Design of ultrasensitive Ag-LaFeO3 methanol gas sensor based on quasi molecular imprinting technology

An ultrasensitive methanol gas sensing device based on the quasi-molecular imprinting technology (quasi-MIT) is studied in this work. We applied the sol-gel method (ALS denotes Ag-LaFeO3 prepared by the sol-gel method) and combustion synthesis (ALC denotes Ag-LaFeO3 prepared by combustion synthesis) to prepare Ag-LaFeO3 based sensors. The morphologies and structures of the Ag-LaFeO3 materials were examined via various detection techniques. The ALSM and ALCM sensor (ALSM and ALCM denotes the devices prepared by coating the ALS and ALC materials with methanol, respectively) fabricated using the sol-gel method and combustion synthesis combined with quasi-MIT exhibit good gas sensing properties to methanol, in contrast with the two devices (ALSW and ALCW denote the devices prepared for coating the ALS and ALC materials with water, respectively) without the use of quasi-MIT. The results show that quasi-MIT introduced the target gas in the fabrication process of the device, playing an important role in the design of the ultrasensitive methanol gas sensor. The sensing response and the optimum working temperature of ALSM and ALCM gas sensor are 52.29 and 155 °C and 34.89 and 155 °C, respectively, for 5 ppm methanol, and the highest response to other gases is 8. The ALSM and ALCM gas sensors reveal good selectivity and response for methanol.

gas-sensitive characteristics and related mechanisms of methanol gas detection by the ALSW, ALSM, ALCW and ALCM were carefully investigated. It was found that ALSM and ALCM exhibited ultrahigh sensitivity.
After the pre-synthesized Ag-LaFeO 3 precursor was obtained, ALS and ALC were sintered in 800 °C to 2 h in the air, with the XRD results showing the crystalline nature of the sample. All peaks are completely identical with the orthorhombic structure of LaFeO 3 as shown in Fig. 1. This diffraction pattern perfectly matches the standard JCPDS card no. 37-1493 18 . No precursor residue was detected, indicating that the samples are highly pure. The amount of Ag is so small (mol (Ag): mol (Ag-LaFeO 3 precursor) = 1%) that it cannot be detected by XRD and FTIR. The lattice parameters obtained from the refinement of the PXRD data are presented in Table 2. It is noteworthy that even though the substitution of Ag does not change the orthorhombic symmetry of the material, it results in a pseudo-tetragonal unit cell with the cell parameter 'a' almost equal to 'c' . It can also be seen from the data presented in Table 2 that the lattice parameters of ALS and ALC show a slight change compared to other reports in the literature [19][20][21][22][23][24][25][26] .
To check the purity of the samples and the interaction between the functional groups, we analyzed the Ag-LaFeO 3 cross-linker, ALS and methacrylic acid (MAA) by FT-IR spectroscopy in the 500-4000 cm −1 range as shown in Fig. 2. The FT-IR spectrum of the synthesized Ag-LaFeO 3 cross-linker (curve (a)) exhibits a peak    Curve (b) shows the curve of ALS, where the disappearance of the peaks at ~1210 cm −1 and ~1709 cm −1 compared to the MAA curve suggested an interaction between Ag-LaFeO 3 and MAA. The mutual effect is attributed to the coordination between the groups in MAA and lanthanum in the Ag-LaFeO 3 cross-linker, and it is proven that the metal carbonyl complex is formed 28 . The pattern and microspore structure of the ALC and ALS were detected via SEM and are shown in Fig. 3. Fig. 3(a,b) shows the appearance of the ALC as a tile, and Fig. 3(f,g) (at low magnification) exhibits stone-like morphologies of the ALS. Fig. 3(c,h) shows featured outer-sides of the ALC and ALS. The ALC shows bulk with porous features composed of many interconnected smaller sized particles, and the thickness of this sample is 0.8 μm. The outside of ALS is solid with the thicknesses of 2.1 μm. The enlarged image of the selected areas of the ALC and ALS samples ( Fig. 3(d,i)) shows a porous structure with a micron-scale size. The number of the pores on the surface of ALS is greater than that on the ALC surface. High-powered SEM images in Fig. 3(e,j) are also clearly displayed. Porous morphologies of the samples would make the target gases diffuse in or escape from the inner parts of the sensing films faster than for their solid counterparts and consequently will improve the responses to the target gases.
