Background

Gas sensors have attracted a widespread attention due to their potential applications in detecting toxic, noxious, flammable, and explosive gas [1]. At present, the metal oxide semiconductors occupy an important position in various sensors due to their simple preparation process, lower cost, and higher sensitivity to the target gases [2,3,4]. Tin dioxide (SnO2), a multifunctional n-type material with a direct band gap of 3.6 eV [5], has been widely used in both fundamental study and practical applications, such as gas sensors [6], catalysis [7], and optoelectronic devices [8]. Especially, SnO2 has been regarded as the most potential sensing material owing to its natural non-stoichiometry [9], high sensitivity, fast response/recovery speed, and high chemical stability [10].

It is well known that the gas-sensing mechanism of metal oxides is related to the adsorption and desorption processes of the target gas on the sensor surface, giving rise to a change of the electrical conductivity [11]. These processes strongly depend on the size, morphology, and dimension as well as crystalline structure of the samples [12]. There are two main ways to effectively enhance the sensing performance of SnO2 [13]. One is to synthesize composed materials based on SnO2, such as fabrication of p-n junctions, surface decoration, or doping [14]. The other is to prepare various pure SnO2 materials including nanotubes [15], nanorods [16], nanospheres [17], hollow structures [14], and nanoflowers [18], which have unique nanostructures, high specific surface area, and strong electron capture abilities [19]. Recently, three-dimensional (3D) hierarchical SnO2 nanostructures have drawn much attention because of their better gas-sensing performance caused by large specific surface area and rapid gaseous diffusion compared with 1D and 2D nanostructures [20]. Various techniques have been used to fabricate 3D nanostructures of SnO2 [21], such as chemical vapor deposition [22], solvothermal synthetic method [23], template method [24], sol–gel method [25], and hydrothermal route [26]. Among them, solvothermal and hydrothermal routes with low cost [27], high yields, and simple manipulation have been proved to be the promising methods to synthesize 3D hierarchical SnO2 nanostructures. For instance, Dong et al. prepared hollow SnO2 nanospheres with the diameter ranging from 200 to 400 nm using a solvothermal synthetic method [28]. Li et al. fabricated a novel snowflake-like SnO2 hierarchical architecture with excellent gas-sensing properties via a facile hydrothermal method [29]. Moreover, Chen et al. successfully synthesized hierarchical flower-like SnO2 blooming nanoflowers constructed by self-assembly of many regular-shaped nanosheets through conventional hydrothermal method [30].

The practical application of SnO2 sensors is still limited to a certain extent due to the relative higher working temperature and poorer selectivity to test gases [31]. In order to improve the gas-sensing properties, researchers have paid attention to controllable synthesis of 3D flower-like SnO2 nanostructures with surfactant effects [32], yet a significant challenge is posed because of the variety of surfactants.

In the present study, we report a well-controlled optimization of 3D hierarchical SnO2 nanoflowers based on self-assembly of thin nanosheets with the help of different surfactants under hydrothermal condition. Our systematically comparative gas-sensing study between fabricated sensors focuses on the promotion effect of surfactants on sensor behaviors. The results show that the amphiphilic non-ionic surfactants, such as PVP and Triton X-100, can be potential candidates to optimize the morphology of 3D nanoflowers with high porosity and large specific surface area. Particularly, the sensor based on PVP exhibits high response, fast response time, and good selectivity to H2S at a relative lower temperature. In addition, a possible well-controlled growth mechanism of SnO2 nanostructures is proposed.

Methods/Experimental

Trisodium citrate dihydrate and tin chloride dihydrate from Sinopharm Chemical Reagent Co., Ltd. were used as precursors for SnO2 synthesis. Polyethyleneimine, hexamethylene tetramine, TritonX-100, and polyvinylpyrrolidone were purchased from Aldrich Chemistry and used as the structure-directing agents. Distilled water was used throughout the experiments. All chemicals were of analytical grade and used as-purchased without any further purification.

