Hierarchical Fe₂O₃@WO₃ nanostructures with ultrahigh specific surface areas: microwave-assisted synthesis and enhanced H₂S-sensing performance

Abstract

Hierarchical Fe₂O₃@WO₃ nanocomposites with ultrahigh specific areas, consisting of Fe₂O₃ nanoparticles (NPs) and single-crystal WO₃ nanoplates, were synthesized via a microwave-heating (MH) in situ growth process. WO₃ nanoplates were derived by an intercalation and topochemical-conversion route, and the Fe₂O₃ NPs were in situ grown on the WO₃ surfaces via a heterogamous nucleation. The water-bath-heating (WH) process was also developed to synthesize a Fe₂O₃@WO₃ nanocomposite for comparison purposes. The techniques of X-ray diffraction (XRD), X-ray photoelectron spectrum (XPS), scanning electron microscopy (SEM) and transmission electron microscopy (TEM) were used to characterize the samples obtained. The results show that α-Fe₂O₃ NPs with a size range of 5–10 nm are uniformly, tightly anchored on the surfaces of WO₃ nanoplates in the Fe₂O₃@WO₃ samples obtained via the MH process, whereas the α-Fe₂O₃ NPs are not uniform in particle-sizes and spatial distribution in the Fe₂O₃@WO₃ samples obtained via the WH process. The BET surface area of the 5wt%Fe₂O₃@WO₃ sample derived by the MH process is as high as 1207 m² g⁻¹, 5.9 times higher than that (203 m² g⁻¹) of the corresponding WO₃ nanoplates. The dramatic enhancement in the specific surface area of the Fe₂O₃@WO₃ samples should be attributed to the hierarchical microstructure, which makes the internal surfaces or interfaces in aggregated polycrystals be fully outside surfaces via a house-of-cards configuration, where the single-layered and disconnected Fe₂O₃ NPs are tightly anchored on the surfaces of the WO₃ nanoplates. The gas-sensing properties of the Fe₂O₃@WO₃ sensors were investigated. The gas-sensors based on the Fe₂O₃@WO₃ obtained via the MH process show a high response and selectivity to H₂S at low operating temperatures. The 5%Fe₂O₃@WO₃ sample shows the highest H₂S-sensing response at 150 °C. Its response to 10 ppm H₂S is as high as 192, 4 times higher than that of the WO₃-nanoplate sensor. The improvement in the gas-sensing performance of the Fe₂O₃@WO₃ nanocomposites can be attributed to the synergistic effect in compositions and the hierarchical microstructures with ultrahigh specific surface areas.

---

1. Introduction

Metal oxide semiconductors (MOSs) have been extensively investigated for gas-sensing applications due to their unique electrical and optical properties.^1–3 The gas-sensing performance of a MOS sensor relies upon the interaction of the target gases with adsorbed oxygen on the oxide surface according to the related “electron depletion layer” models.⁴ Yamazoe and Shimanoe^5,6 further formulated theoretical equations to account for the gas-sensing response and proposed a new type of “volume depletion”, which indicated that decreasing the crystallite size and increasing the adsorbed oxygen can increase “volume depletion” and then improve gas performance. Thus, developing MOS nanocrystals with small particle sizes and higher specific surface areas is the frontier in gas-sensing material investigation. Recently, hierarchical functional nanocomposites with a tunable dimension and structure have become a research hotspot for its potential high surface area with promising applications in high-performance gas-sensors.^2,7,8 Particularly, the hierarchical binary oxide nanocomposites can provide high sensitivities and fast response due to the formation of electric junctions at the interface of the heterostructures and a more extended depletion layer.^9,10

Tungsten trioxide (WO₃), an n-type semiconductor with band gaps of ∼2.7 eV, has been extensively studied as gas-sensing materials because of its high response to various gases.^11–16 Several WO₃-based binary oxide composites such as WO₃/ZnO,¹⁷ NiO/WO₃,¹⁸ CuO/WO₃ (ref. 19) and SnO₂/WO₃,²⁰ have been reported as high-performance gas sensors. The plate-like heterogeneous NiO/WO₃ nanocomposites were synthesized by annealing Ni(OH)₂ and H₂WO₄ nanoplates in air and their gas-sensing sensitivity towards NO₂ were enhanced due to their p–n heterogeneous characteristics.¹⁸ Kida et al.²⁰ introduced SnO₂ nanoparticles into WO₃ lamella-based films obtained by mixing two suspensions containing WO₃·nH₂O and SnO₂ nanoparticles, and the enhanced gas-sensing response is owe to the porosity improvement. But unfortunately, the control in microstructure and distribution of the second phase is always extremely difficult by the RF sputtering technique¹⁹ or simply mixing their precursors.^18,20 As a result, the advantages of potential high surface areas cannot be fully embodied in practical applications. Simple and efficient methods are urgent for the construction of WO₃-based binary oxide heterostructures. Recently, microwave-assisted techniques have been used to synthesize inorganic materials for the potentials in controlling microstructures and morphology due to the selective absorption of microwave energy.²¹

Hematite (α-Fe₂O₃), with a band gap of 2.0–2.2 eV, is suitable to implement a host/guest n/n junction architecture with WO₃.²² The scaffold host material of WO₃ has a larger band gap than α-Fe₂O₃, and the conduction band of WO₃ is lower in energy than that of α-Fe₂O₃, allowing efficient electron transport across the host/guest interface. Fe₂O₃/WO₃ nanocomposites have been fabricated and used as nano-electrodes,²³ water splitting^22,24 and visible-light-driven photocatalysis.²⁵ The α-Fe₂O₃ nanocrystals has also been investigated as a promising gas-sensing material especially for H₂S detection.^26–31 Thus, the combination of WO₃ and α-Fe₂O₃ is expected to be an efficient gas-sensing system for H₂S-detection. However, to our best knowledge, the investigation on α-Fe₂O₃/WO₃ nanocomposites for gas-sensing application has not been reported to date.

