Bilayer Polyaniline-WO3 Thin-Film Sensors Sensitive to NO2.

In this work, a novel bilayer polyaniline-WO3 (PANI-WO3) thin film on the fluorine-doped tin oxide (FTO) glass substrate was prepared by hydrothermal synthesis and in situ chemical oxidative polymerization methods. Until now, no one has ever made attempts to use the PANI-WO3 composite on the FTO glass substrate to detect NO2 gas. The composite showed excellent sensing performance for NO2 detection at an operation temperature of 50 °C and a detection limit of 2 ppm. With regard to the PANI-WO3 hybrid, the response value for NO2 at 30 ppm is 60.19 and is three times higher than that for pure WO3 at 50 °C. Besides, the PANI-WO3 hybrid had excellent stability. The improvement of gas sensing was assigned to the creation of p-n heterojunctions between p-type PANI and n-type WO3, larger specific surface, increase of oxygen vacancies, and a wide conduction channel provided by PANI.

Introduction

With industrial development and rapid economic growth, more and more environmental issues, such as global warming, ozone depletion, and air pollution, gradually loom up.[1] Nitrogen dioxide (NO2), a reddish-brown poisonous gas with pungent smell, is recognized as one of the most hazardous airborne contaminants by the World Health Organization (WHO).[2−4] As is well known, NO2 generated from factories, automotive exhaust, and thermal power plants poses a significant threat to the environment and human health.[5−7] Especially, NO2 is the leading cause of acid rain and photochemical smog. Furthermore, people who are continuously exposed to low concentrations of NO2 can damage the respiratory system, increasing the risk of asthma, emphysema, and bronchitis, and even cause death.[8] Therefore, it is necessary to manufacture highly sensitive, selective, and stable gas sensors of NO2 for environmental protection and human health.

In recent years, n-type semiconducting metal oxides, for instance, tin oxide (SnO2),[9] zinc oxide (ZnO),[10] indium oxide (In2O3),[11] copper oxide (CuO),[12] titanium oxide (TiO2),[13] and tungsten oxide (WO3),[14] have occupied a pivotal place in the field of gas sensing because of low cost, high sensitivity, simplicity, and easy integration in electronics.[15] Among them, tungsten oxide (WO3) with a wide band gap of 2.6–2.8 eV is attracting more and more attention because of structural diversity, facile preparation, outstanding stability under extreme circumstances, and prominent gas sensing for NO2 and volatile organic compounds.[12,14,16,17] Hence, WO3 is a very prospective n-type semiconductor material for the detection of NO2.

It cannot be neglected that gas sensors based on WO3 have the drawbacks of high resistance, high operating temperature, and low selectivity. Instead, conducting polymers, for instance, polyacetylene, polypyrrole, polyaniline (PANI), and polydiacetylene, have the advantages of high conductivity, low-temperature operation, and low consumption.[18−20] Among them, PANI has been deeply studied and widely applied.[21] Thus, nanocomposites of WO3 and PANI can get better gas-sensing properties than single ones.[22−25]

Different preparation methods can result in different morphologies and microstructures. Some methods such as sol–gel[26] and hydrothermal[17] methods can regulate the morphology, while others such as thermal spray,[27] physical vapor deposition,[28] and chemical vapor deposition[29] can control the specific thickness of films. Compared with traditional ceramic electrodes, WO3 film electrodes have lower thickness and probably become thick film electrodes with low sensitivity and high resistance because of fewer influencing factors on the paste-coating process. Haining et al(30) directly had walnut-like In2O3 nanostructures grown on the interdigital electrode substrate via a hydrothermal approach. With the irradiation of UV light of 1.2 mW/cm2, In2O3 showed excellent NO2 gas-sensing performance, including high response (219–50 ppm NO2), superior selectivity, and great stability. Anyway, growing on the substrate requires a seed layer and the coating process was repeated five times, both were complicated. Shendage et al(31) fabricated nanostructured WO3 thin films without any seed layer on the soda lime glass substrate via a hydrothermal approach. The optimal sensing response was 10 for 5 ppm NO2, but an operating temperature of 100 °C was still relatively high. Zhang et al(27) fabricated Au-incorporated WO3 nanoporous thin films via layer-by-layer stacking of a sacrificial colloidal template. The response of this film was 96 for 1 ppm NO2 at an operating temperature of 150 °C. The layer-to-layer accumulation was so intricate that there were too many uncontrollable factors in the middle reaction process. Tae et al(32) prepared Au-incorporated WO3 thin films via a solution process and spin coating. This film showed a selectivity of 194.27 for 5 ppm NO2 at an operating temperature of 150 °C. The speed, time, uniformity, and viscosity of thin films prepared by spin coating are difficult to control precisely. To sum up, the abovementioned research studies mainly used multistep preparation of films or had best performance at high temperature. In order to improve these problems, we have taken the following action.

