Remarkably Enhanced Methane Sensing Performance at Room Temperature via Constructing a Self‐Assembled Mulberry‐Like ZnO/SnO2 Hierarchical Structure

Development of metal oxide semiconductors‐based methane sensors with good response and low power consumption is one of the major challenges to realize the real‐time monitoring of methane leakage. In this work, a self‐assembled mulberry‐like ZnO/SnO2 hierarchical structure is constructed by a two‐step hydrothermal method. The resultant sensor works at room temperature with excellent response of ~56.1% to 2000 ppm CH4 at 55% relative humidity. It is found that the strain induced at the ZnO/SnO2 interface greatly enhances the piezoelectric polarization on the ZnO surface and that the band bending results in the accumulation of chemically adsorbed O2− ions close to the interface, leading to significant improvement in the sensing performance of the methane gas sensor at room temperature.


Introduction
[3][4] Being colorless and odorless, methane (CH 4 ) can form an explosive mixture with ambient air when its concentration reaches up to 5%. [5]Reliable and sensitive sensors are therefore required to monitor the concentration of methane in the environment in real time.Methane is of a special tetrahedral nonpolar molecular structure, which is extremely stable due to the hydrocarbon (C-H) bond energy being as high as 413 kJ mol −1 .This makes the sensing of methane very difficult, particularly at room temperature.It is usually believed that noble metals promote the catalytic oxidation at low temperatures and enhance the low temperature sensing response.The CH 4 sensing response of the recently developed sensors containing noble metals was still low at room temperature.For example, the response of the sensor based on Pd-doped SnO 2 /reduced graphene oxide was lower than 10% to 14 000 ppm of CH 4 , and the corresponding response time was longer than 5 min. [6]Currently, various types of ultrahighly sensitive CH 4 sensors based on metal oxide semiconductors work only at high temperature from 100 °C to 420 °C. [7]Additional heater is therefore required, resulting in complexity in device design and fabrication, high power consumption and many safety issues.
Recent studies have showed that, with the assistance of light irradiation, the response of a metal oxide-based CH 4 sensor can be significantly enhanced at room temperature.Wang et al. [8] reported that ZnO nano-sheets can sense CH 4 at room temperature with UV irradiation, with the response being 48% to a concentration of 1000 ppm.Chen et al. [9] found that the oxygen vacancy-enriched ZnO/Pd hybrid worked at a temperature as low as 80 °C with a response of 36.8% to 1000 ppm CH 4 under visible light illumination.Xia et al. [10] further demonstrated that the response of Pd-decorated ZnO/rGO to 10 000 ppm CH 4 at room temperature increased from 4.1% to 63.4% under Vis-light illumination.It was shown that the light-active catalysis, the efficient charge transfer, and the multiple heterojunctions formed within the hybrids synergistically enhanced the response at room temperature. [10]The synthesized ZnO/ g-C 3 N 4 porous hollow microspheres showed a response of 42% at room temperature under UV illumination, with the response time being reduced to 28 s. [11]It was also demonstrated that the oxygen vacancies formed at the hybrid ZnO/g-C 3 N 4 interface attracted more chemically adsorbed O À 2 ions at the interface under UV irradiation, thus promoting the rate of CH 4 oxidation. [11]lthough the light assisted CH 4 sensors exhibit attractive sensing performance at room temperature, the additional power consumption and extra light stability make the device control more complicated.It is desirable to design a material with high CH 4 sensing performance without the assistance of heat or light.The main breakthrough is depended on finding a suitable energy substitution to provide the extra energy for the activation of CH 4 dissociation at low temperature.ZnO is a polar semiconductor with strong piezoelectric property due to its lack of central symmetry in the wurtzite crystal structure. [12]The polarization electric field can be built under strain condition.It is expected that, by selecting suitable metal oxide which is to form multiheterojunction interface with ZnO, the deliberately created strain located at the oxide/ZnO interface may help enhance the polarization electric field and then provide the extra energy for the CH 4 dissociation.ZnO/SnO 2 hybrid composites have been extensively studied in the past Development of metal oxide semiconductors-based methane sensors with good response and low power consumption is one of the major challenges to realize the real-time monitoring of methane leakage.In this work, a selfassembled mulberry-like ZnO/SnO 2 hierarchical structure is constructed by a two-step hydrothermal method.The resultant sensor works at room temperature with excellent response of ~56.1% to 2000 ppm CH 4 at 55% relative humidity.It is found that the strain induced at the ZnO/SnO 2 interface greatly enhances the piezoelectric polarization on the ZnO surface and that the band bending results in the accumulation of chemically adsorbed O À 2 ions close to the interface, leading to significant improvement in the sensing performance of the methane gas sensor at room temperature.
decades.The diverse nano-structures produced by different methods make the hybrid composites function effectively in sensing of different gases, such as NO 2 , ethanol, and CO. [13,14] As far as the authors are concerned, however, there is not yet any report on CH 4 sensing at room temperature.
In this work, a ZnO/SnO 2 hierarchical structure was deliberately constructed.The self-assembled hexagonal ZnO nanorods growing along its [001] direction on AZO glass substrates were applied as the trunks, and the SnO 2 nanoparticles were self-assembled on the ZnO trunks to form a mulberry-like structure.The self-assembled nanorods structure is expected to provide an excellent carrier transport to enhance the gas response at low temperature, as in the TiO 2 system reported previously. [15,16]Heterojunction barriers and polarization field are expected to be created at the interface in the specific mulberry-like structure.It has been found that, the resultant sensor fabricated in this work showed an excellent CH 4 sensing performance at room temperature, providing a promising way in realizing the real-time monitoring of methane leakage with low power consumption.

