Synthesis and Ethanol Sensing Properties of Novel Hierarchical Sn 3 O 4 Nanoflowers

Due to the metastable property and arduous preparation, to control the size and shape of intermediate Sn 3 O 4 nanocrystals to tune functional properties still poses great challenge, and the physical and chemical properties are not fully investigated. Here, we report a simple one-pot template-free hydrothermal route to fabricate Sn 3 O 4 flower-like hierarchical structures self-assembled by aligned high-density nanoslices. In order to explore the growthmechanism, a series of sampleswith various hydrothermal timewere prepared and examined by FESEM and Raman. Results show that the hydrothermal time influences the phases and morphology of the final products. Particularly, a sensor based on these Sn 3 O 4 was implemented to investigate the potential of Sn 3 O 4 for the ethanol detection, revealing that this material reacts to ethanol in a linear way with high response yet at lower temperature (190C) than that of the well-known SnO 2 . Also, this intermediate tin oxide with rational control over dimension andmorphology provides new opportunities for practical applications in gas sensing towards other reducing gases.


Introduction
SnO  , with coordination structures including stoichiometric compounds SnO 2 and SnO, has drawn particular attention of researchers due to its excellent properties and great potential applications, such as lithium-ion battery, solar cell, photocatalyst, and gas sensor [1][2][3][4][5].In addition, some nonstoichiometric compounds such as Sn 2 O 3 , Sn 3 O 4 , and Sn 5 O 6 , containing both the stannous (II) ion and the stannic (IV) ion, were predicted by theory [6].In the same manner as other mixed-valence metal oxides such as Fe 3 O 4 and Co 3 O 4 , these mixed-valence tin oxides might have a unique property and application in catalyst and lithium-ion battery [7,8].Among these multivalence tin oxides, Sn 3 O 4 was first discovered by Lawson in 1967 [9].In 2010, Sn 3 O 4 was synthesized through the use of a carbothermal reduction method requiring a high temperature at 1,200 ∘ C and nitrogen atmosphere, which is difficult to implement [10].For nearly two years, Sn 3 O 4 has inspired new interests of researchers due to its promising properties in catalysis and gas sensor.In 2014, hierarchical Sn 3 O 4 has been synthesized for the first time by a hydrothermal method and degraded all methyl orange within twenty minutes of being illuminated by visible light [11].Manikandan et al. found that the H 2 evolution rate of Sn 3 O 4 was greater than 5 molh −1 , which is competitive to the performance of reported visible-light-sensitive photocatalysts [12].In 2015, researchers have demonstrated that the Sn 3 O 4 nanobelt exhibits the ultrahigh sensitivity towards NO 2 at 200 ∘ C, over twice that of SnO 2 [13].However, few reports have been focused on the study of Sn 3 O 4 , especially on the hydrothermal time's effects on the final product, due to its metastable property and arduous preparation.
Hydrothermal synthesis, due to its many advantages, such as short processing time, low growth temperature, easy control over experimental parameters, and convenience for large-scale synthesis, has been used to grow various advanced nanomaterials.In this study, we provided a simple, yet efficient, approach to prepare novel Sn 3 O 4 nanoflowers consisting of nanoslices with a great crystallinity by facile hydrothermal synthesis mean.To propose the possible growth mechanism, we further investigated how the hydrothermal time influenced the final products, which

Sample Preparation.
In a typical preparation, 0.09 g SDS was melted in 20 mL NaOH (0.8 M) solution.Then, 20 mL SnCl 2 (0.2 M) solution was added into the above solution and stirred for 30 minutes.All the chemicals used were analytic pure without further purification.The acquired solution was transferred to a 50 mL Teflon-lined stainless autoclave, which was subsequently put into the electric oven, at the maintained temperature of 160 ∘ C, for 24 hours.After cooling down to the room temperature naturally, the yellow precipitate was collected and washed several times with distilled water and dried at 70 ∘ C for 12 hours.

Characterization.
The morphology of the samples was analyzed by FESEM with an energy dispersive X-ray (EDX) spectroscopy (Field Emission Scanning Electron Microscopy, Hitachi S-3500) and TEM (Transmission Electron Microscopy, JEM-2010 200 KeV).The component, structure, crystallinity, and phase purity of the samples were examined by X-ray diffraction (X'Pert MPD), HRTEM (High Resolution Transmission Electron Microscopy), SAED (Selected Area Electron Diffraction), and Raman Spectroscopy (Horiba Jobin Yvon HR 800 Raman spectrometer, France).The excitation wavelength was 532 nm.

