The α-Fe₂O₃ nanorhombs were prepared via a solvothermal route.^[18] Typically, 0.202 g of Fe(NO₃)₃ 9H₂O, 0.300 g of poly(N-viny1-2-pyrrolidone) (PVP, M_w = 30000), and 0.277-g CH₃COONa were dissolved in 30-ml methanol aqueous solution with constant stirring at room temperature. The volume ratio of DI water and methanol is 2:1. After stirring for 1 h, the brownish solution was then turned into a Teflon-lined stainless steel autoclave of 50-ml capacity. The sealed tank was put into an oven and heated at 200 °C for 24 h. After reaction, the autoclave was cooled to room temperature naturally. The reddish brown precipitates were collected by centrifugation, washed with DI water and ethanol several times, and dried at 60 °C in air. Finally, the product was put into a muffle oven for calcination in air atmosphere at 450 °C for 4 h with the heating rate of 3 °C/min.

Transmission electron microscopy (TEM) images were taken with a JEOL JEM-2100 microscope to characterize the morphology of α-Fe₂O₃ nanorhombs. Powder x-ray diffraction (XRD) analyses were performed on a Bruker D8 Advance diffractometer with Cu Kα radiation ( ). X-ray photoelectron spectra (XPS) were performed on an RBD upgraded PHI-5000 C ESCA system (PerkinElmer). XRD and XPS are used for elemental and structural analyses.

The gas sensor devices were based on an SOI (silicon on insulator) MEMS microheater technology. Figure 1(a) displays the schematic structure of the microheater, and the corresponding optical micrograph is shown in Fig. 1(b). The circuitry, both for heater temperature control and sensing layer resistance readout, was integrated on-chip. A pair of heating electrodes to provide highly uniform working temperature of the sensing material at low power, and interdigitated Au electrodes to form a good Ohmic contact with the sensing material were suspended in a dielectric membrane in the microheater device. The gas sensors were prepared by integrating α-Fe₂O₃ nanorhombs onto the center of a microheater. The α-Fe₂O₃ nanorhombs were sonicated into suspension in an aqueous solution of isopropyl alcohol (0.5 mg/mL). A 0.1- drop was placed onto the microheater chip with the heater working at 100 °C and maintained for 2 h to promote solvent evaporation and material deposition at the center of the microheater.

Fig. 1.

	Figure Option ViewDownloadNew Window
	Fig. 1. (a) Schematic structure and (b) optical micrograph of the microheater, (c) the working principle of the gas sensing measurement.

The gas sensing properties were evaluated in a homemade measuring system via a static process. For the test, a reference resistor (R_ref.) was put in series to form a measurement circuit. The circuit diagram is illustrated in Fig. 1(c). The response of the sensor in air or in a target gas could be measured, by monitoring the voltage across the reference resistor. Response is defined by the percentile resistance change when the sensor is exposed to an analyte gas as follows:

Morphology of the as-synthesized α-Fe₂O₃ is characterized using TEM. Figure 2(a) and 2(b) show representative low magnification TEM images of the α-Fe₂O₃ nanorhombs. The nanoparticle size ranges from 45 nm to 65 nm. The measured spacing between the fringes in the lattice is approximately 0.27 nm (Fig. 2(c)), which matches well the (104) d-spacing for the α-Fe₂O₃. The distinct lattice fringes indicate the high crystallinity of the grains.

Fig. 2.

	Figure Option ViewDownloadNew Window
	Fig. 2. The α-Fe2O3 nanorhombs: (a) and (b) low magnification TEM images, (c) high resolution micrograph, (d) x-ray diffraction spectrum.

The x-ray diffraction pattern (Fig. 2(d)) shows characteristic peaks for α-Fe₂O₃ at , 33.2°, 35.7°, 40.9°, 49.5°, 54.1°, 62.5°, and 64.1°, which respectively correspond to the (012), (104), (110), (113), (024), (116), (214), and (300) planes of the hexagonal α-Fe₂O₃ (JCPDS 79–0007). This XRD results agree with the previous reports.^{[28, 29]} No other phases are observed, indicating the purity and single phase of the prepared α-Fe₂O₃ nanorhombs. The strongest peak at is corresponding to the α-Fe₂O₃ (104), which is consistent with its TEM result (Fig. 2(c)).