Synthesis of SnO2 Nanoflowers with Different Architectures

A typical synthesis procedure with simple hydrothermal method can be described as follows (Fig. 1): firstly, 5 mmol of NaOH was added into an 80-ml mixture of anhydrous ethanol and deionized water (1:1) under magnetic stirring. Then, 20 mmol Na3C6H5O7·2H2O and 10 mmol SnCl2·2H2O were dissolved into the mixed solution successively under vigorous stirring for 1 h at room temperature. The mixed solution was then transferred into a 100-mL Teflon-lined stainless steel autoclave, and maintained at 180 °C for 12 h, and then cooled down to room temperature naturally. After reaction, the obtained precipitate was collected by centrifugation, washing with deionized water and anhydrous ethanol for several times, and dried at 60 °C for 6 h. The SnO2 nanoflowers were finally obtained after calcining the precipitate in a muffle furnace under air ambient condition at 500 °C for 2 h. In order to synthesize SnO2 nanoflowers with different microstructures, different surface active agents (1.0 g) were introduced into the solution respectively before the dissolution of Na3C6H5O7·2H2O. In this work, four different kinds of surfactants were used, including PVP, PEI, HMT, and TritonX-100, and the corresponding final products are named as SPVP, SPEI, SHMT, and STritonX-100, respectively, while the product without surfactant is signed as S0.

Fig. 1
figure 1

Schematic illustration of the formation process of the hierarchical flower-like SnO2 nanostructures using different kinds of surfactants

Characterizations

It is well known that the gas-sensing properties of gas sensors are highly related to the morphology, size, and dispersibility of nanomaterials. The as-prepared products were analyzed in terms of their structures and morphologies by means of polycrystalline X-ray diffraction (XRD, Germany Bruker AXS D8 Advance), scanning electron microscopy (SEM, USA FEI Sirion 200), and field emission transmission electron microscopy (FETEM, USA Tecnai G2 F20 S-TWIN). The surface area is measured using Elemental Analyzer (USA ASAP 2460) based on the Brunauer-Emmett-Teller (BET) method.

Sensor Fabrication and Gas-Sensing Test

The gas sensor was fabricated using screen printing method on top of an alumina tube (seen in Fig. 2a). Typically, a proper amount of as-prepared powder was firstly mixed with anhydrous ethanol to form a slurry suspension. Subsequently, the slurry suspension was coated onto the alumina tube by a small brush, which is supported by two Au electrodes and four Pt conducting wires. Next, a Ni-Cr heating wire was inserted into the alumina tube to control working temperature by tuning the heating voltage. Finally, the product was aged at 80 °C for 72 h before testing.

Fig. 2
figure 2

a Schematic diagram of the gas sensor configuration. b Electric circuit diagram of sensor device

The gas-sensing properties were measured using chemical gas sensor-4 temperature pressure small (CGS-4TPs) intelligent gas-sensing analysis system (Beijing Elliott Technology Co., Ltd., China) under the lab conditions. Figure 2b displays the typical schematic electrical circuit. Rs is the resistance of sensor and Rl is the load resistance, and a heating voltage (Vh) is used to adjust the working temperature. In the present work, the response of the sensor was defined as S = (Rs − Rg)/Rg, where Rs is the initial resistance and Rg is the resistance after injection of gases. The response and recovery times are defined as the time taken by the sensor to achieve 90% of the total resistance change in the case of adsorption and desorption, respectively.

Results and Discussion

Structural and Morphological Characterization

The crystalline phase of the as-prepared SnO2 products was identified by power X-ray diffraction as shown in Fig. 3. From the XRD pattern, all the observed diffraction peaks can be easily assigned to the tetragonal rutile structure of pure SnO2 with the standard JCPDS file card no. 41-1445, and no other peaks can be identified due to impurities. The sharp peaks indicate the high degree of crystallinity of our SnO2 samples, and no remarkable shift is detected in the diffraction peaks, revealing that the samples are of high purity.