Microstructural Characterization
Figure 1a gives the XRD patterns obtained for the samples of AZO substrate, ZnO, SnO 2 and ZnO/SnO 2 films over the 2θ range from 20°to 50°.A sharp peak is seen to appear 2θ = 34.2°for the AZO substrate, indicating that the AZO film is well orientated along [002] ZnO direction.A peak is at the similar position in the XRD spectra for the other three samples, respectively, suggesting that the ZnO, SnO 2 and ZnO/ SnO 2 films are highly orientated with the ZnO growing along its [002] and SnO 2 growing along its [101] direction, respectively.The enlarged XRD spectra at the local region of the peak in Figure 1b show that the ZnO (002) plane is located at 2θ = 34.35°,while the SnO 2 (101) plane is located at 2θ = 34.55°.It is interesting to notice that the sharp peak of the ZnO/SnO 2 film can be decomposed into three peaks.In addition to the two peaks located at the same position as ZnO and SnO 2 films, the new emerged peak is located at 2θ = 34.47°,indicating that there exists compressive and tensile strain on the ZnO and SnO 2 side of the interface, respectively.These may be caused by the anti-site substitution between Sn 4+ (atomic radius 0.069 nm) and Zn 2+ (atomic radius 0.074 nm) at the interface.The interplanar spacing of ZnO (002) in the strained part is calculated (based on the new emerged peak) to be about 96% of the unstrained part, suggesting the existence of compressive stress along the [002] direction, which changes the polarization electric field within the ZnO nanorods.
The SEM images in Figure 1c(i-iv) show the morphologies of the ZnO and ZnO/SnO 2 films viewed from the surface and cross-section.The orientated hexagonal ZnO nanorods have grown along its c-axis and are uniformly and well aligned with the AZO seed layer in the ZnO film, in agreement well with the XRD results in Figure 1a.In average, the ZnO nanorods are about 800 AE 50 nm in length and 100 AE 30 nm, based on statistical analysis inserted in the figure.In the ZnO/SnO 2 film, the spherical SnO 2 nano-particles decorate the sides and the top surface of the ZnO nanorods, forming a mulberry-like hierarchical structure.It is believed that at the beginning of the reaction, a lot of oxygen dangling bonds are formed due to the weak corrosion of ZnO surface.Acting as the nucleation sites, the dangling bonds enable the nucleation and rapid growth of SnO 2 nanoparticles on the ZnO nanorods, leading to the formation of the mulberry-like hierarchical structure.The firm contact between SnO 2 and ZnO provides a fast carrier transport path.