Fabrication of Gas
Sensor.We fabricated the gas sensors by uniformly dispersing the mixture of powders and ethanol on the surface of a ceramic tube with a pair of previously printed gold electrodes that were connected by four platinum wires.Ni-Cr heating wire was inserted into the tube to manipulate the operating temperature by tuning the heating voltage.Figure 1 displays the schematic image of the asprepared gas sensor.Next, the ceramic tube was fixed by wielding the ends of the six wires to the terminals of the pedestal and was aging inside the electric oven at 250 ∘ C, for 24 hours, before it was put into the chamber of the computercontrolled gas sensing machine (WS-30A gas sensing system, Weisheng Electronics Co., Ltd.).The environment relative humidity is 49% and ambient temperature is 19 ∘ C. The sensitivity of the sample is defined as   /  , where   and   stand for the resistance of sensor in the air and in the ethanol, respectively.The response and recovery times are defined as the time taken by the sensor reaching 90% of the equilibrium resistance.the nanoflowers are self-assembled by well-defined thin nanoslices with sharp edges and smooth surfaces and the size of each nanoslice is about 5 m.

Results and Discussion
The morphology was further verified by TEM (inset of Figures 2(b) and 2(c)).The inserted image in Figure 2 is a low-magnification TEM image, which further confirms that the nanoflowers are self-assembled by nanoslices being consistent with the result of FESEM.The thickness of a single slice that is approximately 8-15 nm was acquired by the highmagnification TEM image (Figure 2(c)).Further structural analysis of the Sn 3 O 4 nanoslices was carried out by HRTEM (Figure 2(d)).From the HRTEM image, the measured fringe spacing of the nanoslice is 0.28 nm, which is very consistent with the -value of 0.283 nm corresponding to (−210) crystal graphic planes of triclinic Sn 3 O 4 (PDF 16-0737).Figure 2(d) is SAED patterns of a single sheet, clearly revealing that the nanoslices are monocrystal.
The structure of the as-acquired sample was measured by powder XRD as shown in Figure 3.All peak positions and intensities are consistent with the triclinic crystal structure Sn 3 O 4 (PDF 16-0737), except for two weak peaks marked by ( * ) and (I).These weak peaks reveal that the phase mainly consists of Sn 3 O 4 .The peak marked by ( * ) is consistent with SnO (PDF 06-0395).The EDX spectrum (inset of Figure 3) confirms that the sample only consists of two elements, tin and oxygen (Cu is from substrate), which excludes the possibility that the peak (I) is caused by other elements.Because, theoretically, the structure of SnO  reflects the arrangements of oxygen atoms and vacancies on the oxygen sublattice of the rutile SnO 2 , the peak (I) is possible to be caused by the periodic oxygen defects near the surface or some new phase of SnO  in the Sn 3 O 4 structure, which agrees with the previous work [6,14].As a result, considering the uniform morphology of the as-prepared Sn 3 O 4 , we deduce that those weak peaks of XRD may be evidence of the heterojunctions of tin oxides.
Figure 4 was a Raman spectrum at room temperature, which provides structural information of the as-prepared sample.The two sharp peaks, centered at 143 and 172 cm −1 , respectively, are two characteristic phonon modes of Sn 3 O 4 and other typical Raman modes of other tin oxides, mainly SnO (113 and 211 cm −1 ) and SnO 2 (631 cm −1 ), are absent, further revealing the high chemical and structural purity [10,11,15].The peak at 241 cm −1 also belongs to the phase of Sn 3 O 4 according to the previous work.A shoulder peak at 133 cm −1 appears in most of the Raman spectra of Sn 3 O 4 , which may be another characteristic peak of Sn 3 O 4 [10,11,16].
In order to investigate the growth mechanism of nanoflowers, a series of samples were acquired by altering hydrothermal time from 6 hours to 24 hours.All the samples were identified by the SEM and Raman spectra to determine chemical composition and morphologies, as shown in Figure 5.The SEM images show that, as the hydrothermal time is prolonged, the morphologies of the samples alter from the initially smooth, thick plates, to the porous, rough plates, and, ultimately, to the hierarchical nanoflowers.Also, the structure and chemical constitution of each sample from different hydrothermal stages vary.When the hydrothermal time is shorter than 6 h, the Raman spectrum shows a single peak at 210 cm  testified to be SnO and Sn 3 O 4 by the Raman peaks in 170 cm −1 and 210 cm −1 .Any sample with the hydrothermal time longer than 12 h sample is in the phase of Sn 3 O 4 .Based on those SEM images and structural information of the samples, a possible growth mechanism is proposed and displayed in Figure 5.