In this paper, we develop a simple microwave-assisted process to fabricate uniform-distributed α-Fe₂O₃ nanoparticles (NPs) on WO₃ nanoplates, forming Fe₂O₃@WO₃ nanocomposites with ultrahigh surface areas that are highly sensitive to H₂S gas. The two-dimensional WO₃ nanoplates are synthesized via a robust intercalation and topochemical conversion route.³² Furthermore, the as-obtained WO₃ nanoplates have a high diameter-to-thickness ratio and single-crystalline structure, which is suitable as an efficient substrate to construct hierarchical nanocomposites.³³ The Fe₂O₃ NPs are in situ formed on the surfaces of WO₃ nanoplates via the microwave heating followed by calcination. Using the above WO₃ nanoplates as the substrate, the α-Fe₂O₃ NPs are uniform anchored on their surfaces to form hierarchical Fe₂O₃@WO₃ nanostructures. The hierarchical Fe₂O₃@WO₃ nanostructures not only prevent the aggregation of the α-Fe₂O₃ NPs, but also provide efficient diffusion paths and adsorption sites for gas molecules because of their ultrahigh surface areas.

The simple microwave heating process developed here offers several advantages: fast synthesis and no high pressure. The WO₃ nanoplate is kept stable during the fast synthesis by avoiding dissolution in alkaline conditions. In addition, the microwave process may promote the heterogeneous nucleation of Fe(OH)₃ NPs on the surfaces of WO₃ nanoplates due to selective absorption of microwave energy, leading to an uniformly distributed α-Fe₂O₃ NPs. The gas-sensing properties of the as-obtained Fe₂O₃@WO₃ nanomaterials are comparatively evaluated with an emphasis on H₂S-sensing detection. The effects of α-Fe₂O₃ amounts, operating temperatures on H₂S-sensing properties, and the related mechanisms are carefully investigated.

2. Experimental section

2.1 Synthesis of WO₃ nanoplates

WO₃ nanoplates were synthesized according to the literature with some modification.^32,34 Typically, tungstic acid (5.8 g, chemically pure) reacted with n-octylamine (50 ml, analytically pure) in heptane (350 mL, analytically pure) at room temperature for 72 h, forming tungstate-based inorganic–organic hybrid belts. The as-obtained hybrids were treated in a ∼38% HNO₃ aqueous solution for 48 h to remove organic species, forming H₂WO₄ nanoplates. WO₃ nanoplates were finally obtained by calcining the above H₂WO₄ nanoplates at 400 °C for 2 h, and used as the supports for immobilizing α-Fe₂O₃ nanocrystals to construct hierarchical nanostructures.

2.2 Synthesis of hierarchical Fe₂O₃@WO₃ nanostructures

Fe₂O₃@WO₃ nanostructures were synthesized via two methods: the microwave heating (MH) process and the water-bath heating (WH) process. For the MH process, WO₃ nanoplates (0.1 g), Fe(NO₃)₃·9H₂O (0.025 g, analytically pure) and urea (0.01 g, analytically pure) were mixed in 50 mL distilled water in a conical flask with an ultrasonic treating for ∼2 h, and the as-obtained mixture was then placed in a microwave oven (500 W, 2.45 GHz) and heated for 4 min, followed by cooling treating in an ice bath. The brick-red solid particles were collected, washed and then dried in a vacuum oven at 60 °C for more than 12 h. Finally, the Fe₂O₃@WO₃ sample with a theoretical α-Fe₂O₃ content of 5 mass%, i.e., 5%Fe₂O₃@WO₃, was obtained by calcining the above precipitates in air at 500 °C for 3 h. The Fe₂O₃@WO₃ samples with various α-Fe₂O₃ contents, i.e., 2.5% Fe₂O₃@WO₃ and 9% Fe₂O₃@WO₃, were synthesized using the similar process, just changing the amounts of Fe(NO₃)₃ and urea. The molar ratio of urea to Fe(NO₃)₃ was kept ∼2.6 for all of the samples. For the WH process, the mixture containing 0.1 g of WO₃ nanoplates, 0.025 g of Fe(NO₃)₃·9H₂O and 0.01 g of urea was treated in a water bath at 85 °C for 3 h under a stirring condition. The precipitates, obtained by centrifugation, washing with distilled water and ethanol, and drying in a vacuum oven at 60 °C for more than 12 h, were finally calcined at 500 °C for 3 h. The 5%Fe₂O₃@WO₃ nanocomposite was obtained through the water-bath heating process.

2.3 Synthesis of Fe₂O₃ nanocrystals

Pure Fe₂O₃ nanocrystals were synthesized via a hydrothermal process. Typically, Fe(NO₃)₃·9H₂O (0.25 g) and urea (0.1 g) were dissolved in 50 mL water and stirring for 30 min to form a transparent solution. After treated in the microwave oven for 4 min, the above mixture was transferred into a 100 ml Teflon lined stainless autoclave and maintained at 150 °C for 4 h. The resulting solid was collected, washed and dried at 60 °C for ∼12 h. Finally, the α-Fe₂O₃ nanocrystals were obtained by calcining the above solid in air at 500 °C for 3 h.

2.4 Materials characterization

The phases of the samples were confirmed via X-ray diffraction (XRD) analysis by using a Rigaku D/Max-3B X-ray diffractometer with Cu Kα radiation (XRD, λ = 0.15406 nm). The morphology and microstructures were characterized on a scanning electron microscope (SEM, JEOL JSM-5600, Japan) with an acceleration voltage of 15 kV, and on a transmission electron microscope (TEM, FEI Tecnai-G2, USA) with an acceleration voltage of 200 kV. For SEM observation, the samples were coated with a thin Pt film. The Brunauer–Emmett–Teller (BET) surface areas were measured using a Quantachrome Nova2000 sorption analyzer. The X-ray photoelectron spectrum (XPS) spectra of the Fe₂O₃@WO₃ sample were recorded on a multipurpose X-ray photoelectron spectroscope (Kratos Amicus, Manchester, UK) with a microfocused monochromatic X-ray source of Al Kα, using adventitious carbon (C1s 284.8 eV) as the calibration reference.