TEM provides insight into the morphology of the ALC and ALS as shown in Fig. 4. The images clearly show that the ALS ( Fig. 4(a)) exhibit irregular bulk shape, which is consistent with the results of SEM. As shown in the image, the ALS consisted of many particles with the diameters as high as tens of nanometers. Fig. 4(d) shows the bulk structural characteristics of the ALC. Based on the image shown in Fig. 4(e), ALC is found to be comprised of a large number of interstitial Ag-LaFeO 3 particles. There are many pores on the surface of the bulk with the pore size of approximately 50 nm, which is in accordance with the results observed in the high-magnification SEM image ( Fig. 3(j), 400000x). High-magnification TEM images of the ALC and ALS are shown in Fig. 4(b,f) and the clearly observed the lattice stripes indicate good crystallinity. To obtain a clearer image of the lattice fringes, Fig. 4(c,g) shows a more distinct image of the selected areas. The fringe spacing of ALS and ALC are 0.35 nm and 0.32 nm, respectively.
The specific surface areas of the ALC and ALS mesoporous materials are determined by the N 2 adsorption desorption isotherm curve and the distribution curves of the pore size are shown in Fig. 5. According to the International Union of Pure and Applied Chemistry, the H3 hysteresis loop indicates the pores are slit-like and formed by particle aggregation 29 . The ALS and ALC show obvious hysteresis loops at the P/P 0 ranges of 0.6-1. Thus, the existence of pores is proven.
The BJH method ( Fig. 5(a,b), inset) shows that ALS and ALC have porous structures. The size of the hole ranges from 2 nm to 90 nm, and the main peak is positioned at 45-47 nm. The S BET values of the ALS and ALC were 5.66 m 2 g −1 and 4.72 m 2 g −1 , respectively. The BET analysis indicated that ALS has a larger surface area, which means that ALS can provide sufficient surface for the gas reaction.
The factors important for the gas-sensing mechanism include the surface control procedure, particle size, surface condition and adsorption and desorption of oxygen 30 . The characteristics of the gas sensor are affected by the surface states of the ALS and ALC. The XPS spectra of ALC and ALS were obtained in order to analyze the valence states of each element. Fig. 6(a,e) shows the XPS results for the ALS and ALC, with the standard peak of C1s at 284.8 eV. The spectra confirm the presence of Ag (3d, 4p), La (3d, 4p, 4d), Fe (2p), and O (1 s), indicating the high purity of the product. The XPS La 3d and Fe 2p energy levels display that the lanthanum ions and iron ions exist in the +3 state 31 . For the La 3d core level peaks ( Fig. 6(b,f)), two peaks of lanthanum are seen from the peak; one is 3d 5/2 of La 3+ at 833.6 eV, and the other one is 3d 3/2 of La 3+ ions at 850.5 eV, which is due to the spin-orbit splitting of lanthanum oxide. Fig. 6(c,g) shows the XPS patterns of Fe 2p in ALS and ALC, and the two peaks of Fe 2p 3/2 and Fe 2p 1/2 were surveyed at 709.7 and 723.6 eV, respectively, which is due to the Fe 3+ in the oxides 32 . The O 1 s XPS patterns (Fig. 6(d,h)) are wide as well as asymmetrical. At least two O chemical states are observed; one is O L at 528.2 eV, and the other one is O H at 530.5 eV, where O L is attributed to lattice oxygen, and O H is ascribed to the chemically adsorbed O − or OH − . Therefore, the O 1 s XPS patterns fit two chemical conditions by Gauss' rule 33 . is attributed to the OH − resulting mainly from the chemically adsorbed water; this is due to lanthanum oxide, which can easily absorb moisture in the air. This metal-support interaction is a major element in the analysis of the mechanism of gas sensor. Figure 7(a,b) shows the XPS full spectra of Ag-LaFeO 3 and LaFeO 3 , respectively. It can be seen from Fig. 7(c,d) that after the introduction of Ag, the binding energy corresponding to Fe and La shifted to a small direction. This shows that the doping of Ag is successful, but the amount of doping is too small to be detected. Therefore, the high magnification of Ag 3d is not available.
In order to evaluate the response of the gas sensor to methanol, the response of the prepared materials to 5 ppm methanol was detected at the operating temperatures in the 5-250 °C range. The gas response value reaches the maximum at the operating temperatures of 155 °C, 175 °C, 155 °C and 155 °C for ALSW, ALSM, ALCW and ALCM, respectively; as the operating temperature increases, the response decreases, as shown in Fig. 7(a-d). Therefore, 175 °C, 155 °C, 155 °C and 155 °C were used as the best operational temperatures of the ALSW, ALSM, ALCW and ALCM sensors, respectively.