Consider the following:
(A) hydrolysis of SnCl2 During the process of hydrothermal synthesis, the SnO comes into being first and, subsequently, the oxidation process begins.Obviously, the oxidation would increase the coordination number of tin atoms, altering the density of the samples.As a result, the microcracks gradually appeared on the surface or at the edge of the thick plates, rendering them coarse.Afterwards, the nonstop growing cracks, caused by the phase transition, result in the big plates breaking into the final flower-like small pieces.
In order to research the potential application of assynthesized samples, a gas sensor was manufactured.It is well recognized that the operating temperature can greatly influence the gas sensor sensitivity.To determine the optimum temperature, the sensitivities of the gas sensor to 100 ppm ethanol at different temperatures ranging from 130 to 250 ∘ C were measured and shown in Figure 6(a).We can find that the sensitivity value reaches its maximum value of 18.75 at 190 ∘ C and then decreases with further increasing the temperature.Generally, the optimum temperature for tin dioxides is about 300 ∘ C [17,18], while the as-prepared Sn 3 O 4 exhibited sensitivity towards ethanol that is similar or even stronger than the reported SnO 2 , but at a much lower optimum temperature 190 ∘ C [18,19].
To further investigate the relationship between the gas concentration and the gas response, we present, in Since the sensor sensitivity was defined as the ratio of the resistance in the air to the resistance in the gas, from Figure 6(d), we can deduce that the resistance of the materials drops, when the concentration of the ethanol increases, which agrees with the well-known gas sensitive behavior of ntype semiconductor materials [13].Consequently, a probable mechanism of Sn 3 O 4 is put forward, based on the representative mechanism of the n-type semiconductor.As we know, the fluctuation of the metal oxide-based sensors is governed absorption and depletion of the gas molecules on the surface of the material.During the process when the sensors are aging at the atmosphere, the surface of the material will absorb O 2 and the absorbed O 2 produces like O − , O 2− , and O 2− , which trap electrons from conduction bands and generate an electron depletion on the surface, leading to high resistance [20,21].When the sensors are exposed to the reducing gas under appropriate operating temperature, for example, the ethanol in this paper, the reactions between the absorbed O 2 and the gas will take place on the surface of the materials, which decrease the concentration of the oxygen ions and thus release the restrained electron to the conduction band of the materials, giving rise to the decrease of the resistance.As observed in the SEM images, the unique 3D hierarchical shape presents a mesoporous structure.This mesoporous structure could provide abundant surface state and defects, as well as an enormous amount of mesopores with a highly active surface area.This large, specific surface area could more readily absorb gas molecules, which leads to the high gas sensor sensitivity and a faster response speed.In addition, the heterojunctions of tin oxides may play a role on its gas sensitive behavior.

Conclusions
To summarize, we successfully synthesize the novel hierarchical Sn 3 O 4 via facile one-pot hydrothermal route and investigate the effects of the hydrothermal time on the phase of the final product, and its gas sensing properties.Experiment data shows that the hydrothermal time is the key factor to control the final chemical composition, especially when the metal element has many valences.In particular, the systematical study on the gas sensing indicates that, compared with traditional monovalence tin oxides, Sn 3 O 4 reacts to the increase of the ethanol increase in a linear manner with a high response yet at a much lower temperature, which shows the great potential of the Sn 3 O 4 for gas sensors towards other reducing gases.

Figures 2 (Figure 1 :
Figures 2(a) and 2(b) depicted the morphology of the as-prepared sample examined by FESEM.The lowmagnification FESEM image (Figure 2(a)) indicates a representative overview of the 3D uniform hierarchical flower-like architecture.Close examination by the highmagnification SEM, shown in Figure 2(b), reveals that

Figure 5 :
Figure 5: Growth mechanism of as-prepared sample; (a) morphology variation in response with operating time and pictures of possible growth mechanism (the red and the green represent the Sn 3 O 4 and SnO, resp.) and (b) the Raman curve of samples at different hydrothermal times 6 h, 12 h, 18 h, and 24 h at room temperature.

Figure 6 :
Figure 6: Gas sensitive response of as-prepared samples; (a) sensitivity variation in response with operating temperature; (b) the sensor response curve of gas sensor; (c) typical transient response curves of the sensors; and (d) dynamic sensing transient of the sensor to 100 ppm for ethanol.
Figure 6(b), the gas response to ethanol with operating temperature of 190 ∘ C. As the gas concentration rises from 50 ppm to 400 ppm, the response of the sensor increases in nearly linear manner, without clues of the saturation, which is necessary to monitor the concentration of the ethanol.The transient response characteristics of the flower-like architecture toward ethanol of different concentrations at 190 ∘ C were investigated and the results are displayed in Figure 6(c).The sensitivity increases dramatically as soon as the concentration of the gas increases, which exhibits high response and recovery property.Figure 6(d) shows the dynamic sensing transient of the sensor at the optimum temperature.By careful measurement, the response time of Sn 3 O 4 and recovery time toward ethanol are approximately 20 and 60 seconds, respectively, which can meet the practical application.