In this research, PANI–WO3 film was synthesized on the fluorine-doped tin oxide (FTO) glass substrate by a hydrothermal and in situ chemical oxidative method for the detection of NO2 at 50 °C. Until now, there has been no related report to use the combination of PANI, WO3, and FTO to reduce the resistance of WO3 itself and facilitate the transmission of electrical signals from the detection circuit. Gas-sensing properties of sensitivity, selectivity, stability, and response/recovery kinetics were observed. A comparative research between pure WO3 and PANI–WO3 composites was investigated to show the effect of PANI on gas sensing. Meanwhile, the gas-sensing mechanism is discussed in this article. We hope this research can provide a new idea to prepare WO3 film gas sensors with high sensitivity, selectivity, and stability.

Experimental Section

Synthesis of Composite Materials

Preparation of composite materials was divided into two parts: preparation of pure n class="Chemical">WO3 films and preparation of composites. The detailed preparation is in Scheme . All the chemicals from Aladdin were of analytical grade without further purification.

Scheme 1

Schematic Illustration of Preparing PANI–WO3 Composites

Synthesis of WO3 Films

The FTO glass substrates with length × width × thickness of 1.5 × 1 × 0.1 cm were purchased and washed with ethanol and deionized water. WO3 films were prepared by hydrothermal synthesis. Sodium tungstate dehydrate (Na2WO4, 0.18 g) was dispersed into 60 mL of deionized water with strong magnetic stirring for 10 min. Then, a certain amount of 3 M hydrochloric acid (HCl) was dropped into the preceding solution to adjust pH = 1. The acidified solution was continuously stirred vigorously for 20 min. After that, 0.20 g of ammonium oxalate [(NH4)2C2O4] was added into the solution with constant stirring for 10 min. The FTO glass substrate was washed with ethanol and deionized water in turn in an ultrasonic bath for 10 min. The cleaned substrate was put vertically in the autoclave containing the reaction solution. The sealed autoclave was kept at 160 °C for 6 h and then cooled down to room temperature in the oven. The white film was formed and washed with ethanol and deionized water several times to remove impurities. Finally, the film was sintered at 500 °C for 2 h. The yellow WO3 film was acquired.

Synthesis of PANI–WO3 Films

The PANI–WO3 composites were prepared through the in situ chemical oxidative polymerization route First, 110 μL of aniline and 1.5 mmol of ammonium persulfate were separately added into 15 mL of 1 M HCl solution and treated with ultrasonic wave for 10 min. Second, two solutions were placed in an ice bath for 15 min and then mixed with the abovementioned solution. Next, the previously prepared yellow WO3 films were put into the mixed solution under ice bath conditions for 5, 10, 15, 20, 30, and 40 min. Finally, the PANI–WO3 film was washed several times with deionized water and ethanol and dried at 60 °C. Bilayer PANI–WO3 films with different deposition times (5, 10, 15, 20, 30, and 40 min) were recorded as PW1, PW2, PW3, PW4, PW5, and PW6, respectively.

Characterization

Crystalline phases of the WO3 film were characterized by X-ray diffraction (XRD, SmartLab). The morphology of the WO3 film was investigated by field emission scanning electron microscopy (FESEM, JSM-7610F) equipped with energy dispersive X-ray spectrometry (EDS) and transmission emission microscopy (TEM, FEI Tecnai G2 F20). The band gap of WO3 was measured by UV–visible spectrophotometry (UV–vis, TU1901). The element oxidation state was evidenced by X-ray photoelectron spectroscopy (XPS, Nexsa X).