Oxygen Species
The high resolution XPS spectra of the films are shown in Figure 2. The two peaks located at around 1020.4 and 1044.5 eV in the Zn 2p spectrum are assigned to Zn 2p 3/2 and Zn 2p 1/2 of Zn 2+ species respectively (Figure 2a).The peak position of the Zn2p 3/2 in ZnO/SnO 2 is red shift by about 0.11 eV compared with that of the ZnO film.The core levels of Sn 3d 5/2 and Sn 3d 3/2 are located at 485.67 and 494.12 eV in the ZnO/SnO 2 film, showing a blue shift by about 0.78 and 0.69 eV respectively, compared with that of the SnO 2 (Figure 2b).The large blue shift suggests that there exists a strong interaction between the Zn 2+ and Sn 4+ species.The blue shift of Sn 3d peaks and the red shift of Zn 2p peaks in ZnO/SnO 2 indicate that the charge is transferred from SnO 2 to ZnO.
The typical O 1s peak of the ZnO, SnO 2 and ZnO/SnO 2 films can be decomposed into two peaks (Figure 2c).The low-energy one at ~529 eV is attributed to the lattice oxygen in ZnO (L O ), while the highenergy one at 531 eV is assigned to the adsorbed oxygen in oxygen vacancies (V O ). [17,18] No peak related to the adsorbed water molecules (H-O-H) (at ~532.5 eV) appears in the three samples.Besides, the concentration of V O on the surface in the SnO 2 film is 17.7%, much lower than that of the ZnO film (28.5%).However, in the ZnO/SnO 2 film where the SnO 2 nano-particles fully cover the surface of ZnO nanorods, the V O content is 24.9%.This suggests that more oxygen vacancies are formed at the SnO 2 side due to the anti-site substitution between Sn 4+ and Zn 2+ at the ZnO/SnO 2 interfaces.The large amount of V O acted as the active sites and thus allowed more O À 2 chemically adsorbed close to the interfaces.
The EPR spectra in Figure 3a verify further the formation of the oxygen vacancies.The similar EPR signals locate around g = 2.003 in the ZnO and SnO 2 and ZnO/SnO 2 samples, demonstrating the existence of single-electron-trapped surface defects, V O or O À s . [19]In addition, no signal related to lattice electron trapping sites, such as Zn + or V À Zn , is observed at g = 1.960, suggesting that the surface oxygen vacancy is the dominant defect type. [18]he PL spectra of ZnO, SnO 2 and ZnO/SnO 2 films are shown in Figure 3b.There are two main peaks located at 378 and 620 nm in the ZnO film, corresponding respectively to the recombination at nearband edge and the transitions of involved oxygen vacancy (V O ). [20] No PL peak was observed for the SnO 2 film.In the ZnO/SnO 2 film, the emission at the near-band edge is completely suppressed, indicating that the photo generated electron-hole pairs are effectively separated.The position of the V O defect-related peak is found to red shift to 640 nm, and its intensity decreases by 90%.From the XPS results, it is believed that the V O move to SnO 2 in the ZnO/SnO 2 film and thus the PL intensity greatly decreases.

Electronic Structure
The electronic structure of the ZnO/SnO 2 film was determined by the analysis combining UV-Vis absorption spectroscopy and XPS valence edge spectroscopy.Both ZnO and SnO 2 are direct bandgap Energy Environ.Mater.2024, 7, e12624 semiconductors, so the optical band energy can be obtained by Tauc plots using the Kubelka-Munk equation (Equation 1). [21] wherein α, hv, A and E g are the optical absorption coefficient, the photon energy, the proportional constant and the optical band energy, respectively.There are two absorption edges in the ZnO film, one is 3.30 eV related to the ZnO film, which was consistent with the near-band edge emission of PL spectra; and the other is 3.03 eV, which origins from the light trapping effect of the ZnO nanorods.
The band gap of SnO 2 film is 3.53 eV.The valence spectrum gives the information of the energy difference between the Fermi level and the top of the valence band (Figure 4b), and then the band alignment at the interface between ZnO and SnO 2 is obtained, as shown in Figure 4c.At the SnO 2 /ZnO interface, there exist transfer of electrons from SnO 2 to ZnO and transfer of holes in the opposite direction under the build-in electric field, resulting in an effective charge separation.This is in a good agreement with the PL results (Figure 3b) and the XPS analysis (Figure 2).