Fig. 3
figure 3

XRD patterns of the SnO2 samples with different morphologies. a S0, b STritonX100, c SHMT, d SPEI, and e SPVP

Figure 4a shows the SEM image of the product without surfactant. A hierarchical flower-like architecture can be observed and the unique nanoflowers are assembled by ultrathin nanosheets with an average thickness of 20 nm around. Unfortunately, these nanosheets are closely staggered with respect to each other, which results in a sharp decrease in its reaction spaces. Figure 4be shows the morphologies of the products obtained by introducing different surface active agents while keeping other experimental conditions unchanged. One can see that upon addition of TritonX-100 surfactant (Fig. 4b), the nanosheets are loosely intersected with each other, and some mesopores are shaped at the edge of the nanosheets. When HMT was added into the reaction mixture as the surface active agents (Fig. 4c), it can be seen that the nanosheets are arranged randomly and a number of smaller nanosheets are formed between ultrathin nanosheets. Figure 4d shows the SEM images of the products obtained by introducing PEI surfactant in the precursor solution, which reveals that the nanosheets with smooth surfaces are arranged orderly and are vertically intersected with each other, leaving a larger reaction space. Figure 4e, f presents the typical SEM images of the products obtained upon addition of PVP surfactant under the same conditions. One can see that the nanosheets are uniformly distributed along the radius across the whole sample to form a flower-like structure. Moreover, compared with other structures of STritonX-100, SHMT, and SPEI, the nanosheets of SPVP are enclosed into an inverted triangle cone with a relative larger hollow space (Fig. 4e). Further enlarged image reveals that the flower-like architectures are assembled by mesoporous nanosheets to form an open porous hierarchical structure, and each nanosheet was fabricated with numerous mesopores (Fig. 4f).

Fig. 4
figure 4

The SEM images of the SnO2 nanoflowers with different morphologies. a S0, b STritonX-100, c SHMT, d SPEI, and e, f SPVP

In order to further investigate the microstructures and crystalline properties of the nanoflowers, low-magnification TEM and typical HRTEM combined with the selected area electron diffraction (SAED) analysis techniques are employed. From TEM images (Fig. 5ae), it can be seen that the nanoflowers with an average diameter of 3 μm are assembled of numerous individual nanosheets, and its morphology and size are similar to the SEM images. Especially, the TEM image of SPVP (Fig. 5e) shows that the most flower-like structure with the uniform dark color in the middle region is constructed from the well dispersion of numerous uniform nanosheets along the radius direction. Combining the SEM with TEM measurements, a conclusion can be drawn that the structures obtained upon the addition of PVP surfactant are the most stable. The high-resolution TEM (HRTEM) images show that for samples S0, SHMT, SPEI, and SPVP, the observed 0.335-nm lattice spacing is consistent with the (110) lattice plane of tetragonal rutile SnO2 (Fig. 5f only shows a typical HRTEM image for SHMT as a representative). The exposure of (110) lattice plane reveals that (110) lattice plane is the most stable plane for SnO2 in air, which is consistent with the theoretical study. It should be noted that STritonX-100 is a special case in this work (Fig. 5b). Upon addition of TritonX-100 surfactant, the growing and the dispersion of nanosheets are randomly leading to a relative larger diameter (3~4 μm ) of the nanoflowers compared with other samples. Moreover, its HRTEM image shows that the calculated lattice spacing is 0.264 nm, which is corresponding to (101) lattice plane of tetragonal rutile structure of SnO2. Furthermore, SAED pattern reveals that SPVP has a nearly perfect single crystalline structure, and the diffraction spots correspond to (110),(\( 1\;\overline{1}\;0 \)), (\( \overline{1}\;1\;0 \)), and (200) lattice planes of SnO2 (Fig. 5h). Contrarily, for other samples such as S0, SHMT, SPEI, and STritonX-100, the SAED pattern shows a polycrystalline structure, and the diffraction ring is indexed to the (110), (101), and (211) planes of tetragonal rutile structure of SnO2 (Fig. 5g).