Gas-Sensing Tests

Gas sensing of thin films was measured using a JF02F gas-sensing test system from Kunming Guiyan jinfeng Technology Co., Ltd. This system adopted the dynamic gas distribution method under the control of a computer to test various related performance indicators of the sensors, which could eliminate the method error brought by the traditional static gas distribution. The sensing element was placed in the gas chamber onto the heating plate. The film was placed on the heating plate in the test chamber and accessed to test circuits through four copper probes. Introducing different volume ratios of NO2 and air from different gas paths to the chamber could achieve a specific concentration of the target gas. The response was defined as S = Rgas/Rair, where Rgas and Rair were the resistances of sensors under exposure to NO2 and air, respectively. The initial resistance of samples is shown in Table .

Table 1

Initial Resistance of Pure WO3, PW1, PW2, PW3, PW4, PW5, PW6, and PANI

sample	resistance (Ω)
WO3	98,250
PW1	38,449
PW2	28,429
PW3	22,628
PW4	17,183
PW5	9603
PW6	9020
PANI	3582

Results and Discussion

XRD Analysis

Figure exhibited the crystalline structure of pure WO3, PANI, and the PW3 composite. In Figure a, peaks appearing in pure WO3 can be well indexed to standard monoclinic WO3 with lattice constants a = 7.30084 Å, b = 7.53889 Å, and c = 7.68962 Å (JCPDS 83-0950). The main diffraction peaks of pure WO3 at 2θ = 23.1, 23.5, 24.3, 33.26, and 34.16° were attributed to the diffraction intensity of (002), (020), (200), (022), and (202) crystal faces of pure WO3. As for the XRD pattern of PW3 shown in Figure b, all peaks were similar to the peaks of pure WO3, which matched well with WO3 (JCPDS 83-0950), but the intensity of compounds was weaker, thanks to the existence of PANI. In Figure c, the XRD pattern of PANI showing a wide peak in the range of 17–33°was related to the periodic arrangement of the polymer chain.[33]

Figure 1

XRD pattern of (a) pure WO3, (b) PW3, and (c) PANI.

Morphological Analysis

The morphology and crystal structure were characterized by SEM and TEM techniques. Figure showed the FESEM images of pure WO3, PANI, and PW3. Figure a depicted that the uniform flower-like WO3 possessed an average diameter of about 4 μm, which was assembled by a number of small nanosheets in different directions. From the image of PANI in Figure b, it was apparent that PANI had a one-dimensional nanofiber structure with a diameter of roughly 100 nm. As is exhibited in Figure c, PANI nanofibers were randomly uniformly distributed on the surface of flower-like WO3, which revealed the composite structure of PW3. From Figure d–g, the EDS maps of PW3 further confirmed that PANI was grown on the surface of flower-like WO3. In addition, the distribution of W, O, and N were entirely consistent with the result observed in Figure . Also, the N element confirmed the existence of PANI.

Figure 2

FESEM images of (a) pure WO3, (b) PANI, and (c) the PW3 hybrid; (d–g) EDS elemental mapping images of W, O, and N for the PW3 hybrid.

TEM images in Figure further illustrated morphologies of the PW3 hybrid. According to Figure , the translucent PANI was covered on the surface of flower-like WO3, in accordance with the SEM result in Figure c–g. Figure b–d provided some information regarding high-resolution TEM (HRTEM) images of the PW3 hybrid. There was no distinct boundary between amorphous PANI and crystalline WO3, as shown in Figure b, which created the necessary conditions to form p–n junctions between PANI and WO3. Furthermore, the lattice spacings of 0.262, 0.365, and 0.384 nm matched well, respectively, with (202), (200), and (002) planes of monoclinic WO3 (JCPDS 83-0950).

Figure 3

(a) TEM image of the PW3 hybrid; (b–d) HRTEM images of the PW3 hybrid.

(a) TEM image of the n class="Chemical">PW3 hybrid; (b–d) HRTEM images of the n class="Chemical">PW3 hybrid.

XPS Analysis

The XPS surface analysis was carried out to further determine the chemical compositions. Figure indicates the existence of W and O in pure WO3 and the existence of W, O, N, and C in the PW3 hybrid, which further confirmed that no other impurity was imported in the prepared samples.

