Ultrasensitive room-temperature ethanol detection based on Au functionalized nanoporous SnO2/C60/SnO2 composites

We have realized extremely sensitive, selective sub-ppm level ethanol gas detection through an Au functionalized nanoporous SnO2/C60/SnO2 sensing composite. This new type of hierarchically nanoporous SnO2 composite film supports both the Au@SnO2/C60 layer with closely packed open macropores (∼300 nm) and the bottom SnO2 layer with highly ordered nanopores (8–10 nm). The structure, crystallinity and composition of the hierarchical structured Au@SnO2/C60/SnO2 nanocomposite films are characterized by XRD, Raman, HRTEM, and SAED. The interaction between the Au/SnO2 and C60/SnO2 heterojunctions in light of transfer of charge carriers and modulation of potential barriers plays a crucial role in enhancing the detecting performance. The sensing properties of the Au@SnO2/C60/SnO2 nanocomposite sensor are investigated by exposing it to ethanol gas with a concentration range from 0.5 ppm to 50 ppm. Furthermore, the gas sensor exhibits ultrahigh sensitivity to ethanol gas with a response value of up to 16.8 at 0.5 ppm and a short recovery time of 9 seconds at room temperature.


Introduction
Ethanol as one of the most commonly and widely used chemical compounds, is applied in many elds such as the chemical industry, medicine, and the food industry. Furthermore, ethanol fuel is also a clean, efficient, and promising candidate for a renewable energy source that will overcome current challenges such as air pollution and the global energy crisis. In this growing ethanol fuel economy, safety challenges involving ethanol production, transportation, and storage require precise quantitative detection of ethanol vapour at sub-ppm level. Besides 1D nanomaterials, 1 hollow spheres, 2 and core-shell materials, 3 ordered nanoporous SnO 2 based sensors have undergone extensive development, because they usually offer acceptable sensing performance due to large active surface area and efficient gas diffusion induction. However, they suffered from an obvious drawback of high operating temperature, resulting in high power consumption and difficulty in integration. 4 Therefore, the design and synthesis of effective materials and devices for high sensing performance at room temperature are of crucial important for practical applications.
Recently, SnO 2 decorated with nanoscale carbon materials such as carbon nanotube, graphene, and fullerene have attracted much attention for gas sensor. [5][6][7] In particular, C 60 appears to be attractive for its special electric property because it is chemically stable, metallic impurity free and relatively simple to implement and gives rise to reproducible electrical responses. 8,9 Zito et al. synthesized the C-SnO 2 nanocomposite forming p/n heterojunction to get enhanced sensing performance for ethanol gas. It was proposed that the formation of p/n junction was responsible for the improved sensing performance of SnO 2 based sensing materials. 10 Qu et al. reported porous C 60 /SnO 2 composites exhibited high sensitivity to 5 ppm ethanol gas with operating temperature 100 C. 11 However, C 60 -based sensors with p/n heterojunction still had some disadvantages in modifying the diffusion and adsorption of gas molecules on the sensing materials surface, which led to the long responserecovery time. The response time to 5 ppm ethanol gas was as long as 100 second. For improving the response performance, Nguyet et al. fabricated n-p-n heterojunctions of C/SnO 2 nanowires to obtain high enhancement in gas sensing performance. 12 Therefore, the n/p/n heterojunction between SnO 2 nanoparticles and C 60 on the porous nanostructure would be able to enhance the ability of the gas molecules to diffuse and adsorb on the sensing materials, thus promoting the sensor performance.
On the other hand, several strategies demonstrated that the incorporation metal nanoparticles such as Au, Pt, Pd, and Ag to the metal oxides surface can effectively reduce the operating temperature and improve the lower limit detection, sensor response, sensitivity, and selectivity. [13][14][15][16] Au nanoparticles have received the most attention from the scientic community, especially for sensor applications, because of their chemical inertness, high adsorption, and excellent catalytic activity, [17][18][19][20] where the Au modied SnO 2 /C 60 /SnO 2 nanocomposite lm based gas sensor can be exploited to detect ethanol gas at room temperature.