2.5 Gas-sensing measurement

The details for the sensor fabrication and the gas-sensing testing process were similar to our previous report.³⁵ The sensors were fabricated using a brush-coating method. Simply, the samples consisting of Fe₂O₃@WO₃ nanostructures, WO₃ nanoplates or α-Fe₂O₃ nanocrystals were mixed with a small amount of water to form the corresponding pastes, which were then coated onto the surfaces of alumina microtubes to form continuous thin films. The as-obtained thin films were heated at 300 °C for 5 h in air before gas-sensing test. The gas-sensing properties of the as-obtained sensors were evaluated using a commercial WS-30A system equipped with a computer terminal. The testing system was placed in a ventilating cabinet with a large draught capacity. The target substances for gas-sensing evaluation were injected into the closed chamber using a micro-injector. The harmful gases (e.g., H₂S, CH₄, CO, H₂ and SO₂) and some common organic vapors (e.g., methanol, ethanol, isopropanol, methanol, acetone and benzene) were chosen as the target substances to evaluate the gas-sensing performance of the Fe₂O₃@WO₃ sensors, with an emphasis on H₂S-sensing testing. The H₂S detection were operated at 100–250 °C with a H₂S concentration [H₂S] range of 0.5–10 ppm, whereas the other target substances with a concentration of 100 ppm were used the sensing gases (or vapors) operating at 150 °C.

A sensor (R) was connected with a standard resistor (R₀) in series, and the total voltage (U₀) applied on the sensor and standard resistor was fixed to 5 V. The WS-30A system measured the voltage (U) loaded on the standard resistor (R₀), and the resistance (R) of the Fe₂O₃–WO₃ sensor were calculated according to R = (5 − U) × R₀/U.¹² The sensor response (S) of a sensor based on an n-type semiconductor was defined as S = R_a/R_g for reducing gases and S = R_g/R_a for oxidizing gases, where R_a and R_g were the resistances of the sensor in air and in target gas, respectively. The response (or recovery) time was defined as the time in which the sensor reached 90% of the saturated signal upon exposure to (or removing) the target gas.^33,35

3. Results and discussion

3.1 Synthesis and characterization

Hierarchical Fe₂O₃@WO₃ nanocomposites are synthesized via a microwave heating process (Scheme 1). Firstly, Fe³⁺ ions are absorbed on the surfaces of WO₃ nanoplates because of the electrostatic interaction between the Fe³⁺ ions and the hydroxyls of WO₃ nanoplates. Urea molecules, CO(NH₂)₂, are possibly enriched around the surfaces of WO₃ nanoplates due to the combination of Fe³⁺ ions with –NH₂ groups in urea molecules. It is known that the quick heat and non-thermal effects of microwave radiation. When microwave heating is conducted, many hot spots generated rapidly on the surface of WO₃ nanoplates due to their selective absorption of microwave energy. Urea molecules around the WO₃ nanoplates are hydrolyzed to generate a large amount of OH⁻ ions as the temperature increases to ∼80 °C in a short time (i.e., 4 min). As a result, Fe(OH)₃ clusters are predominately formed on the surfaces of WO₃ nanoplates via a heterogeneous nucleation mechanism, and then grow up rapidly during the microwave heating. The Fe(OH)₃ nanocrystals are in situ formed and tightly immobilized on the surfaces of WO₃ nanoplates, and the hierarchical Fe(OH)₃@WO₃ nanostructures can be expected. Hierarchical Fe₂O₃@WO₃ nanocomposites are then obtained by calcining the Fe(OH)₃@WO₃ at 500 °C for 3 h in air. On the basis of the above analysis of the reaction mixture, the possible reaction can be described as eqn (1)–(3).

CO(NH₂)₂ + 3H₂O → 2NH₄⁺ + 2OH⁻ + CO₂↑ (1)

Fe³⁺@plate-WO₃ + 3OH⁻ → Fe(OH)₃@plate-WO₃ (2)

2Fe(OH)₃@plate-WO₃ → Fe₂O₃@plate-WO₃ + 3H₂O (3)

Scheme 1 Schematic illustration of the microwave-heating (MH) synthesis of the hierarchical Fe₂O₃@WO₃ nanocomposites with ultrahigh specific surface areas.

Fig. 1 shows the typical XRD patterns of WO₃ nanoplates, α-Fe₂O₃ nanocrystals and 5%Fe₂O₃@WO₃ nanocomposite. Fig. 1a shows a XRD pattern of the pure Fe₂O₃ sample. The major peaks can be readily indexed to the (012), (104), (110), (113), (024), (116) and (214) reflections of the rhombohedral α-Fe₂O₃ phase (JCPDS card no. 33-0664). Fig. 1c shows a typical XRD pattern of the WO₃ nanoplates, and it can be readily indexed to a triclinic WO₃ phase according to the literature (JCPDS card no. 32-1395). Fig. 1b shows the XRD pattern of the 5%Fe₂O₃@WO₃ nanocomposite. The major peaks of the Fe₂O₃@WO₃ sample are similar to the those of WO₃ nanoplates, and the weak peak at around 33.1° in 2θ can correspond to the (104) plane of α-Fe₂O₃ phase. The weakened peaks of α-Fe₂O₃ phase in Fig. 1b are mainly because of the lower content of α-Fe₂O₃ in the Fe₂O₃@WO₃ nanocomposite.

Fig. 1 Typical XRD patterns of (a) α-Fe₂O₃ nanoparticles obtained by the hydrothermal process, (b) 5%Fe₂O₃@WO₃ nanocomposite obtained by the MH process, and (c) WO₃ nanoplates.

The morphologies and microstructures of WO₃ nanoplates, Fe₂O₃ nanocrystals and 5%Fe₂O₃@WO₃ nanocomposite were characterized using SEM and TEM techniques. Fig. 2 shows the typical SEM and TEM images of the samples. Fig. 2a gives the SEM image of WO₃ nanoplates with dimensions of 100–700 nm in lateral size and 10–30 nm in thickness, which is similar to our precious report.³² It needs to be noted that the surfaces of the pure WO₃ nanoplates are smooth, and the small particles are sparse. Fig. 2b shows the SEM image of α-Fe₂O₃ sample, and the pure α-Fe₂O₃ sample consists of particulate nanocrystals with a mean size of about 30 nm.