Figure 4

W 4f XPS spectrum of (a) pure WO3 and (b) the PW3 hybrid; O 1s XPS spectrum of (c) pure WO3 and (d) the PW3 hybrid; N 1s spectrum of the (e) PW3 hybrid; and C 1s spectrum of the (f) PW3 hybrid.

Both of the high-resolution spectra of W 4f in Figure a,b exhibited two characteristic peaks at binding energies of 35.6 and 37.6 eV with an energy difference of 2 eV, corresponding to spin–orbit splitting of W 4f7/2 and W 4F5/2 components, respectively, in n class="Chemical">WO3, which illustrated that the binding energy position conformed to the W6+ state.[34]

The O 1s spectra of two samples in Figure c,d were well fitted into three peaks (Oa, Ob, and Oc) by the Gaussian fitting method.[26] The Oa peaks of two samples at lower binding energy (530.5 eV) were the characteristic peak of WO3, which was assigned to the inherent oxygen in the WO3 matrix. The Ob peak in Figure c at a binding energy of 531.2 eV was attributed to the oxygen-deficient region because of oxygen vacancies, while the Ob peak in Figure d at a binding energy of 531.2 eV was attributed to the C=O band and oxygen-deficient region. The Oc peaks of two samples at higher binding energy (532.5 eV) were associated with chemisorbed surface oxygen. It was obvious that PW3 had more Ob and Oc species than pure WO3. However, it is well known that oxygen vacancies on the surface promote the adsorption of gas molecules. Thus, PW3 could adsorb more oxygen species on the surface, which was helpful to detect NO2.

The N 1s spectrum of PW3 in Figure e was decomposed into two peaks at 399.3 and 400.6 eV, which were respectively allocated to the benzenoid ring in PW3 and the radical cationic nitrogen atoms (=NH+=).[35] The N 1s spectrum clearly evidenced the existence of PANI. The C 1s spectrum in Figure f decomposed into three peaks at 284.8, 286.4, and 288 eV, which were respectively attributed to C–C, COOH, and C=O bands.[21] Hence, XPS surface analysis fully verified the existence of PANI and the formation of the PW3 hybrid.

Thermogravimetric Analysis

The thermogravimetric (TG) analysis of the PW3 hybrid was measured to investigate the mass ratio of PANI and the PW3 hybrid. From Figure , the TG curves of PANI and PW3 were further sorted into three different parts. The first part showed that both PANI and PW3 had a little mass loss before 100 °C because of evaporation of surface adsorbed water molecules. The main process was in the second part. As for PANI, the weight gradually decreased between 100 and 296 °C owing to the elimination of PANI acidified by hydrochloric acid and then fell to nearly 0 at 582 °C on account of the degradation of PANI. As for PW3, it had similar decrease in mass percentage. The weight of PW3 had a slow mass loss from 100 to 349 °C and eventually fell to 93% at about 467.5 °C. In the third part, both PANI and PW3 remained level with the increasing temperature. The abovementioned data confirmed that PANI accounted for 7% of the total mass of the PW3 composite.

Figure 5

TG curves of pure PANI and the PW3 hybrid.

TG curves of pure n class="Chemical">PANI and the n class="Chemical">PW3 hybrid.

Gas-Sensing Measurements

To find out the optimum mass ratio of PANI–WO3 composites, the research compared the responses of sensors based on different samples to 2–30 ppm at an operating temperature of 50 °C. The specific response values are shown in Table . As shown in Figure a, responses of sensors based on pure WO3, PW1, PW2, PW3, PW4, PW5, PW6, and PANI possessed a trend of “increase-maximum-decrease” to the same concentrations of NO2 with the increasing weight proportion of PANI. Besides, PW3 had a higher response to 10–30 ppm NO2 than pure WO3. It was distinct that the PW3 hybrid sensor showed the highest response of 60.19–30 ppm at 50 °C among all of the samples shown in Figure b, which was three times higher than that of pure WO3. The reason why responses of PW4, PW5, and PW6 were poorer than those of PW3 was mainly assigned to two aspects: on the one hand, adding an excess of PANI resulted in entirely covering the surface of WO3 and thus most of the NO2 gas was taken in PANI which had a bad effect on gas sensing. On the other hand, there were plenty of holes given by extra PANI which made the depletion area thicker at the p–n junction, which impeded the selectivity dramatically. Hence, we could draw a conclusion that an excess of PANI obstructed the response of PANI–WO3 to NO2.