Fig. 5
figure 5

Low-magnification TEM images of SnO2 samples. a S0, b STritonX-100, c SHMT, d SPEI, and e SPVP. f HRTEM micrograph showing enlarged lattice images of SHMT. g SAED patterns of S0. h SAED patterns of SPVP

Growth Mechanism of the SnO2 Nanoflowers

Based on the experimental observations and analysis above, it is believed that the surfactants play a significant role in the formation of various SnO2 nanoflowers [33]. The possible growth mechanism of hierarchical sheet-flower SnO2 nanostructures is briefly illustrated in Fig. 1. In this work, all SnO2 nanoflowers are synthesized using SnCl2 as precursor [34]. Under the hydrothermal conditions, the overall reaction for the growth of SnO2 crystals with high temperature and pressure can be expressed as follows [35]:

$$ {\mathrm{SnCl}}_2+2{\mathrm{OH}}^{-}\to \mathrm{Sn}{\left(\mathrm{OH}\right)}_2+2{\mathrm{Cl}}^{-} $$
(1)
$$ \mathrm{Sn}{\left(\mathrm{OH}\right)}_2\to \mathrm{Sn}\mathrm{O}+{\mathrm{H}}_2\mathrm{O} $$
(2)
$$ \mathrm{SnO}+\frac{1}{2}{\mathrm{O}}_2\to {\mathrm{SnO}}_2 $$
(3)
$$ \mathrm{Sn}{\left(\mathrm{OH}\right)}_2+\frac{1}{2}{\mathrm{O}}_2+{\mathrm{H}}_2\mathrm{O}\to \mathrm{Sn}{\left(\mathrm{OH}\right)}_4\to {\mathrm{SnO}}_2+2{\mathrm{H}}_2\mathrm{O} $$
(4)

During the whole process, three chemicals greatly affect the morphology growth of SnO2 nanoflowers, including NaOH, sodium citrate, and the surfactant. Firstly, a number of tiny primary nanocrystals were formed due to the hydrolysis of Sn2+ in a basic ethanol-water solution as well as its rapid reaction with OH ions from NaOH. It should be noted that the basic ethanol-water environment is significant for stimulating the SnO2 nucleation and growth [36]. The addition of sodium citrate plays a crucial role in the space distribution of precursors due to its strong coordinating ability, which can promote an anisotropy in the fast growth and aggregation of SnO2 nanosheets with the driving force of decreasing surface energy and accelerate the assembling of nanosheets into the stable hierarchical blooming nanoflowers [37].

Generally, the addition of surfactants is favorable for the enlargement of the surface area as well as the enhancement of the surface activity [38]. Among the surfactants used in this work, PEI is one kind of cationic surfactants. When PEI is added into the reaction solutions, due to the existent N+ ions with a hydrophilic tail, PEI will preferentially adsorb on a certain crystal facet, which is conducive to the nucleation of SnO2 nanocrystals as well as the orderly growth of SnO2 nanosheets with a directional selectivity. Both PVP and TritonX-100 are amphiphilic non-ionic surfactants, which can serve as a soft template in the fabrication of mesoporous materials. Let us take PVP for example to explain the growth mechanism of porous structures on the SnO2 nanosheets as follows: when PVP is added into the solution, PVP molecules self-assemble into spherical micelles because of the strong hydrophobic attraction between the straight alkyl tails. Due to its amphiphilicity, the hydrophilic radical will move toward the direction of aqueous solution, and the hydrophobic radical will move in the opposite direction, leading to the formation of inorganic domains around periodically arranged PVP micelles. Then, Sn2+ and OH ions are easily adsorbed on the outer surfaces of these micelles via electrostatic interactions until SnCl2 is oxidized into SnO2 nanosheet, which are followed by the self-assembly of nanosheets into blooming nanoflowers with the help of sodium citrate. Finally, the soft-template PVP micelles are removed during the calcination process, yielding hierarchical SnO2 nanoflowers with mesoporous structures. Although both PVP and Triton X-100 contributed to the formation of porous structures, it should be noted that PVP can also play a role of dispersing agent, which makes the SnO2 nanosheets grow more uniformly and separately owing to the strong interactions and short electrostatic interaction distances between the SnO2 nanosheets and PVP.