Fig. 2 (a–c) Typical SEM images of (a) WO₃ nanoplates, (b) α-Fe₂O₃ nanoparticles and (c) 5%Fe₂O₃@WO₃ nanocomposites via the microwave-heating process; (d) EDS spectrum of the 5%Fe₂O₃@WO₃ nanocomposite (the inset is the corresponding SEM image); (e and f) TEM observation of the 5%Fe₂O₃@WO₃ nanocomposite: (e) low-magnification TEM image and (f) high-resolution TEM image (the inset is the corresponding FFT pattern).

Fig. 2c shows a typical SEM image of the 5%Fe₂O₃@WO₃ sample. One can see that the α-Fe₂O₃ NPs with a mean size of about 9 nm are uniformly immobilized on the surfaces of WO₃ nanoplates. Fig. 2d presents the energy dispersive X-ray spectroscopy (EDS) spectrum of the 5%Fe₂O₃@WO₃ nanocomposite, and the elements of W, Fe and O are detected in the sample. Fig. 2e show the typical low-magnification TEM image of the 5%Fe₂O₃@WO₃ nanostructure. One can further confirmed that the α-Fe₂O₃ NPs with a size range of 5–10 nm are uniformly and tightly immobilized on the surfaces of WO₃ nanoplates, similar to the SEM observation. Fig. 2f shows the high resolution TEM (HRTEM) image. The lattice fringes with a crystalline interplanar spacing of 0.376 nm can be attributed to the (020) plane of the WO₃ phase, and the lattice fringes with a crystalline interplanar spacing of 0.27 nm should be attributed to the (104) planes of the α-Fe₂O₃ phase. The corresponding two-dimensional fast Fourier transforms (FFT) of the lattice image given in the inset of Fig. 2f corroborate the co-existence of WO₃ and α-Fe₂O₃ phases. It should be emphasized that the α-Fe₂O₃ NPs are tightly anchored on the surfaces and edges of WO₃ nanoplates, and no discrete α-Fe₂O₃ NPs are found in the large TEM observation field. From the SEM and TEM images, one sees that the α-Fe₂O₃ NPs with very small sizes are not aggregated because of the support effect of WO₃ nanoplates, suggesting that the hierarchical nanostructures derived from zero-dimensional nanoparticle and two-dimensional nanoplates can be an ideal configuration to prevent the small nanoparticles from aggregation.

The microstructures of the Fe₂O₃@WO₃ composites were also characterized using the Brunauer–Emmett–Teller (BET) nitrogen sorption-desorption measurement. Fig. 3 shows the typical N₂ adsorption–desorption isotherms and the BJH pore-size distribution curves of the 5%Fe₂O₃@WO₃ composite and the support of WO₃ nanoplates. One can find both the WO₃ nanoplates (Fig. 3a) and the 5%Fe₂O₃@WO₃ composite (Fig. 3b) show a similar type II isotherm with no obvious hysteresis, and there are no obvious peaks in the pore-size distribution curves (insets in Fig. 3a and b). According to the above BET results, the WO₃ nanoplates and the 5%Fe₂O₃@WO₃ composite are of a non-porous structure.³⁶ The BET specific surface area of the WO₃ nanoplates is 203 m² g⁻¹, similar to our previous report.³² The introduction of a small amount of α-Fe₂O₃ NPs on WO₃ nanoplates using the MH process highly enhances the BET specific surface areas. For example, the 5%Fe₂O₃@WO₃ composite has a BET specific surface area as high as 1207 m² g⁻¹, which is 5.9 times higher than that of the corresponding WO₃ nanoplates. To the best of our knowledge, it is possibly the highest value in BET surface areas for a non-porous metal oxide material. Taking the SEM and TEM observations into account, the dramatic enhancement in the specific surface area of the Fe₂O₃@WO₃ samples should be attributed to the hierarchical microstructure, which makes the internal surfaces or interfaces in aggregated particles fully be outside surfaces via a house-of-cards configuration, where the single-layered and disconnected α-Fe₂O₃ NPs (several nanometers in size) are tightly anchored on the surfaces of the ultrathin WO₃ nanoplates.

Fig. 3 Nitrogen (N₂) adsorption–desorption isotherms of (a) WO₃ nanoplates and (b) 5%Fe₂O₃@WO₃ nanocomposites obtained by the microwave-heating process. The insets are their corresponding BJH pore-size distribution curves.

Elemental compositions and chemical states of the hierarchical Fe₂O₃@WO₃ nanocomposites were analyzed using the XPS technique. Fig. 4 shows the typical XPS spectra of the 5%Fe₂O₃@WO₃ nanocomposite. A survey spectrum shown in Fig. 4a confirms that there are 4 elements of W, Fe, O and C detected. The elements of W, Fe and O should belong to the Fe₂O₃@WO₃ sample, and the C species is due to the carbon tape used to attach the sample powders during the XPS measurement. Fig. 4b shows the W 4f spectrum, which has two peaks at 35.4 eV and 37.5 eV, corresponding to the W4f_7/2 and W4f_5/2 of the crystalline WO₃ species, respectively.²⁸ Fig. 4c shows the Fe 2p spectrum, and there are two peaks at 711.2 and 725.3 eV, belonging to Fe 2p_3/2 and Fe 2p_1/2 of α-Fe₂O₃, respectively.³⁷ The O 1s spectrum in Fig. 4d shows a wide peak with a large hump on the high-energy side, and the wide peak can be fitted to be three sub-peaks at 530.4, 532.2 and 533.4 eV. The intense peak at 530.4 eV can be assigned to the lattice oxygen in the crystalline WO₃ and α-Fe₂O₃ species. The peak at 532.2 eV may be due to the oxygen ions (O₂⁻, O⁻and O²⁻) adsorbed on the surfaces of the hierarchical Fe₂O₃@WO₃ nanostructure, whereas the peak at 533.4 eV is usually ascribed to the adsorbed H₂O molecules.^38,39 The intense intensities of the O 1 s XPS peaks at 532.2 and 533.4 eV suggest that there is a large amount of adsorbed oxygen and H₂O molecules on the surfaces of the Fe₂O₃@WO₃ nanocomposite because of its ultrahigh BET specific surface area (1207 m² g⁻¹).

Fig. 4 XPS spectra of the 5%Fe₂O₃@WO₃ nanocomposite obtained by microwave-heating process: (a) a survey scan, (b) W 4f, (c) Fe 2p and (d) O1s.