Table 2

Responses of Sensors Based on Pure WO3, PW1, PW2, PW3, PW4, PW5, PW6, and PANI to 2–30 ppm NO2 at 50 °C

response	2 ppm	4 ppm	6 ppm	8 ppm	10 ppm	20 ppm	30 ppm
WO3	1.16	1.28	1.93	4.05	5.64	11.39	16.9
PW1	1.27	2.026	3	3.59	6.83	12.22	19.4
PW2	1.15	1.21	1.44	2.25	4.06	10.21	19.16
PW3	1.18	1.23	1.45	1.66	2.39	14.47	60.9
PW4	1.123	1.14	1.31	1.62	1.983	2.734	6.42
PW5	1.11	1.12	1.21	1.36	1.71	2	2.50
PW6	1.03	1.04	1.08	1.18	1.58	1.15	1.19
PANI	1	1	1	1	1	1	1

Figure 6

(a) Responses of sensors based on pure WO3, PW1, PW2, PW3, PW4, PW5, PW6, and PANI to 2–30 ppm NO2 at 50 °C; (b) Responses of sensors based on pure WO3, PW1, PW2, PW3, PW4, PW5, PW6, and PANI to 30 ppm NO2 at 50 °C.

Figure shows typical dynamic sensing response curves of pure WO3, PW3, and PANI with 2–10 ppm NO2 at an operating temperature of 50 °C. The lines in Figure a led us to the conclusion that the responses of sensors based on pure WO3 and PW3 increased with increasing NO2 concentrations from 2 to 10 ppm at an operating temperature of 50 °C, whereas the sensor based on PANI had no sensitivity to NO2 gas. According to Figure b, the sensor based on PW3 could recover to the original response value after exposure to 2–10 ppm NO2. In contrast, pure WO3 almost possessed no recovery to the original response after exposure to 2–6 ppm and possessed recovery to half of the original value after exposure to 8–10 ppm. Hence, PW3 had a better recovery property than pure WO3. In addition, Table demonstrates a comparison of the sensing performances of the sensors in this work and in literature studies.[5,27,30−32] As shown in the table, the sensors based on PANI–WO3 hybrids for NO2 have lower working temperature than those reported in literature studies. Therefore, they are the promising sensing materials for the detection of NO2.

Figure 7

(a) Responses of sensors based on pure WO3, PW3, and PANI to 2–10 ppm NO2 at 50 °C; (b) responses of sensors based on pure PW3 to 2–10 ppm NO2 at 50 °C.

Table 3

Responses of PANI–WO3 to NO2 at 50 °C in the Present Work and Those Reported in the Literature Studies

sensor material	NO2 concentration (ppm)	operating temperature	response	reference
Au@WO3	1	100 °C	20	(5)
WO3 films	1	150 °C	96	(27)
In2O3 films	50	UV light of 1.2 mW/cm2	219	(30)
WO3 films	5	100 °C	10	(32)
Au–WO3	5	150 °C	194.27	(32)
PANI–WO3	2	50 °C	1.18	this work

Additionally, the stability of the sensor based on the PW3 hybrid was measured to 30 ppm NO2 over the period of 35 days at 50 °C. The response of PW3 to 30 ppm NO2 was tested once a week in the span of 21 days and then finally tested after 14 days, as shown in Figure . In the third week, there was an obvious decline probably because of sensing material aging. After that, the response of PW3 remained stable. Therefore, the stability of the sensor based on the PW3 hybrid was great on average.

Figure 8

Sensing stability of sensors based on PW3 to 30 ppm NO2 at 50 °C.

Sensing stability of sensors based on n class="Chemical">PW3 to 30 ppm n class="Chemical">NO2 at 50 °C.

Sensing tests were also performed to evaluate the selectivity of the sensor toward NO2, NH3, SO2, ethanol, and acetone gases. As shown in Figure , pure WO3 and the PW3 hybrid showed a response of 17–30 ppm and a response of 60.19–30 ppm, respectively. Furthermore, both showed no significant response to NH3, SO2, ethanol, and acetone gases. It clearly demonstrated that PW3 exhibited the highest response toward 30 ppm compared to other test gases. Consequently, the sensor based on the PW3 hybrid had great selectivity.