Gas-Sensing Properties

As reported previously, the hierarchical flower-like nanostructures were favorable for the absorption and diffusion of probe gases in sensor materials. To shed light on the promotion effect of surfactants and corresponding morphology on sensor behavior, a systematically comparative gas-sensing study between fabricated sensors is performed in this work.

Gas-Sensing Behaviors of Fabricated Sensors to Ethanol

The optimum operating temperature is a key factor for the application of semiconductor oxide gas sensors. Firstly, the responses of sensors to 100 ppm ethanol gas at various operating temperatures from 180 to 360 °C are tested as shown in Fig. 6a. It is clearly observed that all these sensors exhibit a similar gas-sensing behavior, i.e., the response values first increase with a rise of temperature, reach a maximum value at 270 °C, and then decrease gradually with a further increasing of temperature. Therefore, 270 °C can be chosen as the optimized operating temperature for gas-sensing study of all fabricated flower-like SnO2 sensors in our work. The reason for the dependence of the response on the temperature is as follows: When the operating temperature is too low, a relative smaller response value is assigned to the inert response due to chemical activation, while for too high operating temperature, the absorbed gas target molecules can escape from the sensors before reactions, resulting in a poor response as well. In addition, it can be seen from Fig. 6a that of all the five SnO2 sensors based on different surfactants, SPVP shows the highest response to ethanol gas and the largest gas response value (38). The maximum response values of other four sensors are 27 for SPEI, 16 for SHMT, 11 for STritonX-100, and 8 for S0.

Fig. 6
figure 6

a The response of the sensors to 100 ppm ethanol under different operating temperatures (180–360 °C). b The dynamic response curves of the sensors to ethanol with different concentrations (10–150 ppm) at 270 °C. c Response versus time curves of the sensors to 10–200 ppm ethanol consecutively at 270 °C. d Dynamic sensing transient of the sensors to 100 ppm ethanol at 270 °C

Figure 6b shows the response of all SnO2 sensors toward ethanol in the concentration range 10~150 ppm at the optimal work temperature 270 °C. It can be clearly observed that the response of all sensors increases rapidly with the gas concentration below 50 ppm, and this trend becomes smooth from 50 to 150 ppm, tending to be saturated at about 100 ppm. As expected, the surfactants and the induced morphologies can produce a large influence on the gas-sensing of fabricated sensors. Among these fabricated sensors, the SPVP sensor exhibits the best sensing behavior toward ethanol gas, and SPEI comes second. To get a deep insight of gas-sensing mechanism, BET (Brunaure-Emmett-Teller) nitrogen adsorption-desorption is also performed to determine the specific surface areas of these samples, as shown in Table 1. One can see that SPEI has the largest specific surface area (38.4 m2 g−1) with overall majority. It is noteworthy that despite the relative smaller surface area (15.5 m2 g−1), SPVP is the best candidate for ethanol gas sensor due to its perfect flower-like architecture with orderly stacking self-assembly and relative higher porosity, providing more active adsorption sites for ethanol molecules. Even at low ethanol concentration as 10 ppm, the sensitivity of S0, STritonX-100, SHMT, SPEI, and SPVP sensors can reach 2, 4, 7, 9, and 11, respectively, indicating their potential application for ethanol sensors even at low concentrations.

Table 1 The BET for the prepared samples

Figure 6c displays dynamic gas-sensing response and recovery curves of fabricated sensors toward ethanol with an operating temperature of 270 °C, from which one can see that the responses of all fabricated sensors increase with increasing the ethanol concentration, and a remarkable modulation of resistance is achieved at about 100 ppm. The responses show a drastic rise once the sensor was exposed to target gases and then dropped to its initial value in air. As shown in Fig. 6d, the response and recovery time to 100 ppm ethanol are about 16 s and 28 s for S0, 14 s and 18 s for STritonX-100, 11 s and 15 s for SHMT, 9 s and 11 s for SPEI, and 5 s and 8 s for SPVP, respectively. It is obvious that the SPVP sensor has the best response/recovery characteristics compared with other sensors.