CH 4 Gas Sensing Properties
The  concentration can be derived by the Langmuir isotherm adsorption model of the unimolecular layer as described in Equation ( 2). [22] where R is the response and C CH4 is the CH 4 concentration (in ppm), R s is the saturation response and a is constant related to the adsorption coefficient.As shown in Figure 5b, there is a very good agreement between the experimental data and the fitting curve, indicating that the Langmuir isotherm adsorption model is reliable to explain the CH 4 response behavior, with the values of R s and a being worked out as 65% and 0.004 in the case of the ZnO/SnO 2 sensor.The lowest limit of detection (LOD) can be theoretically calculated by Equation (3) based on the response on low concentration region as inserted in Figure 5b, [23] LOD ¼ 3R noise =L slope (3)   where R noise is measured noise of the sensors, and L slope is the slope of fitting curve.R noise value of ZnO/SnO 2 sensor is calculated as 0.045 based on Equation (4) below: where, R i is the experimental data (i.e.various response of CH 4 concentration) and R is the fitting values.Therefore, the LOD value of the ZnO/SnO 2 sensor is 9.2 ppm.
The sensing performance shows good repeatability and stability, as demonstrated by the repeatedly measured response to 800 ppm CH 4 using cycle tests for five times, as shown in Figure 5c.The device maintains good response about 75% of the original value after 50 days storage in the ambience without any preservation.Figure 5d gives the result of the selectivity test of the sensor to various gases, including H 2 , CO, NO 2 , NH 3 , H 2 S, and C 2 H 6 at the same concentration of 800 ppm.It is obvious that the ZnO/SnO 2 sensor exhibits superior CH 4 selectivity to the other gases.Figure 5e,f show the dynamic response of the ZnO/SnO 2 to 800 ppm CH 4 at various relative humidity.The initial resistance decreases gradually as the humidity increases and stabilizes under high humidity.The response is 52% at 36% humidity, and it decreases as the humidity increases.It is noticed that the good response as high as 40% is realized even under the working condition of 100% saturated humidity, demonstrating that the ZnO/SnO 2 sensor is universally applicable under various humidity conditions.
Table 1 gives a direct comparison of the CH 4 sensing performance between the ZnO/SnO 2 sensor and the other metal oxide-based sensors.Significantly, the ZnO/SnO 2 sensor developed in this study works at room temperature without the assistance of Vis-or UV-light, with a higher and faster response, even superior to the other sensors working at a higher temperature.The outstanding performance makes the ZnO/ SnO 2 sensor one of the most promising candidates for commercial application in CH 4 sensing.