In the present work, we synthesize nanoporous SnO 2 /C 60 / SnO 2 sensing lm decorated with Au applying the water vapour post-synthetic hydrothermal treatment, in order to form n/p/n heterojunction for improving sensing behaviour. This new type of hierarchically nanoporous SnO 2 composite lms supporting both the Au@SnO 2 /C 60 layer with closely packed open macropores ($300 nm) and the bottom SnO 2 layer with highly ordered nanopores (8-10 nm). The generation of p/n heterojunctions between C 60 /SnO 2 nanoparticles can inevitably develop the depletion regions of the junction, decreasing the conduction volume of SnO 2 nanoparticles. Furthermore, the potential barrier (0.17 eV) between SnO 2 /C 60 /SnO 2 heterojunction will be developed at both sides of n/p/n heterojunctions, being accompanied by the bending of the vacuum energy level and the energy band. Therefore, the SnO 2 /C 60 /SnO 2 heterojunctions will provide additional modulation of resistance during the introduction and removal of ethanol gas by changing the n/p/n heterojunction barrier. The addition of Au catalyst will provide a great signicant effect on the enhancement of sensing performance. By the spill over effect, the target gas molecules will be efficiently adsorbed on Au surface. They can be easily transported to the adjacent SnO 2 surface. 21 On the other hand, the Schottky barriers (0.55 eV) generated at Au-SnO 2 heterojunctions will provide additional modulation of resistance during the adsorption and desorption of ethanol gas by changing the heterojunction barrier. For hierarchical Au@SnO 2 /C 60 /SnO 2 composite lm, we obtain remarkable development of ethanol sensors, and the sensor response is up to 16.8 for 0.5 ppm ethanol gas, which is a prominent result in terms of ethanol detection at room temperature.

Experimental methods
Step 1: ordered nanoporous SnO 2 sensing lm fabrication. Experimental procedures are described in Fig. 1. The Sn precursor solution was prepared by dissolving 100 mg SnCl 4 (Aldrich, >99%) in 0.016 mol ethanol (>99.5%) in the presence of 4.0 Â 10 À6 mol Pluronic F127 triblock copolymer (Sigma, BioReagent). 0.1 ml 12 M HCl (37 wt%) was then mixed with the above tin precursor solution under constant stirring for overnight. The resulting precursor solution was stable over one month period. The molar ratio of Sn : F127 : EtOH : HCl : H 2 O in the nal Sn precursor solution was 1 : 0.01 : 40 : 3 : 10. The thin lms were prepared by spin-coating 50 ml of the Sn precursor solution onto one sensor device of ca. 10 Â 10 mm 2 with interdigital electrode at 4000 rpm for 30 s under 30% relative humidity for ve times, and then dried at 60 C for 2 h. The lms were then exposed to a water vapour hydrothermal treatment, 80% relative humidity, at 100 C for 96 h. The relative humidity was achieved using a supersaturated salt aqueous solution, kept at 100 C, in a humidity controlled chamber. The lms were then heat treated by annealing at temperature of 300 C for 2 h with an up/down ramp rate of 1.0 C min À1 . In Fig. 1, the step 1 shows the ordered nanoporous SnO 2 thin lm on the sensor device.
Step 2: C 60 /SnO 2 sensing lm fabrication. The C 60 solution was prepared by adding 5 mg C 60 (Aldrich, >99.5%) in 0.016 mol ethanol (>99.5%) in the presence of 3.6 Â 10 À6 mol Pluronic F127 triblock copolymer (Sigma, BioReagent), and then dropped in 50 ml 12 M HCl (37 wt%) with the aid of ultrasound for 5 minutes. The mixture was then under constant stirring for overnight. Composite lms were synthesized by spin-coating 40 ml of the C 60 precursor solution on one above sensor device with ordered nanoporous SnO 2 thin lm ( Fig. 1 step 1) at 4000 rpm for 60 seconds under 30% relative humidity for two times, and then dried at 60 C for one hour. The nanocomposite thin lms were then oxygen-plasma treated (Femto, 40 kHz, 100 W) for 4 minutes.