Efficient control in particle sizes and location distribution of a second phase is very important in the construction of hierarchical nanostructures. In the present Fe₂O₃@WO₃ system synthesized via the MH process, the particle sizes and number density of the α-Fe₂O₃ NPs anchored on the surfaces of WO₃ nanoplates can be readily adjusted by changing the Fe-ion concentration in the precursors, i.e., by changing the contents of α-Fe₂O₃ in the Fe₂O₃@WO₃ composites. The typical SEM images of the 2.5%Fe₂O₃@WO₃ and 9%Fe₂O₃@WO₃ composites are shown in Fig. 5a and b, respectively. One can see clearly that the number densities of α-Fe₂O₃ NPs in 5% Fe₂O₃@WO₃ (Fig. 2c) and 9% Fe₂O₃@WO₃ composites (Fig. 5b) are higher than that of the 2.5%Fe₂O₃@WO₃ composite (Fig. 5a). The particle sizes of the α-Fe₂O₃ NPs in the Fe₂O₃@WO₃ composites are statistically analyzed according to the SEM observations (Fig. 2c, 5a and b) are shown in Fig. 5c. The mean particle size of the α-Fe₂O₃ NPs in the Fe₂O₃@WO₃ nanocomposites increases from ∼9 nm to ∼13 nm when the apparent content of Fe₂O₃ NPs increases from 2.5% to 9%. It should be noted that the particle sizes of the α-Fe₂O₃ NPs measured according to the SEM images are a little larger than those (∼5 nm) observed in the TEM images (Fig. 2e and f). The possible reason should be the Pt-coating during the preparation of SEM samples.

Fig. 5 (a and b) Typical SEM images of (a) 2.5%Fe₂O₃@WO₃ and (b) 9%Fe₂O₃@WO₃ nanocomposites synthesized via the microwave-heating process. (c) The sizes of the α-Fe₂O₃ NPs immobilized on the surfaces of WO₃ nanoplates according to the statistical analysis of the SEM observations.

To demonstrate the efficiency and advantages of the microwave heating (MH) process in the construction of hierarchical nanostructures, we comparatively developed a water-bath heating (WH) process to synthesize the Fe₂O₃@WO₃ composites using the similar synthetic parameters except the heating manner. As a typical example, Fig. 6a shows the SEM image of the 5%Fe₂O₃@WO₃ composite obtained via the WH process at 85 °C for 3 h. One can find that most of the α-Fe₂O₃ NPs are loosely attached on the surfaces of WO₃ nanoplates, and the location distribution is random and uneven in Fig. 6a. In addition, many dissociative α-Fe₂O₃ NPs with aggregated structures can be found during the SEM observation (Fig. 6a). Comparatively, the 5%Fe₂O₃@WO₃ composite obtained via the MH process consists of monodispersed α-Fe₂O₃ NPs tightly anchored on the WO₃ nanoplates evenly, and few dissociative α-Fe₂O₃ NPs can be find in the SEM observation. We therefore infer that in the MH process the α-Fe₂O₃ NPs are in situ grown on the WO₃ nanoplates via the heterogeneous nucleation mechanism, whereas in the WH process, the α-Fe₂O₃ NPs are firstly formed mainly via a homogenous nucleation mechanism, and then loosely adsorbed on the surfaces of WO₃ nanoplates. The above inference can also be corroborated by the experimental phenomena: the filtrates are colorless in the MH process, whereas they are light brick-red in the WH process. The light brick-red suspension should be discrete Fe(OH)₃ clusters, which then turn to be α-Fe₂O₃ NPs by calcination.

Fig. 6 (a) Typical SEM image and (b) N₂ adsorption–desorption isotherms of the 5%Fe₂O₃/WO₃ nanocomposite obtained via the water-heating (WH) process at 85 °C for 3 h.

The advantages of the MH process in the synthesis of hierarchical nanostructures are obvious when compared with the WH process. Firstly, the second-phase nanocrystals can be tightly anchored on the supports via the in situ heterogeneous nucleation growth because the high-efficient microwave energy can be selectively reacted with the precursor system. Secondly, the MH process is much rapider than the WH process in the synthesis of hierarchical nanostructures. For example, the synthesis of 5%Fe₂O₃@WO₃ nanocomposite requires more than 3 h in the WH process, but just 4 min in the MH process. In addition, as the N₂ adsorption–desorption isotherms shown in Fig. 6b, the BET specific surface area of the 5%Fe₂O₃@WO₃ composite obtained via the WH process is 300 m² g⁻¹, which is far below that (1207 m² g⁻¹) of the 5%Fe₂O₃@WO₃ composite obtained via the MH process.

3.2 Gas-sensing performance

Hierarchical metal oxide nanocomposites with high specific surface areas are promising in developing high-performance gas-sensors. The Fe₂O₃@WO₃ nanocomposites obtained via the MH process are thus used as the active materials to detect harmful and combustible substances. Fig. 7 shows the typical H₂S-sensing properties of the Fe₂O₃@WO₃ nanocomposites, being compared with the WO₃ nanoplates and α-Fe₂O₃ nanoparticles obtained by a hydrothermal method. Fig. 7a shows the typical H₂S-sensing (U–t) profiles of the sensors operating at 150 °C upon exposure to H₂S gases with a concentration range of 0.5–10 ppm. The U–t curves show the change in voltage (U) loaded on the standard resistor as a function of the time (t) when the target gas are injected and discharged. One can see that the U values sharply rise and drop when the H₂S gas are injected and discharged for the sensors derived from the Fe₂O₃@WO₃ nanocomposites under the testing conditions, indicating that the Fe₂O₃@WO₃ sensors are highly sensitive to low-concentration H₂S gases and their resistances decrease quickly upon exposure to H₂S gases. Comparatively, the responses of the sensors derived from the pure WO₃ nanoplates and the α-Fe₂O₃ nanoparticles are relatively weak. The response details upon exposure to a 10 ppm H₂S gas are shown in Fig. 7b. One can see that the Fe₂O₃@WO₃ nanocomposites with various amounts (2.5–9 wt%) of α-Fe₂O₃ NPs show much faster and higher response than WO₃ nanoplates and Fe₂O₃ nanoparticles. Fig. 7c shows the plots of the response (R_a/R_g) of the sensors versus the concentration of H₂S gases ([H₂S]/ppm). The responses of all the sensors increase with the increases of [H₂S]. For a given [H₂S] value, the sensitivity in a descending order is 5%Fe₂O₃@WO₃ > 2.5%Fe₂O₃@WO₃ > 9%Fe₂O₃@WO₃ > WO₃ > α-Fe₂O₃. Typically, the response of the 5%Fe₂O₃@WO₃ sensor upon exposure to a 10 ppm-H₂S gas is as high as 192, up to 4-fold that (43) of the WO₃-nanoplate sensor and 17-fold that (11) of the α-Fe₂O₃-nanoparticle sensor. For a given [H₂S] value, the responses of the Fe₂O₃@WO₃ sensors increase firstly and then decrease with the increase of the α-Fe₂O₃ amounts, and the optimal amount of α-Fe₂O₃ is about 5 wt% under the present test condition.