Then, the response and selectivity of the four sensors were analyzed based on the data in Fig. 8. The responses of ALSM (52.29, Fig. 8(a)) and ALCM (34.89, Fig. 8(d)) are obviously superior to those of ALSW (23.6, Fig. 8(b)) and ALCW (23.2, Fig. 8(c)) for 5 ppm methanol at the best working temperature, compared to the reported values, and a good response is observed, as shown in Table 3 [34][35][36][37][38][39][40][41] . All four sensors exhibit good selectivity to methanol. ALSM exhibits the best gas sensing properties among all devices (Fig. 8(e)) because the methanol molecules were introduced into the process of equipment preparation, which result in more methanol gas scattered on the surface of ALSM when ALSM was alternately placed in methanol gas and air. This will improve the gas response to methanol. Fig. 9 shows the relationships between the response and methanol concentration for the ALSM, ALSW, ALCW and ALCM in the methanol concentration range from 1 to 5.5 ppm. It can be seen in Fig. 9(a 1 -d 1 ) that the responses increased linearly with the methanol concentration increasing from 1 to 5.5 ppm. The sensors can be used for real-time detection of methanol at the optimal working temperature. It can be found in Fig. 9(a 2 -d 2 ) For sensors based on metal oxides, several factors affect the gas sensing properties, such as surface area, porous structure and particle size. This further influences the spread and adsorption of the gases 42 . Methanol can interact with Ag-LaFeO 3 in the preparation of the materials, which will further optimize the structure of ALSM. The structure of the materials was further optimized by introducing methanol into the device fabrication. This mechanism of the sensors ALSM and ALCM is defined as the excimer imprinting technique. Then, the mechanism of MIT was analyzed in detail [43][44][45] . ALSM and ALCM show better responses than ALSW and ALCW. This means that the preparation method of the device plays an important role in improving the response of the gas

Detection limit (ppm) References
Na  sensor. The gas sensing properties of the device are improved by the large surface area and multiple recognition sites of the gas sensing material 46,47 . The working principle of the sensor is based on the change of the conductivity caused by the reaction between the gas molecule and the surface of the material. The porous structure of the material promotes the reaction of the gas and further improves the gas sensing property of the material. During the test process, the sensor was exposed to air at first, and oxygen adsorbed on the surface of LaFeO 3 combined with the free electrons trapped on the This process will form a thin space charge layer, and relatively low barriers are formed to lower the resistance 49 . When the material is in the presence of methanol (CH 3 OH), the reducing gases react with the adsorbed oxygen forming CO 2 and H 2 O and releasing the electrons 50 . The space charge layer on the SLFO surface becomes thick, and the sensor resistance increases (Fig. 10). The reaction process is as follows: The design of a high-sensitivity methanol gas sensor based on quasi-MIT mechanism is reasonable. Furthermore, the pore number and specific surface area of ALSM are larger than those of ALCM, which accelerates the target gases to diffuse in or escape from the inner regions; therefore, ALSM displays a better gas response than ALCM. Our method is proven to be innovative for the design of a high-sensitivity methanol gas sensing device. This work provides important guidance for the design of high-response gas sensors.

Methods
Preparation of ALS. The chemical reagents used in the experiment were analytical reagents that were used without further treatment. The solvent is highly pure water (18 MΩ cm at 25 °C). In the prototype process, citric acid, lanthanum nitrate, iron nitrate, silver nitrate, PEG and deionized water were mixed under stirring at 90 °C for 10 hours as solution A. Then, solution A was placed in the microwave synthesizing device (CEM) for 4 hours at 75 °C. Methanol mixed with MAA was treated for 30 min by magnetic stirring and was let to stand to form solution B. Then, azobisisobutyronitrile (AIBN) was mixed with methanol, solution B and solution A. The final solution was treated by magnetic stirring at 50 °C for 12 h with a water bath and was then dried. Finally, the white powder of ALS was obtained after annealing at 800 °C for 2 h in an oven.
Preparation of ALC. The ALC powder was synthesized via the combustion synthesis; here, citric acid, lanthanum nitrate, iron nitrate, silver nitrate, PEG and deionized water were mixed under stirring at 90 °C for 10 hours as solution D. The final mixture was treated for 30 min by magnetic stirring, and the prepared samples were oven-dried at 85 °C. The xerogel was heated at 300 °C for 30 min, and then sintering at 800 °C was carried out to obtain ALC. Device fabrication. The mesoporous material was fabricated via the sol-gel method (ALS) as well as combustion synthesis (ALC). The as-synthesized samples (ALS, ALC) were mixed with deionized water (ALSW and ALCW) and methanol (ALSM and ALCM) and ground to a mushy solution. For the detailed description of the fabrication method of the sensors, see ref. 51 . Characterization. X-ray diffraction (XRD) spectra were obtained using a Japan AXS D/MAX-3BX advanced device handled at 25 mA and 35 kV with Cu Kα scattered at 1.5406 Å wavelengths. The detailed morphology features and crystallization conditions of the material were surveyed via field emission scanning electron microscopy (FESEM, S-3400N) as well as transmission electron microscopy (JEM-2100, Hitachi, Japan). The functional group was identified by Fourier transform infrared spectroscopy (FT-IR, FTS-40). The BET surface area (S BET ) of the material was measured using a Quadrasorb-evo instrument.