Step 3: Au@SnO 2 /C 60 /SnO 2 sensing lm fabrication. Au@SnO 2 /C 60 /SnO 2 sensing thin lm was fabricated through a simple route. The Au-Sn precursor solutions were prepared by dissolving 100 mg SnCl 4 (Aldrich, >99%) in 0.016 mol ethanol (>99.5%) in the presence of 3.6 Â 10 À6 mol Pluronic F127 triblock copolymer (Sigma, BioReagent). 0.5 ml 0.2 mg ml À1 colloidal gold solution (XFNANO) was dropped in 50 ml 12 M HCl (37 wt%) with the aid of ultrasound for 5 minutes, then mixed with the above Sn precursor solution under constant stirring for overnight. Au@SnO 2 /C 60 /SnO 2 sensing thin lms were prepared by spincoating 40 ml of the Au-Sn precursor solution onto one sensor device with nanocomposite lm ( Fig. 1 step 2) at 4000 rpm for 60 s under 30% relative humidity for four times, and then dried at 60 C for 1 h. The sensor device exposed to a water vapour hydrothermal treatment, 95% relative humidity, at 120 C for 96 h. The nal nanocomposite thin lms were then oxygenplasma treated (Femto, 40 kHz, 100 W) for 6 minutes.

Materials characterization
WAXRD data were obtained by a Bruker D8 Advance X-ray diffractometer with Cu Ka (0.15406 nm) radiation. High resolution transmission electron microscopy (HRTEM) and scanning transmission electron microscopy in high angle annular dark eld mode (STEM-HAADF) were performed using a FEI Titan 80-300 equipped with a eld emission gun operated at 300 kV; lm parts were scratched from the substrate and collected on an amorphous holey carbon lm on a copper grid. Raman spectra were recorded with a LabRAM HR UV-vis (Horiba Jobin Yvon) Raman microscope (Olympus BX41) with a Symphony CCD detection system using a HeNe laser at 632.8 nm. The spectra were taken from material removed from the substrate. The Fourier transform infrared spectra (FTIR) of the samples were recorded on an AVATAR370FT-IR spectrophotometer using conventional KBr pellets.

Sensing measurement
During the sensing measurement, liquid VOCs inputted through a sample inlet and led down to a heater, which vaporized it. A fan ensured that the vapour was homogeneously distributed. The gas sensing properties were determined in a sample cell consists of a sample chamber and has a gas inlet and outlet. 22 A certain concentration of VOCs gas or pure air is periodically passed into the test chamber based on computercontrolled mass ow controllers (MFCs), and the total ow rate is maintained at 1000 sccm. Resistance changes upon sample exposure to gases recorded by a high resistance meter Keithley 6517B. The sensor response is dened as R a /R g , where R a and R g are the sensor resistances in air and in the target gas, respectively. This parameter is positive for n-type VOCs sensing behaviour. Here, the response or recovery time dened as the time taken for the sensor to achieve 90% of its maximum response or decreases to 10% of its maximum response, respectively.

Results and discussion
Scanning transmission electron microscopy (STEM) image ( Fig. 2a) depicts the structure and morphology of the resulting nanoporous Au@SnO 2 /C 60 /SnO 2 sensing lm. It can be seen that there are bigger nanopores with the average diameter of 300 nm exhibiting on the surface of sensing composite lm. The pore walls with an average thickness of 160 nm are made up of nanoparticle. The presence of C 60 and Au nanoparticles in the pore wall is further conrmed by the high resolution TEM image (Fig. 2b) showing the region enclosed within the square in upper right of Fig. 2a. The image shows that Au nanoparticles have uniform particle size, and small size of SnO 2 nanoparticles distributed among Au nanoparticles. The high magnication HRTEM image shows the lattice spacing of 0.33 nm and 0.24 nm, corresponding to the (110) plane of cassiterite phase and (111) plane of gold, respectively. Under these bigger nanopores, the high magnication TEM image (Fig. 2c) of selected region in the lower le of Fig. 2a shows that smaller nanopore arrays are packed in an ordered arrangement with a repeating distance. A fast Fourier transformation of the smaller nanoporous SnO 2 layer depicted in the insert image of Fig. 2c, indicates orthorhombic symmetry of a [111] oriented Fmmm nanostructure. 