CH 4 Gas Sensing Mechanism
It is general accepted that the CH 4 gas sensing performance of a sensor depends on the thermos-activated oxidation reaction with the target gas CH 4 and the carrier concentration of the semiconductor, both being closely related to the working temperature of the sensors.Hence, CH 4 molecules are activated usually with the assistance of heat or light, which supplies the energy required for breaking the C-H bonds.The activated CH 4 molecules reacts with the oxygen species (O À 2 , O − or O 2− ) adsorbed on the surface of metal oxides, producing CO 2 and H 2 O, and releasing electrons in the meantime.At a temperature lower than 100 °C, the adsorbed oxygen species on the surface are O À 2 , and  Energy Environ.Mater.2024, 7, e12624 the reaction is described by Equation ( 5). [26]he released electrons inject into the semiconductor, reducing the resistance and thus producing the CH 4 sensing response.
The key point to realize the CH 4 sensing at room temperature is to lower the activation energy or to provide enough energy required for the CH 4 dissociation.ZnO is a polar semiconductor, and the strain induced polar electric field provides the additional energy required for the CH 4 dissociation.
The polarization of ZnO was detected by Raman spectroscopy.The typical Raman spectrum of the AZO in Figure 6a is similar to that of the low-Al:ZnO sample, as described in the previous study. [27]There are four main peaks located at 105, 281, 443 and 587 cm −1 , related to the E 2g (low), A 1g (TO), E 2g (high) and E 1g (LO) mode, respectively.The intensity of E 2g modes is related to the orientated growth of ZnO in its c axis direction.A 1g (TO) is the Zn-O stretching vibration along the c axis direction.The E 1g (LO) mode is forbidden in a perfect crystal, but activated by the electric field on the surface of the ZnO nanorods. [28]he Raman spectra of the ZnO and ZnO/ SnO 2 films show similar peaks to those of the AZO.Although no SnO 2 related peak was observed in the ZnO/SnO 2 film due to the small amount of SnO 2 , the ZnO related peaks shift due to the influence of the interaction between ZnO and SnO 2 .The much enhanced E 2g peak of the ZnO compared with that of the AZO indicates the direc-  tional growth of ZnO nanorods along the c axis, in agreement with the XRD and SEM results (Figures 1 and 2).The position of the A 1g (TO) peak in ZnO is at 283.2 cm −1 , which red shifts to 280.9 cm −1 in the ZnO/SnO 2 film as shown in Figure 6b, suggesting the shortening of the Zn-O bond along the c axis direction.The intensity of the E 1g (LO) peak gradually increases as ZnO is growing on AZO and then SnO 2 nanoparticles are decorating the ZnO, indicating that the polarization electric field is gradually enhanced.All these features in the Raman spectroscopy demonstrate that the ZnO nanorods are compressed along the c axis direction and that the polar electric field is the strongest in the ZnO/SnO 2 film.
The piezoelectric property of the ZnO/SnO 2 films was measured by piezoelectric forces microscopy (PFM) and compared with that of the ZnO film.The corresponding PFM images are shown in Figure 7a,b.The typically butterfly shaped loop, as shown in Figure 7c, indicates the existence of the piezoelectric effect.The averaged phase angle is 183°and 175°for ZnO and ZnO/SnO 2 films, respectively (Figure 7d).The phase angle close to 180°verifies that the response was piezoelectric instead of electrostatic.It has been reported that the polarization coefficient d 33 of the pristine bulk ZnO was ~9.9 pm V −1 , while that of the films with well orientation was ~12.4 pm V −1 . [29]In this work, the average d 33 of the ZnO film is 18.5 pm V −1 , and that of the ZnO/SnO 2 film is further enhanced up to 24.1 pm V −1 .The lattice strain will induce the permanent local electric dipoles, and the well aligned ZnO nanorods results in the great enhancement of the piezoelectric polarization (d 33 ) of the ZnO/SnO 2 .This functions like the assistance of heat or light, providing the additional energy required for CH 4 dissociation, and thus realizing the enhancement of the CH 4 sensing at room temperature.
The activation energy is further measured and calculated according to the Arrhenius equation.The gas sensing performance of ZnO and ZnO/SnO 2 sensors were measured at 25 °C, 100 °C and 150 °C, respectively.The plot of ln normalized resistance change rate (d(R/R 0 )/ dt) as a function of 1/T produces a straight line (Figure 8).From the slope based on Arrhenius equation, the activation energy (ΔE a ) for the process involved was calculated as 2.91 kJ mol −1 for ZnO/ SnO 2 , smaller than one third of 10.18 kJ mol −1 , the corresponding value for ZnO.This further proves that the enhanced polarization field in the new mulberry-like hierarchical structure can greatly reduce the activation energy, and thus significantly enhances CH 4 sensing performance at room temperature.
Oxygen vacancies usually provide the site for O À 2 adsorptions on the surface.The analysis on the energy band structure indicates that the electron charges will transfer from SnO 2 to ZnO, and the accumulated electrons will readily attract the O 2 molecules chemically adsorbed close to the interface to form O À 2 .This enables the reaction of CH 4 and O À 2 more likely to occur at the interface, where the polar electrical field is the strongest.

Conclusions
In this study, methane sensing performance at room temperature is remarkably enhanced via constructing a self-assembled mulberry-like ZnO/SnO 2 hierarchical structure.The synergistic effect of the strongest polarized electric field and the maximum density of chemically adsorbed O À 2 in the region close to the ZnO/SnO 2 interface provides not only the energy required for activation of CH 4 dissociation, but also the large number of the reaction sites, leading to a significantly enhanced response at room temperature.This work provides an  effective way to integrate the polarized electric field into the gas sensing to reduce the reaction temperature down to room temperature.