Fig. 7 (a and b) Typical U–t response profiles of the Fe₂O₃@WO₃ (WH) nanocomposite operating at 150 °C upon exposure to H₂S gases with various concentrations: 0.5–10 ppm (a) and 10 ppm (b); (c) the plots of the [H₂S]-dependent responses of the WO₃-based sensors operating at 150 °C (A: 5%Fe₂O₃@WO₃, B: 2.5%Fe₂O₃@WO₃, C: 9%Fe₂O₃@WO₃, D: WO₃, E: α-Fe₂O₃). (d) The operating-temperature-dependent responses of the sensors derived from WO₃, 5%Fe₂O₃@WO₃ and α-Fe₂O₃ upon exposure to 5 ppm H₂S.

Operating temperature usually highly influences the response of a gas sensor based on metal oxide semiconductors. Fig. 7d shows the typical responses of the sensors derived from the WO₃ nanoplates, α-Fe₂O₃ nanoparticles and 5%Fe₂O₃@WO₃ nanocomposite upon exposure to 5 ppm H₂S operating at different temperatures (i.e., 100–250 °C). Both the WO₃ and Fe₂O₃@WO₃ sensors have a maximum response at around 150 °C, and their responses are 27 and 115, respectively. The response changes of the 5%Fe₂O₃@WO₃ sensor are much larger than those of the WO₃-nanoplate sensor from 100 to 250 °C. For the α-Fe₂O₃-nanoparticle sensor, the responses are less than 10 at the operating-temperature range of 100–250 °C.

The response and recovery times, embodying the response speeds, are important parameters in evaluating the performance of a gas-sensor. Fig. 8 shows the response and recovery times of the sensors derived from the WO₃ nanoplates and 5%Fe₂O₃@WO₃ nanocomposite upon exposure to 5 ppm H₂S operating at 100–250 °C. One can see that a higher operating temperature is helpful to shorten the response and recovery times, and that the 5%Fe₂O₃@WO₃ sensor has shorter response and recovery times than the WO₃ sensor. Operating at 100 °C, the response and recovery times of the 5%Fe₂O₃@WO₃ sensor are 17 s and 150 s, respectively (Fig. 8b), much shorter than those (239 s and 560 s) of the WO₃ sensor (Fig. 8a). When the operating temperature is higher than 150 °C, the response time of the 5%Fe₂O₃@WO₃ sensor is as short as 3–9 s, and its recovery time is about 10–38 s. The rapid response of the Fe₂O₃@WO₃ sensors should be attributed to the hierarchical nanostructure and ultrahigh specific surface areas, providing efficient diffusion paths and adsorption sites for gas molecules.

Fig. 8 Response times and recovery times of the sensors derived from (a) WO₃ nanoplates and (b) 5%Fe₂O₃@WO₃ (MH) nanocomposite upon exposure to 5 ppm H₂S at an operating temperature range of 100–250 °C.

The selective response of a gas sensor is also important in practical applications. Fig. 9 shows the selectivity of the sensors based on the WO₃ nanoplates and 5%Fe₂O₃@WO₃ nanocomposite operating at 150 °C. Various gases and organic vapors, including CH₄, CO, H₂, SO₂, methanol, ethanol, isopropanol, methanol, acetone and benzene, with apparent concentrations of 100 ppm, are used as target substances to evaluate the selective response performance, comparing with 2 ppm H₂S. One can see that the WO₃-based sensors have the highest responses to 2 ppm H₂S, and the responses are less than 10 upon exposure to the other substances. The 5%Fe₂O₃@WO₃ sensor has a higher selectivity than the WO₃-nanoplate sensor under the same test conditions. The hierarchical Fe₂O₃@WO₃ nanocomposites are therefore suitable for H₂S detection at low operation temperatures (e.g., ∼150 °C) due to its high response and selectivity to H₂S.

Fig. 9 Selective response at 150 °C of the sensors derived from the WO₃ nanoplates and 5%Fe₂O₃@WO₃ composite obtained by microwave-heating process.

To demonstrate the advantages of the microwave-heating process, we compare the H₂S response of the 5%Fe₂O₃@WO₃ nanocomposites obtained by microwave-heating process (5%Fe₂O₃@WO₃-MH) and water-heating process (5%Fe₂O₃@WO₃-WH). Fig. 10 shows the typical responses of the 5%Fe₂O₃@WO₃-MH and 5%Fe₂O₃@WO₃-WH sensors operating at 150 °C upon exposure to H₂S gases with various concentrations (0.5–10 ppm). It is clear that the Fe₂O₃ NPs highly enhance the H₂S-sensing performance of WO₃ nanoplates, and that the microwave-heating process is more favorable in forming higher H₂S-sensing Fe₂O₃@WO₃ nanocomposites than the water-heating process. The possible reasons for the enhancement of the MH-process are its efficient control in more uniform distribution and smaller particle sizes of Fe₂O₃ NPs anchored on the surfaces of WO₃ nanoplates, resulting in ultrahigh specific surface areas of the Fe₂O₃@WO₃-MH samples.