23 It is proved that this sensing lm is composed of Au@SnO 2 /C 60 layer covering the surface of ordered nanoporous SnO 2 layer. The elemental mappings of the nanocomposite lm are obtained by energy-dispersive spectroscopy (EDS) for Sn, C, and Au as shown in Fig. S1. † Clearly, C and Au are evenly distributed throughout the Au@SnO 2 /C 60 layer. The STEM image (Fig. 2d) illuminates the cross-section structure of Au@SnO 2 /C 60 /SnO 2 sensing lm. Further EDX line measurement taken from the selected area (marked by a red line in Fig. 2d) exhibits the presence of C 60 layer in the composite lm, indicating the formation of SnO 2 /C 60 /SnO 2 heterojunctions. Notably, the unique electron coupling at this SnO 2 /C 60 /SnO 2 interface is able to effectively facilitate the electron transfer from the sensing lm to integrated electrode. 24 As shown in Fig. 3a, Au@SnO 2 /C 60 layer is directly grown on ordered nanoporous SnO 2 thin lm using post-synthetic hydrothermal treatment and then oxygen-plasma treated. For further investigating the composite sensing lm, the structure and crystallinity of the synthesized thin lm are conrmed by Xray diffraction, FTIR, and Raman studies.  indicating the successful formation of cassiterite SnO 2 . The mean crystallite size of the synthesized SnO 2 nanoparticles, which is calculated using Scherrer's formula: D s ¼ 0.9 Â l/(b Â cos q), where D s is the crystallite size, l is the wavelength of Xrays (l Cu ¼ 0.15418 nm), b is the full width at half maximum (FWHM) of the diffraction peak, and q is the Bragg diffraction angle of the XRD peak. The mean crystallite size is 3.1 nm, which is in good agreement with the HRTEM studies (Fig. 2b). Fig. 3c shows the IR spectra in the range of 400 to 3400 cm À1 for IR-active fundamental vibrations of C 60 (a), SnO 2 (b), and Au@SnO 2 /C 60 /SnO 2 nanocomposites (c). The peak at 1631 cm À1 can be attributed to O-H bending vibrations. For Au@SnO 2 /C 60 / SnO 2 nanocomposite lms (c), the bands at 527, 577, 1186 and 1421 cm À1 are attributed to the internal modes of the C 60 molecule. Two new peaks display at 1659 and 1720 cm À1 for oxygen-plasma treated C 60 , resulting from the stretching vibration of carboxyl groups formed during the treatment. It should be mentioned that these surface functional groups provide active sites for connection with SnO 2 nanocrystals. The IR-active fundamental vibrations of SnO 2 are located in the range of 400-1000 cm À1 , and the intensive broad bands are ascribed to the stretching vibrations of Sn-O bonds at 556 and 632 cm À1 . For pure SnO 2 nanoparticle, a broad peak at 621 cm À1 can be assigned to the vibration of Sn-O. When SnO 2 is modi-ed with C 60 , covalent bonds (e.g., Sn-O-C-O or Sn-O-C) formed due to interaction between oxygen plasma-treated C 60 and SnO 2 . As a result, the main peak associated with SnO 2 slightly red-shi to low frequency, especially for the band at 632 cm À1 . The IR results further conrm the chemically bonded interaction between C 60 and SnO 2 nanoparticles. Fig. 3d shows the Raman spectra of SnO 2 (a), the synthesized composite lms (b), C 60 (c). For the synthesized composite lm (b), the Raman shis of C 60 are observed at 265 cm À1 , 491 cm À1 , and 1457 cm À1 , which can be associated with H 1g squashing, A 1g breathing, A 2g pentagonal pinch modes of C 60 molecules, respectively. The two strong peaks located at 1380 cm À1 (k-point phonons of A 1g symmetry, D-band) and 1594 cm À1 (E 2g phonons of C sp 2 atoms, G-band) exhibit that partial C 60 transforms into carbon crystallite in nanocomposite lm. 25 Compared to the pure C 60 (c), changes in the relative intensities of the D and G bands (D/G) indicate the changes of the electronic conjugation state of the C 60 during the fabrication procedure. The weak peak located at 632 cm À1 in nanocomposite lm can be assigned to the A 1g modes of the SnO 2 rutile phase. Besides, the other mode named M 2 (located at 566 cm À1 ) being related to the nanostructure of composite lm is also identied in the spectra. 26 For comparison, we examine the sensing performance of C 60 /SnO 2 lm (a), SnO 2 /C 60 /SnO 2 lm (b), and Au-functionalized SnO 2 /C 60 /SnO 2 lm (c), in terms of the ethanol sensing. The typical response curves of three different SnO 2 based gas sensors being measured at various ethanol concentrations are shown in Fig. 4a. The response, S, is dened as R air /R gas for the reducing gases (ethanol) where R gas and R air denote resistance in the presence and absence of test gases, respectively. The measurements are performed at room temperature with ethanol gas exposure concentration ranging from 0.5 ppm to 50 ppm. A clear increase in the sensor response is observed with increasing gas concentration, and of all the materials tested,  Au@SnO 2 /C 60 /SnO 2 composite lm displays the best ethanolsensing performance, with high sensor response of 16.8.6 to 0.5 ppm ethanol gas at room temperature.