Experimental Section
Fabrication of ZnO, SnO 2 and ZnO/SnO 2 films: Aluminum-doped zinc oxide (AZO)-coated glass substrates were used to provide suitable seed layer.Hydrothermal method was used for the growth of ZnO nanorods on the AZO similar to that described in the previous work. [15]Two AZO substrates were placed diagonally with conductive surface downward in a 200 mL stainless steel Teflon-lined autoclave.The precursor solution was prepared by adding 30 mM hexamethylenetetramine (C 6 H 12 N 4 , HMT) and 30 mM Zn(NO 3 ) 2 Á6H 2 O powders into 100 mL deionized water and stirring to form a homogeneous solution.After the hydrothermal reaction at 100 °C for 4 h, the samples were cleaned, dried and then annealed at 400 °C for 20 min in air to form ZnO nanorods films (named as ZnO films).
SnO 2 films were also grown on AZO substrates using hydrothermal method.The SnO 2 precursor solution was prepared by mixing 3 mM SnCl 4 Á5H 2 O and 10 mM NaOH powders with 100 mL deionized water.The hydrothermal reaction lasted 6 h at 180 °C.The samples were cleaned, dried and then annealed at 550 °C for 30 min in air to form SnO 2 films.
On the basis of hydrothermal grown ZnO films, the ZnO/SnO 2 films were fabricated via a second-step hydrothermal process.Two AZO glasses with ZnO nanorods films were put back into the autoclave to allow growth of the ZnO/ SnO 2 nanostructure.The precursor solution and the hydrothermal reaction condition were the same as that used for growing SnO 2 films.
Characterization: The phases and growth orientation of the films were identified using X-ray diffraction spectrometry (XRD, Bruker D8 advanced, Germany) with Cu Kα X-ray source (λ = 0.154 nm).The morphology of the films was examined by field emission scanning electron microscopy (FESEM, Sigma500, Zeiss).Xray photoelectron spectroscopy (XPS, Escalab 250Xi, Thermo Fisher, USA) was conducted to analyze the chemical composition and valence states of elements.Photoluminescence (PL) was conducted by a fluorescence spectrometer (Pico-Quant, FluoTime 300) at the excitation wavelength of 320 nm.The UV-Vis spectroscopy (UV-3600, Shimadzu, Japan) was employed to examine optical absorption properties.Raman spectra were collected by high-resolution confocal micro-Raman spectroscopy (Horiba JY LabRAM HR800, France) with the laser wavelength of 532 nm.Unpaired electrons of the film sample, i.e., oxygen free radicals, were tested by an electron paramagnetic resonance (EPR spectra, Bruker EMXplus spectrometer, Germany).The piezoelectric polarization was measured by piezoelectric forces microscopy (PFM, SPA 400, Seiko Inc.).
Gas sensing test: Platinum interdigitated electrodes were deposited onto the thin films by DC magnetron sputtering as described in our previous work. [30]The sensing performance was assessed in air at room temperature about 25 °C.A Keithley DAQ6510 multimeter was used to measure the resistance change of the sensors.The sensing response (S) is calculated by Equation ( 6): where R air and R gas are the resistances of the sensor in air and in gas atmosphere (i.e.mixed 5% CH 4 and 95% Ar) respectively.The response/recovery time (t res /t recov ) is defined as the time taken for the change of the resistance to reach 90% of ΔR in the case of gas in and gas out, respectively. [26]

Figure 1 .
Figure 1.a) XRD patterns of the AZO, ZnO, SnO 2 and ZnO/SnO 2 films; b) The zoomed XRD spectra; and c) SEM images showing the morphology of ZnO and ZnO/SnO 2 films viewed from the surface and cross-section, respectively.

Figure 5 .
Figure 5. CH 4 gas sensing properties of the ZnO/SnO 2 film.a) Dynamic response change curve to 100-2000 ppm CH 4 ; b) the corresponding response curve fitted by the Langmuir adsorption isotherm equation; c) repeatability to 800 pm CH 4 upon five exposure-release cycles; d) selective measurement at 800 ppm.e) Dynamic response change curve and f) the response under various relative humidity.

Figure 7 .
Figure 7. PFM images of domain structures a) ZnO and b) ZnO/SnO 2 films; c) A representative amplitude by PFM versus applied bias voltage curve; d) phase-voltage hysteresis loop versus applied bias voltage curve.

Table 1 .
Comparison of CH 4 gas sensing properties between the ZnO/SnO 2 sensor and fabricated from varying nanohybrids and nanostructures in previous reports.