristine In2O3 and NiO/In2O3 materials were studied. The values of NiO/In2O3 sensitivity were observed to range from 31.78 to 213.18, whereas the sensitivity of pure In2O3 ranged from 26.58 to 73.78. The higher sensitivity of NiO/In2O3 towards the methanal samples compared to the parallel In2O3 substrate can be attributed to the improved surface area. The higher surface area of NiO/In2O3, arising after the NiO treatment of In2O3 (previously observed in the SEM images in Figure 1), provided a larger surface for the sorption process. Thereby, the response to methanal of p-n heterojunction-like NiO/In2O3 was better than that of the as-prepared pure In2O3 samples. Moreover, at a low concentration of 600 ppm of methanal, the relative response of NiO/In2O3 is about 31.8. When increasing the concentration of methanal, the response of the NiO/In2O3 also sharply increased. For the NiO/In2O3, the response to 1200 ppm methanal is up to 213.2, representing an almost threefold difference of sensitivity when compared to the pure In2O3 at the concentration of 1200 ppm of methanal. Simply, better sensitivity was expected to higher methanal concentration, which might be a result of fast diffusion and reaction rate of methanol with high concentration [47].

In an attempt to investigate the selective detection behavior of NiO/In2O3 sensors, the selectivity of samples against other interference gases was also discussed by measuring the response to 600 ppm methanal, ethanol, ammonia, benzene, toluene, and chloroform at 260°C (Figure 8). It can be seen that the NiO/In2O3 exceeded In2O3 responses for methanol and ethanol only while other gases are comparable and almost negligible. Such phenomenon demonstrated that methanal could be sensitively and selectively detected by employing NiO/In2O3 sensor.

In theory, the negatively charged oxygen has inherent ability to oxidize the reducing gas, generating electrons which decrease the hole concentration in the surface layer through the electron-hole recombination [48]. Therefore, the resistance of gas sensor is added upon exposure to a reducing gas. In the current work, the chemical mechanisms of NiO/In2O3 sensors to methanal can be shown in the following reactions. Ion-sorbed oxygen behaved as the electron extractor from the bulk NiO/In2O3 sensors, and the methanal molecules were postulated to be oxidized by the ion-sorbed oxygen (where O2 was the oxygen in gas and O was ion-sorbed oxygen):

Therefore, the electron extracted by oxygen came back to the bulk, and the resistance of the NiO/In2O3 sensors changed to generate a sensor signal. The proposed process of NiO/In2O3 and reaction mechanism of NiO/In2O3 sensors to methanal gas is shown in Figure 9.

Note that when the background hole concentration was lower, the change of hole concentration had an increasing trend since the gas sensing reaction would result in a higher variation in sensor resistance. Similar results have been reported previously that both the gas response and sensor resistance of In2O3 were enhanced by doping with ZnO [49, 50]. Accordingly, in the case of p-n heterojunction-like NiO/In2O3 sensors, the enhancement of sensor resistance and gas response could be attributed to the lessened hole concentration due to NiO doping.

The prominent increasing of methanal gas response effectuated by decorating p-type NiO petals with n-type In2O3 could be explained by the formation of abundant p-n junctions. The hole accumulation layer for conduction would be narrowed by the hole depletion layer formation of the n-type In2O3 cubes, which offered an explanation for the superior resistance of the NiO/In2O3 sensors. It should be noted that and (the widths of the space charge layer of the n-oxide and p-oxide semiconductor sides, resp.) also relied on (relative ratio of the charge carrier concentration, and is donor density; stands for acceptor density). The equation (given that the value will go up as reduces) meant that NiO/In2O3 not only declined the hole concentration within the hole accumulation layer but also gave rise to the hole depletion layer to be thicker. In turn, the p-conducting channel became narrow. Hereby, a small variation in charge carrier concentration could bring about higher change in sensor resistance.

4. Conclusions

In summary, a highly sensitive and selective methanal sensor was designed via coating the petal-shaped p-type NiO onto the surface of n-In2O3 cubes. The as-synthesized product was characterized using SEM-EDS, XRD, and FT-IR. The experimental results of detecting the gas response verified that the gas sensing characteristic of NiO/In2O3 sensors was significantly different from the pure In2O3 cubes. The gas response value of NiO/In2O3 sensors was 85.55 to 600 ppm methanal gas at 260°C, and this value was almost four times more than pure In2O3 substrate. The prominent increasing of methanal gas response effectuated by decorating p-type NiO petals with n-type In2O3 could be explained by the formation of abundant p-n junctions. The gas responses of NiO/In2O3 sensors to ammonia, benzene, toluene, and chloroform were extremely weak at the similar circumstance. Thus, a superior response, quick recovery, and higher selectivity of gas response to 600 ppm methanal were achieved by decorating In2O3 cubic particles with petal-shaped NiO nanoclusters. Such dual-role nature of decorating the NiO microparticles onto In2O3 cubes therefore provides a novel approach to designing highly sensitive, selective, and rapidly recovering gas sensors based on n-type oxide semiconductors.

Competing Interests

The authors declare that they have no potential or actual conflict of interests pertaining to this submission.

Acknowledgments

This work was supported by Research Project of Foundation of Shaanxi Provincial Department of Education (no. 14JK1657) and the Fundamental Research Funds for the Central Universities (no. 310829161015 and no. 310829162014).