The XPS survey spectrum of the α-Fe₂O₃ nanorhombs is shown in Fig. 3(a). It reveals the presence of Fe, O, and C elements. According to the Fe 2p high-resolution XPS spectrum in Fig. 3(b), the binding energy peaks at 711.4 eV and 724.3 eV with the corresponding satellite peak at 718.6 eV, are attributed to Fe 2p_3/2, 2p_1/2, respectively.^[15] The energy positions of these peaks indicate a Fe valence state of +3. Two obvious peaks were observed in the O 1s high resolution XPS spectrum as shown in Fig. 3(c), which have the binding energy values of 530.5 eV and 532.7 eV. The peak at 530.5 eV is associated with the lattice oxygen atoms in α-Fe₂O₃ and the peak at 532.7 eV is linked with the chemisorbed oxygen species on the surface of α-Fe₂O₃.^[30] The large peak area of chemisorbed oxygen species means many chemisorbed oxygen species on the surfaces of α-Fe₂O₃ nanorhombs.^[15]

Fig. 3.

	Figure Option ViewDownloadNew Window
	Fig. 3. (a) X-ray photoelectron spectroscopy (XPS) survey spectrum, (b) Fe 2p and (c) O 1s high-resolution spectra of the α-Fe2O3 nanorhombs.

The gas sensing performances of the α-Fe₂O₃ sensor are related to the working temperature closely. This is because that the adsorption/desorption rates of oxygen molecules and target gas molecules on the sensing material surface are affected by temperature. The gas sensing properties of the α-Fe₂O₃ sensor are investigated at different operating temperatures of the heater. Nearly linear current versus voltage behaviors at different temperatures suggest an Ohmic contact between α-Fe₂O₃ and sensor electrodes (Fig. 4(a)). The decreasing resistance with increasing the working temperature is consistent with a typical behavior of semiconductor. At a low temperature ( ), the α-Fe₂O₃ sensor cannot show an obvious resistance change to ethanol, because the ethanol molecules do not have enough thermal energy to react with the adsorbed oxygen.^[20] The response values of the α-Fe₂O₃ sensor to 3-ppm ethanol at various working temperatures from 180 °C to 310 °C are shown in Fig. 4(b). The response of the sensor to 3 ppm ethanol increases as the temperature increases, while the response and recovery time decreases. This is due to the high temperature accelerating the adsorption and desorption of ethanol molecules on the α-Fe₂O₃ surface. However, both the response and recovery time are slight at 280 °C and 310 °C, 3 s and 15 s respectively. Considering the response and recovery properties to ethanol and the power consumption of the device, a temperature of 280 °C is taken as the optimum operating temperature, with the power of microheater at 60 mW. Further sensing tests are taken at 280 °C with the ambient humidity of 30%.

Fig. 4.

	Figure Option ViewDownloadNew Window
	Fig. 4. (a) I–V curves of the α-Fe2O3 nanorhombs-based sensor at different working temperatures. (b) Response values of the sensor to 3-ppm ethanol at different working temperatures.

The α-Fe₂O₃ sensor response to varied concentrations of ethanol from 1 ppm to 20 ppm at 280 °C is shown in Fig. 5(a). The resistance of the sensor decreases upon exposure to ethanol gas, consistent with n-type behavior.^[31] For all different concentrations of ethanol, this gas sensor can be fully recovered (e.g., returning back to the baseline) after ethanol gas is quickly replaced by air, indicating the excellent reversibility of the α-Fe₂O₃ nanorhombs-based sensor. The response increases monotonically with the ethanol gas concentration increasing in the low concentration level of 1 ppm–7 ppm, while tending to saturate above 20 ppm (Fig. 5(b) and Fig. A1 in Appendix A). The response time (t_response) is defined as the time taken to reach 90% of the full response after the introduction of the target gas. The recovery time (t_recovery) is defined as the time taken to return to 90% of the baseline resistance after the flow of target gas is stopped. The sensor has ultra-fast response and recovery times for all ethanol concentrations around 3 s and 15 s, respectively.

Fig. 5.

	Figure Option ViewDownloadNew Window
	Fig. 5. The α-Fe2O3 nanorhombs-based sensor working at 280 °C: (a) Resistance versus time for various ethanol concentrations (1 ppm–20 ppm), (b) the corresponding response values, (c) repeatability for 3 ppm ethanol, (d) selectivity for various gases.

Fig. A1.

	Figure Option ViewDownloadNew Window
	Fig. A1. The α-Fe2O3 nanorhombs-based sensor working at 280 °C: (a) Resistance versus time for various ethanol concentrations (1 ppm–100 ppm), (b) the corresponding response values.