Table 2 shows a comparison of ethanol-sensing performances based on different SnO2 fabricated approaches reported in other literatures and this work at the concentration of 100 ppm. One can see that our polyporous SnO2 nanoflower presents remarkably ethanol-sensing behaviors with lower optimal operating temperature and higher response value as well as faster response-recovery time, which could be attributed to the presence of numerous mesopores in SPVP sensor, leading to a high porosity in favor of the adsorption and diffusion of ethanol gas.

Table 2 The comparison of ethanol sensing performances of SnO2 materials based on different fabricated approaches reported in other literatures and this work

Gas-Sensing Behaviors of Fabricated Sensors to H2S

As discussed in the previous subsection, SPVP sensor exhibits the best gas-sensing property to 100 ppm ethanol owing to its high porosity. In order to find out its optimum detecting gas, we test the response of SPVP sensor toward different gases, including acetone, methanol, formaldehyde, and H2S, with a concentration of 100 ppm at various operating temperatures (as shown in Fig. 7a, b). It can be noted that the optimal response appears at 330 °C of methanol, at 210 °C of formaldehyde, at 360 °C of acetone, and at 180 °C of H2S. Furthermore, the maximum response value of SPVP to H2S is estimated to be 368, which is one or two orders of magnitude (\( {\mathrm{S}}_{{\mathrm{H}}_2\mathrm{S}}/{\mathrm{S}}_{\mathrm{ethanol}}=9 \), \( {\mathrm{S}}_{{\mathrm{H}}_2\mathrm{S}}/{\mathrm{S}}_{\mathrm{formaldehyde}}=45 \)) higher than that to other test gases. The lowest optimal work temperature as well as the best response value indicates SPVP has the excellent selectivity to H2S.

Fig. 7
figure 7

a Response of SPVP sensor to 100 ppm ethanol, acetone, methanol, and formaldehyde at different operating temperatures. b Response of SPVP sensor to 100 ppm H2S at different operating temperatures. c Response versus time curves of the sensors to 10–200 ppm H2S consecutively at180 °C. d Dynamic sensing transient of the sensors to 100 ppm H2S at 180 °C

Considering the high response of SPVP sensor to H2S, we also performed a systematic gas-sensing measurement of all other sensors. A dynamic gas-sensing response and recovery curves of fabricated sensors toward H2S at 180 °C are displayed in Fig. 7c. Obviously, the response values of all fabricated sensors show a monotonic increasing function of H2S concentration. For 100 ppm H2S, the response and recovery time are about 9 s and 43 s for S0, 5 s and 30 s for STritonX-100, 14 s and 40 s for SHMT, 8 s and 38 s for SPEI, and 4 s and 20 s for SPVP, while the maximum response values are 35, 132, 41, 49, and 368 for S0, STritonX-100, SHMT, SPEI, and SPVP, respectively. It is obvious that the SPVP sensor has the best response/recovery characteristics and the highest response to H2S gas compared with other sensors, while STritonX-100 achieves the second.

Figure 8 displays the bar graph of response of five fabricated sensors to formaldehyde, methanol, ethanol, acetone, and H2S. All of the gases were tested with a concentration of 100 ppm at the optimal operating temperature. STritonX-100 and SPVP show a distinct response to H2S, while SPEI shows the highest gas response to methanol and acetone. It should be mentioned that the specific surface area and the porosity are two important factors for gas sensors. The larger specific surface area will provide more active sites for adsorption and desorption of test gases, while the larger porosity would induce a greater speed of gas diffusion owing to the presence of mesopores. In comparison, SPEI possesses a relative larger specific surface area than others (seen in Table 1), which shows the highest gas response to methanol and acetone (Fig. 8), while SPVP and STritonX-100 exhibit the higher gas response to H2S due to their polyporous flower-like nanostructures, proving good selectivity of STritonX-100 and SPVP toward H2S. The good selectivity of the samples to H2S can be explained as follows: when the SnO2 sensor is exposed in H2S gas, both chemisorbed oxygen species and SnO2 nanostructure react with H2S during sensing measurement to form SO2 and SnS2, respectively. Compared with SnO2, the body resistance of SnS2 is relatively smaller, leading to the sensitivity enhancement of the gas sensor [39]. On the contrary, the SnO2 sensor does not react with any other target gases, such as formaldehyde, methanol, ethanol, and acetone.