Fig. 10 The plots of the [H₂S]-dependent responses of the WO₃, 5%Fe₂O₃@WO₃-MH and 5%Fe₂O₃@WO₃-WH sensors operating at 150 °C upon exposure to H₂S gases with various concentrations (0.5–10 ppm).

3.3 Possible H₂S-sensing mechanism

The hierarchical Fe₂O₃@WO₃ nanocomposites obtained via the MH process with α-Fe₂O₃ NPs uniformly immobilized on the surfaces of WO₃ nanoplates have a high specific surface area (1207 m² g⁻¹). the H₂S gas-sensing performance of the sensors derived from the Fe₂O₃@WO₃ nanocomposite has been greatly improved, and they show high response, high selectivity and rapid response.

The gas-sensing performance of n-type semiconductors of WO₃ and α-Fe₂O₃ nanocrystals relies upon the interaction of the target gases with adsorbed oxygen species. In air, oxygen molecules are adsorbed onto the surfaces of the α-Fe₂O₃ NPs and WO₃ nanoplates, and are then transferred to O⁻, O²⁻or O₂ ions by trapping electrons from the conductive bands of the n-type oxides (α-Fe₂O₃ and WO₃), forming electron depletion layers with a high resistance (Fig. 11A). When the Fe₂O₃@WO₃ sensors are exposed to H₂S gas, the H₂S molecules react with the chemisorbed oxygen species (eqn (4) and (5)), releasing electrons back to the conduction bands of α-Fe₂O₃ and WO₃ (Fig. 11B). As a result, the concentration of electrons on the surface of the Fe₂O₃@WO₃ nanocomposites increases and the electron depletion layers decrease, forming a low-resistance state.

H₂S + 3O⁻ (ads) → H₂O + SO₂ + 3e⁻ (4)

H₂S + 3O²⁻ (ads) → H₂O + SO₂ + 6e⁻ (5)

Fig. 11 A schematic demonstration of the adsorption and reaction process of O₂ and H₂S molecules at the interface of the hierarchical Fe₂O₃@WO₃ nanocomposite.

As the operating temperature increases, high response is achieved because of the activation of adsorbed molecular oxygen and lattice oxygen (Fig. 11A). This phenomenon continues up to a certain optimum temperature, beyond which exothermic gas adsorption becomes difficult and gas molecules begin to desorb in large quantities, leading to a drop in sensor response.⁴⁰ Thus, the optimum temperature is a balance point between the above two conflicting aspects (i.e., activation and desorption).

The synergistic effect of WO₃ nanoplates and α-Fe₂O₃ NPs is the key factor improving the H₂S-sensing performances of the Fe₂O₃@WO₃ sensors. The additional depletion layers enhance the “volume depletion” because the Fe₂O₃@WO₃ nanocomposites have ultrahigh specific surface areas (i.e., 1207 m² g⁻¹ for the 5%Fe₂O₃@WO₃ sample), which allow them to absorb more gas molecules.⁶ Additionally, the highly monolayer-dispersed Fe₂O₃ NPs on the surfaces of WO₃ nanoplates provides efficient and rapid electron-exchange between the cations: Fe(III) ↔ Fe(II).⁴¹ All the above aspects can improve the gas-sensing performance of the Fe₂O₃@WO₃ sensors. Although the exact mechanism for understanding the enhancement in H₂S-sensing process is not available, the synergistic effect of the α-Fe₂O₃ and WO₃ species, not only in chemical compositions but also in microstructures, should be the essential origin to improve the low-temperature gas-sensing property of the WO₃-based materials.

4. Conclusions

Hierarchical Fe₂O₃@WO₃ nanocomposites have been fabricated via a simple microwave-heating process by in situ growing α-Fe₂O₃ NPs on the surfaces of WO₃ nanoplates. The α-Fe₂O₃ NPs with particle sizes of 5–10 nm are uniformly, tightly anchored on the surfaces of WO₃ nanoplates. The BET specific surface area of the 5%Fe₂O₃@WO₃ composite is up to 1207 m² g⁻¹, 5.9 times higher than that of the WO₃ nanoplates. The Fe₂O₃@WO₃ nanocomposites are particularly sensitive and selective toward H₂S gas. The 5%Fe₂O₃@WO₃ sensor shows the highest H₂S-sensing property at the optimum operating temperature of 150 °C, and its sensitivity upon exposure to a10 ppm-H₂S gas is as high as 192, 4 times higher than that of the WO₃ nanoplates. The improved low-temperature gas-sensing performance ought to be attributed to the synergistic effect of the α-Fe₂O₃ and WO₃ species in chemical compositions and microstructures. The hierarchical nanostructures consisting of zero-dimensional nanoparticles on two-dimensional ultrathin nanoplates and the microwave-heating process developed in this work provide a simple and robust strategy to achieve high-performance gas-sensing materials.

Acknowledgements

This work was partly sponsored by National Natural Science Foundation of China (51172211, 512101207, 51172213), China Postdoctoral Science Foundation (2013M531682, 2014T70682), Foundation for University Young Key Teacher by Henan Province (2011GGJS-001), Program for Science & Technology Innovation Talents in Universities of Henan Province (14HASTIT011), Special Support Program for High-End Talents of Zhengzhou University (ZDGD13001), and Technology Foundation for Selected Overseas Chinese Scholar (Ministry of Human Resources and Social Security). Prof. D. Chen thanks Prof. Bing Zhang (Zhengzhou University) for his kind help in measuring BET surface area.