Moreover, the Au@SnO 2 /C 60 /SnO 2 sensor also depicts the response to formaldehyde, acetone, and toluene gases, as shown in Fig. 4b. The 0.5 ppm ethanol response (16.8) becomes more than four times as high as that of formaldehyde. Thus, the Au@SnO 2 /C 60 /SnO 2 sensor reveals not only high response but also high ethanol selectivity at low concentration. The response curves illuminate excellent stability, as shown in Fig. 4c. The response time dened as the time necessary to reach 90% of the maximum response is about 35 s, and the recovery time is about 9 s. In Fig. 4d, the Au@SnO 2 /C 60 /SnO 2 gas sensor (c) shows a power law relationship between S and C EtOH , and the correlation coefficient (n) is estimated to be 0.31 within the gas concentration ranging from 0.5 to 50 ppm (S ¼ a Â C n ). They are determined by n/p/n heterostructure in the nanoporous structure which strongly inuences carrier mobility in the sensing lm. The high sensitivity of Au@SnO 2 /C 60 /SnO 2 sensor may be due to the high reactivity of ethanol with adsorbed oxygen in the presence of Au-sites on the SnO 2 /C 60 /SnO 2 surface. The quick response and recovery would also be attributed to the nanoporous structure and n/p/n heterostructure of the composite lm. Fig. S2 † exhibits the long-term stability of the Au@SnO 2 / C 60 /SnO 2 sensor exposed to 0.5 ppm ethanol gas. The stability of the sensor is measured over different times on stream. The average and relative standard deviation of the response (S) of the ethanol gas sensor at a testing concentration of 0.5 ppm of ethanol for 48 days are 16.8 and 1.1%, respectively.
The generation of n/p/n heterojunctions (Fig. 5a) can inevitably develop the depletion regions of both sides of junctions, decreasing the conduction volume of SnO 2 nanoparticles. The introduction/removal of ethanol gas will induce the more signicant change of conduction volume in case of SnO 2 /C 60 / SnO 2 composite lms. This will bring about the larger change in resistance, enhancing the sensitivity. Fig. 5b schematically depicts the changes of the electronic energy bands for C 60 /SnO 2 material before and aer the adsorption of ethanol gases. This gure shows two depletion layers-one is on the surface of the SnO 2 nanoparticles, and the other is in the interface between C 60 and SnO 2 nanoparticles. In the n/p/n heterojunctions, the work function of n-SnO 2 and C 60 are 4.55 and 4.72 eV, 27 respectively. Due to the charge transfer, the potential barrier (0.17 eV) will be developed at heterojunctions, being accompanied by the bending of the vacuum energy level and the energy band. The local p/n heterojunctions between C 60 and SnO 2 nanoparticles will provide additional modulation of resistance during the introduction and removal of ethanol gas by changing the heterojunction barrier.