Fig. 8
figure 8

The comparison among sensor response of SnO2 nanomaterials to 100 ppm of various gases at the optimal operating temperature

Good stability and long service duration are expected from the viewpoint of practical application. To verify the stability of the sensor, the successive gas-sensing behavior of SPVP to 100 ppm ethanol was tested under the same conditions after 1 month. The samples were stored in the vacuum drying vessel during the 1-month interval. It can be seen from Fig. 9 that SPVP exhibits an excellent repeatability and stability even after 1 month. The three cyclic curves are similar to that measured 1 month ago, including the response value as well as the response-recovery time.

Fig. 9
figure 9

Stability of SPVP over 1 month of aging for 100 ppm ethanol at 270 °C

Gas-Sensing Mechanism

Up to now, the most widely accepted gas-sensing mechanism of semiconductor oxide is the model based on the electron transfer dynamics during an adsorption–oxidation–desorption process, which can change the resistance value of the sensors [40]. The response of typical n-type semiconductor greatly depends on the electron concentration. As shown in Fig. 10, at elevated temperature, electrons in the valence band are thermally excited to the conductive band. Once the SnO2 sensor is exposed to ambient air, oxygen molecules will be chemisorbed on the surface of SnO2 nanoflowers. Oxygen ions (O2, O, and O2) are then formed by capturing electrons from the conductive band of SnO2 [41], which is accompanied by an effective enlargement of electron-depleted layer. As a typical n-type semiconductor, the broadening of electron-depleted region means the decrease of carrier concentration within SnO2 nanoflowers, which will lead to the increase of resistance of the sensors. Conversely, when the SnO2 sensor is exposed in the reductive ambient, the absorbed oxygen species will quickly react with the target gas, which results in releasing the trapped electrons back to the conduction band and a reduction of the resistance of the sensors. Among the sensors fabricated in this work, SPEI and SPVP show relative better gas-sensing performances. The underlying physical mechanisms are as follows: the gas sensing properties are strongly dependent on the surface special area and the porosity. In comparison, SPEI possesses a relative larger specific surface area than others, which will provide more active sites for adsorption and desorption of test gases. SPVP exhibits a relative higher porosity due to the polyporous flower-like nanostructures, which is favorable to the rapid diffusion of gas (as shown in Fig. 10).

Fig. 10
figure 10

Schematic diagrams on the gas-sensing mechanism of flower-like SnO2 hierarchical nanostructures

Conclusions

We have successfully prepared the 3D hierarchical flower-like SnO2 nanostructures through a simple and low-cost facile hydrothermal route with the assistance of different surfactants. The images of SEM and TEM showed that the fabricated 3D hierarchical SnO2 nanoflowers with an average diameter of 2~4 μm were composed of many 2D nanosheets. The addition of surfactant plays an important role in the formation of nanoflowers. Based on the experimental observations, the possible growth process and gas-sensing mechanism of SnO2 nanoflowers were proposed. As a cationic surfactant, the addition of PEI is conducive to the nucleation of SnO2 nanocrystals as well as the orderly growth of SnO2 nanosheets, leading to a relative larger specific surface area. As amphiphilic non-ionic surfactants, PVP and TritonX-100 can make the nanosheets grow more uniformly and separately, which can serve as a soft template in the synthesis of advanced material, especially in the fabrication of mesoporous materials. In comparison, the sensor with the help of PVP (SPVP) exhibits excellent gas-sensing performances to ethanol and H2S due to its relative higher porosity. Especially, SPVP shows a high response (368), fast response/recovery time (4 s/20 s), and good selectivity toward H2S gas. In addition, it is found that NaOH and sodium citrate are also important for the morphological formation of SnO2 nanoflowers.