References

1. M. M. Arafat, B. Dinan, S. A. Akbar and A. S. M. A. Haseeb, Sensors, 2012, 12, 7207–7258.
2. C. X. Wang, L. W. Yin, L. Y. Zhang, D. Xiang and R. Gao, Sensors, 2010, 10, 2088–2106.
3. Y. Wang, S. Wang, H. Zhang, X. Gao, J. Yang and L. Wang, J. Mater. Chem. A, 2014, 2, 7935–7943.
4. K. Wetchakun, T. Samerjai, N. Tamaekong, C. Liewhiran, C. Siriwong, V. Kruefu, A. Wisitsoraat, A. Tuantranont and S. Phanichphant, Sens. Actuators, B, 2011, 160, 580–591.
5. N. Yamazoe and K. Shimanoe, J. Electrochem. Soc., 2008, 155, 85–92.
6. N. Yamazoe and K. Shimanoe, J. Electrochem. Soc., 2008, 155, 93–98.
7. J. H. Lee, Sens. Actuators, B, 2009, 140, 319–336.
8. A. Gurlo, Nanoscale, 2011, 3, 154–165.
9. K. Potje-Kamloth, Chem. Rev., 2008, 108, 367–399.
10. D. Bekermann, A. Gasparotto, D. Barreca, C. Maccato, E. Comini, C. Sada, G. Sberveglieri, A. Devi and R. A. Fischer, ACS Appl. Mater. Interfaces, 2012, 4, 928–934.
11. B. Cao, J. Chen, X. Tang and W. Zhou, J. Mater. Chem., 2009, 19, 2323–2327.
12. D. Chen, X. Hou, T. Li, L. Yin, B. Fan, H. Wang, X. Li, H. Xu, H. Lu, R. Zhang and J. Sun, Sens. Actuators, B, 2011, 153, 373–381.
13. T. Stoycheva, F. E. Annanouch, I. Gràcia, E. Llobet, C. Blackman, X. Correig and S. Vallejos, Sens. Actuators, B, 2014, 198, 210–218.
14. C. S. Rout, M. Hegde and C. N. R. Rao, Sens. Actuators, B, 2008, 128, 488–493.
15. J. Xiao, P. Liu, Y. Liang, H. B. Li and G. W. Yang, Nanoscale, 2012, 4, 7078–7083.
16. S. Bai, K. Zhang, J. Sun, D. Zhang, R. Luo, D. Li and C. Liu, Sens. Actuators, B, 2014, 197, 142–148.
17. S. An, S. Park, H. Ko and C. Lee, Appl. Phys. A: Mater. Sci. Process., 2012, 108, 53–58.
18. M. Bao, Y. Chen, F. Li, J. Ma, T. Lv, Y. Tang, L. Chen, Z. Xu and T. Wang, Nanoscale, 2014, 6, 4063–4066.
19. N. S. Ramgir, C. P. Goyal, P. K. Sharma, U. K. Goutam, S. Bhattacharya, N. Datta, M. Kaur, A. K. Debnath, D. K. Aswal and S. K. Gupta, Sens. Actuators, B, 2013, 188, 525–532.
20. T. Kida, A. Nishiyama, Z. Hua, K. Suematsu, M. Yuasa and K. Shimanoe, Langmuir, 2014, 30, 2571–2579.
21. Y.-J. Zhu and F. Chen, Chem. Rev., 2014, 114, 6462–6555.
22. K. Sivula, F. L. Formal and M. Grätzel, Chem. Mater., 2009, 21, 2862–2867.
23. W. Luo, T. Yu, Y. Wang, Z. Li, J. Ye and Z. Zou, J. Phys. D: Appl. Phys., 2007, 40, 1091.
24. T. Jin, P. Diao, Q. Wu, D. Xu, D. Hu, Y. Xie and M. Zhang, Appl. Catal., B, 2014, 148–149, 304–310.
25. S. Bai, K. Zhang, J. Sun, R. Luo, D. Li and A. Chen, CrystEngComm, 2014, 16, 3289–3295.
26. Y. Wang, J. Cao, S. Wang, X. Guo, J. Zhang, H. Xia, S. Zhang and S. Wu, J. Phys. Chem. C, 2008, 112, 17804–17808.
27. V. Balouria, A. Kumar, S. Samanta, A. Singh, A. K. Debnath, A. Mahajan, R. K. Bedi, D. K. Aswal and S. K. Gupta, Sens. Actuators, B, 2013, 181, 471–478.
28. C. Zhao, W. Hu, Z. Zhang, J. Zhou, X. Pan and E. Xie, Sens. Actuators, B, 2014, 195, 486–493.
29. P. Sun, Y. Cai, S. Du, X. Xu, L. You, J. Ma, F. Liu, X. Liang, Y. Sun and G. Lu, Sens. Actuators, B, 2013, 182, 336–343.
30. Y. Wang, S. Wang, Y. Zhao, B. Zhu, F. Kong, D. Wang, S. Wu, W. Huang and S. Zhang, Sens. Actuators, B, 2007, 125, 79–84.
31. J. Ma, L. Mei, Y. Chen, Q. Li, T. Wang, Z. Xu, X. Duan and W. Zheng, Nanoscale, 2013, 5, 895–898.
32. D. Chen, L. Gao, A. Yasumori, K. Kuroda and Y. Sugahara, Small, 2008, 4, 1813–1822.
33. D. Chen, L. Yin, L. Ge, B. Fan, R. Zhang, J. Sun and G. Shao, Sens. Actuators, B, 2013, 185, 445–455.
34. D. Chen and Y. Sugahara, Chem. Mater., 2007, 19, 1808–1815.
35. D. Chen, X. Hou, H. Wen, Y. Wang, H. Wang, X. Li, R. Zhang, H. Lu, H. Xu, S. Guan, J. Sun and L. Gao, Nanotechnology, 2010, 21, 035501.
36. K. S. W. Sing, D. H. Everett, R. A. W. Haul, L. Moscou, R. A. Pierotti, J. Rouquerol and T. Siemieniewska, Pure Appl. Chem., 1985, 57, 603–619.
37. Y.-J. Chen, X.-M. Gao, X.-P. Di, Q.-Y. Ouyang, P. Gao, L.-H. Qi, C.-Y. Li and C.-L. Zhu, ACS Appl. Mater. Interfaces, 2013, 5, 3267–3274.
38. A. F. Carley, M. W. Roberts and A. K. Santra, J. Phys. Chem. B, 1997, 101, 9978–9983.
39. J. Deng, L. Zhang, H. Dai, H. He and C. T. Au, Ind. Eng. Chem. Res., 2008, 47, 8175–8183.
40. Z. Li, X. Lai, H. Wang, D. Mao, C. Xing and D. Wang, J. Phys. Chem. C, 2009, 113, 2792–2797.
41. L. A. Patil, M. D. Shinde, A. R. Bari, V. V. Deo, D. M. Patil and M. P. Kaushik, Sens. Actuators, B, 2011, 155, 174–182.

---