In case of Au, Au being a better oxygen dissociation catalyst than SnO 2 enhances the rate of dissociation and diffusion of oxygen species on the surface of SnO 2 , resulting in a greater degree of electron withdrawal from the conduction band of SnO 2 ("spill-over" effect). In the Au/SnO 2 heterojunctions (Fig. 5c), the work function of Au (5.1 eV) is larger than that of SnO 2 (4.55 eV). In order to align the Fermi level, the electrons will ow from SnO 2 to Au sides, as the electrostatic potential of the semiconductor is increased. Being similar to the C 60 -SnO 2 heterojunctions, the Schottky barriers (0.55 eV) generated at Au-SnO 2 heterojunctions will provide additional modulation of resistance during the adsorption and desorption of gas molecules by changing the heterojunction barrier. In equilibrium and in the absence of ethanol gas, because the Au work function is large, a depletion layer will be developed in the SnO 2 nanoparticle. For Au with high work function, this depletion region of electrons is much more pronounced, which gives rise to a larger barrier height. This manifests a larger electrical resistance of the sensing lm in air, as shown in Fig. 5d.
From Fig. 5b, we can see that Au@SnO 2 /C 60 /SnO 2 nanocomposite sensor shows far greater sensitivity to ethanol gas than other gases, manifesting the excellent selectivity to ethanol. On the one hand, it is reported that the O-H band of ethanol is important for the adsorption onto the SnO 2 (110) surface. And the SnO 2 (110) surface is highly sensitive to ethanol due to its powerful attraction to the polar O-H bond. Even more, quantum size SnO 2 nanoparticles (3.1 nm) can provide a catalytic effect to decrease the activation energy of ethanol to oxidize and decompose by rst weakening the O-H bond of the ethanol molecule at room temperature. 28 On the other hand, the Aufunctionalized SnO 2 /C 60 /SnO 2 sensor exhibits an extreme selectivity to ethanol, in comparison to other gases such as toluene. Adsorbed phenyl groups are stable on the surface until 165 K, but the chemisorbed phenyl group will be desorbed from the Au surface at room temperature. With the reasonable assumption that the stable adsorption is the primarily crucial step for subsequent catalytic reactions, Au NPs will provide both high sensitivity and selectivity in sensing ethanol in comparison to toluene at room temperature. 29 Up to the present, signicant amount of research work has been reported on the sensing of ethanol gas (Table S1 †). First,   35 Li et al. utilized Au-decorated SnO 2 nanoparticles for obtaining the higher response of 0.6 at a low concentration of 0.5 ppm. 36 However, the grain size was more than 30 nm, which hindered the further improvement in sensing performance. In this work, the gas sensor exhibits ultrahigh sensitivity to ethanol gas with response value up to 16.8 to 0.5 ppm at room temperature.

Conclusions
In summary, we employ not only the Au-functionalization but also the C 60 -functionalization to synthesize hierarchically nanoporous composite lms for enhancing the sensing performance. The fabricated hierarchically nanoporous Au@SnO 2 /C 60 /SnO 2 nanocomposite are obtained through water vapour hydrothermal treatment and layer by layer selfassembly. Their hierarchical nanopore structure is bimodal with closely packed open macropores ($300 nm) at Au@SnO 2 / C 60 layer and with highly ordered nanoporous with a narrow size range of nanopores mainly from 8-10 nm at the bottom SnO 2 layer. The structure, crystallinity and composition of the hierarchical structure Au@SnO 2 /C 60 /SnO 2 nanocomposite lms are characterized by XRD, Raman, HRTEM, and SAED. Interestingly, the Au@SnO 2 /C 60 /SnO 2 nanocomposite lms are treated with oxygen-plasma treatment to remove the organic template and modify the sensing activity. The Au@SnO 2 /C 60 / SnO 2 nanocomposite sensor demonstrates the ultrasensitive ethanol sensing behaviour at room temperature. The sensing properties of Au@SnO 2 /C 60 /SnO 2 nanocomposite sensors are investigated by exposing it to ethanol gas with a concentration range from 0.5 ppm to 50 ppm. Furthermore, these gas sensors exhibit ultrahigh sensitivity to ethanol gas with response value up to 16.8 to 0.5 ppm and a short recovery time of 9 second at room temperature. The synthesis method of combining the hierarchical nanoporous structure with layer by layer selfassembly is a general technique, and offers a promising strategy for preparing high performance nanoporous crystalline materials for applications including gas sensing, photocatalysis, and 3 rd generation photovoltaics.

Conflicts of interest
There are no conicts to declare.