Figure 9

Selectivity study of sensors based on pure WO3 and PW3 to 30 ppm different gases at 50 °C.

What is more, it was acknowledged that humidity was an important factor of gas sensing, which strongly influenced the performance of the gas sensor. As could be seen from Figure , the sensor based on the PW3 hybrid was investigated under different relative humidities (20–80% RH). The bar chart of response presented a fluctuation with an increase of relative humidity. The response showed a slight decrease in the range from 20 to 60%, while the response exhibited an evident decline. Therefore, the optimum operating humidity was in the range from 20 to 60%.

Figure 10

Responses of sensors based on PW3 under different humidities to 30 ppm NO2 at 50 °C.

Responses of sensors based on n class="Chemical">PW3 under different humidities to 30 ppm n class="Chemical">NO2 at 50 °C.

Gas-Sensing Mechanism

It is generally acknowledged that the sensing mechanism of n-type WO3 is related to the adsorption and desorption of gas molecules on the sensor surface, which causes changes in resistance.[36] When the sensor was based on pure WO3 in air, the oxygen molecules were adsorbed onto the surface of WO3 and then reacted with carries (electrons). Therefore, there a resistance of the sensor is formed in air. In addition, oxygen species vary with temperature, such as O2– (below 100 °C), O– (100–300 °C), and O2– (above 300 °C).[34] When the sensor based on pure WO3 is exposed to NO2, the NO2 molecules not only capture carries (electrons) from WO3 but also react with oxygen species on the surface of WO3. The abovementioned process causes a decrease of electronic concentration on the surface of WO3 and an increase of width of depletion regions and thus results in an increase in resistance of the sensor.[26]

As Figures and 7 show, the response of the sensor based on the PW3 hybrid has a distinguished improvement in gas-sensing properties compared to that of pure WO3, which can be assigned to four reasons. First, the structure of the PW3 hybrid obtained by growing PANI on the surface of flower-like WO3, according to the SEM image in Figure c, had a larger specific area than pure WO3,[37] which is favorable to assimilate more NO2 gas molecules on the surface of the PW3 hybrid to improve the sensitivity. Second, the acidified PANI, one of the conducting polymers, has a wide conduction channel which leads to a lower resistance in the PW3 hybrid than in pure WO3. In other words, it can enlarge the range of response values. Third, from XPS analysis, it is obvious that PW3 had more oxygen vacancies than pure WO3, which promotes the adsorption of gas molecules.[26] Finally, the most important reason is that the formation of the p–n junction at the hetero-interface between p-type PANI and n-type WO3 occurs, as shown in Figure . The acidified PANI provides some free holes, while the flower-like WO3 offers a number of free electrons. When nanofiber PANI grows on the surface of flower-like WO3, electrons and holes will diffuse in the opposite direction owing to the different Fermi energies. In the end, the p–n junction reaches equilibrium and a narrow depletion area is formed at the hetero-interface. When the PW3 hybrid is exposed to NO2, NO2 will break the old equilibrium and create new balance, resulting in enhancing the width of the depletion layer at the hetero-interface. Consequently, the resistance of the PW3 hybrid is larger than that of pure WO3 and thus improves gas-sensing performance of the PANI–WO3 hybrid.

Figure 11

(a) Schematic diagram of PANI–WO3 composites; (b) schematic illustration of the sensing mechanism of PANI–WO3 composites.

Conclusions

Bilayer PANI–WO3 thin-film sensors have been successfully synthesized on the FTO glass substrate via hydrothermal synthesis and in situ chemical oxidative polymerization methods. Our results reveal that the sensor has ultrasensitivity to NO2 gas at an operating temperature of 50 °C. The limit of concentration is 2 ppm. In addition, the sensor, whose structure was obtained by growing nanofiber PANI on the surface of flower-like WO3, showed good stability and selectivity. The improvement of gas sensing was assigned to the creation of p–n heterojunctions between p-type PANI and n-type WO3, larger specific surface, increase of oxygen vacancies, and a wide conduction channel provided by PANI. Our study gives a new idea to prepare a low-resistance WO3-based gas sensor for NO2 detection and other related application.