Repeated response/recovery measurement of the α-Fe₂O₃ sensor to 3-ppm ethanol gas for 7 cycles are shown in Fig. 5(c), which indicates well repeatable behavior. Additionally, the selectivity of the α-Fe₂O₃ sensor is displayed in Fig. 5(d) among the gases of methanol, acetone, isopropyl alcohol (IPA), formaldehyde, and ammonia gas at higher concentrations, while the measurement of ethanol gas is at a lower concentration of 3 ppm. The results confirm the great selectivity of the α-Fe₂O₃ sensor to ethanol molecules.

Excellent long-term stability of gas sensors is another essential requirement for reliable applications. The responses of the α-Fe₂O₃ sensor to 3-ppm ethanol gas were measured over a period of 50 days, while the sensor is always on at the working temperature of 280 °C. The response values and the reproducible resistance time are shown in Fig. 6(a) and Fig. A2 in Appendix A. An insignificant response average variation of around 4% was observed (Fig. 6(a)). Hence, the α-Fe₂O₃ nanorhombs-based sensor exhibits a great long-term stability.

Fig. 6.

	Figure Option ViewDownloadNew Window
	Fig. 6. The α-Fe2O3 nanorhombs-based sensor working at 280 °C: (a) long-term stability to 3-ppm ethanol, and (b) response to various ethanol concentrations (1 ppm–10 ppm) in different humidity conditions.

Fig. A2.

	Figure Option ViewDownloadNew Window
	Fig. A2. Repeatability for 3-ppm ethanol of the fresh α-Fe2O3-based sensor and the one over working at 280 °C for 50 days.

Humidity is a critical issue for all semiconducting solid-state sensor devices due to the varying humidity in realistic environment. As shown in Fig. 6(b), the response and recovery properties of the α-Fe₂O₃ sensor to ethanol gas at relative humidity (RH) of 30%, 50%, and 70% were tested. Here, RH was adjusted by humidifier and air conditioner, and monitored by a hygrometer. In this case concentrations were varied from 1 ppm to 10 ppm. Passing from 30% to 70% RH, the response average variation is below 6% for all the analyzed ethanol gas concentrations, indicating saturation of the material with respect to humidity at the operating temperature of 280 °C. This means the acceptable immunity to humidity of the α-Fe₂O₃ sensor at 280 °C. It can be used to effectively detect ethanol gas in realistic environment, and it is not vital to know the exact RH value. However, baseline of the α-Fe₂O₃ sensor has a slight drift at 70% RH (Fig. A3 in Appendix A). Therefore, ethanol sensing performances of the α-Fe₂O₃ sensor in high humidity should be improved.

Fig. A3.

	Figure Option ViewDownloadNew Window
	Fig. A3. Resistance versus time for various ethanol concentrations (1 ppm–20 ppm) of the α-Fe2O3-based sensor at 280 °C & 70% RH.

The gas sensing mechanism of α-Fe₂O₃ sensor has been clarified in many previous reports.^{[13, 31]} The model is based on the change in resistance caused by the adsorption and desorption of gas molecules on the surface of sensing films. In air ambient, the oxygen molecules are adsorbed on the α-Fe₂O₃ surface. The adsorbed oxygen molecules capture electrons from the conduction band of the α-Fe₂O₃ surface layer, which results in the formation of chemisorbed oxygen ions O⁻ at the operating temperature of 280 °C, according to the following equation:^[32]

Hence, an electron depletion layer is formed on the surface of α-Fe₂O₃, and a potential barrier is generated contributing to a high resistance state.

Upon exposure to ethanol gas, the ethanol molecules react with the adsorbed O⁻ ions on α-Fe₂O₃ surface and release the trapped electrons back to conduction band of α-Fe₂O₃, resulting in a reduced potential barrier energy. Accordingly, the sensor’s resistance is decreased. These can be expressed according to the following reactions:

To demonstrate the feasibility of real application of the α-Fe₂O₃ sensor, the sensor was integrated with a prototype of electronic circuit, which can transmit sensing data to mobile terminal and show the ethanol gas concentration values of the atmosphere. Figure 7 displays the working principle of a gas sensor network. Gas leakage or breath alcohol can be monitored or tested easily.

Fig. 7.

	Figure Option ViewDownloadNew Window
	Fig. 7. Gas sensor network connected by a mobile terminal. The integrated circuit is the picture of real products.