ZIF-8/SnO₂ based high sensitivity ethylene gas sensor with Au-GO doped

Abstract

Metal oxide semiconductor gas sensors exhibit significant advantages in gas detection due to their high sensitivity and low cost. However, challenges such as poor selectivity and insufficient stability remain critical scientific issues. In this study, tin dioxide nanomaterials with a unique structure were successfully synthesized using ZIF-8 as a template. Further modification with gold-decorated reduced graphene oxide yielded a nanocomposite that demonstrated rapid response, high sensitivity, and excellent selectivity for low-concentration ethylene detection. The crystal structure, morphology, elemental composition, and pore size distribution of the materials were systematically characterized using XRD, FESEM, EDS, UV-Vis spectroscopy, and N₂ adsorption-desorption analysis. Gas sensing tests revealed that the sensor exhibited a response value of 5.35 to 20 ppm C₂H₄ at an optimal operating temperature of 280 °C, with response and recovery times of 14 s and 17 s, respectively. The selectivity ratio for ethylene over the second most sensitive gas was 3.26, highlighting its superior specificity. Additionally, the sensor demonstrated good stability and repeatability, providing a cost-effective solution for real-time ethylene monitoring in humid environments.

Introduction

Ethylene (C₂H₄) plays a pivotal role in industrial and agricultural sectors, serving as a critical raw material in polymerization processes, plastics production, and fruit ripening regulation^1,2. Its precise detection is essential for optimizing industrial efficiency, ensuring agricultural product quality, and mitigating safety risks due to its flammability³.

Tin dioxide (SnO₂), a widely studied n-type semiconductor with a bandgap of 3.6 eV, has been extensively explored for gas sensing owing to its chemical stability, high electron mobility, and low fabrication cost⁴. For example, Pil Gyu Choi et al.⁵ synthesized SnO₂ nanosheets with remarkable selectivity towards olefin gases (C₂H₄, C₃H₆, C₄H₈), and higher response values than those of alkane gases (CH₄, C₂H₆, C₃H₈, n-C₄H₁₀).

Despite the fact that pure SnO₂ gas sensors are sensitive to a variety of gases, but generally poor in terms of selectivity as well as response and high relative operating temperature^6,7. To address these limitations, strategies such as noble metal doping, heterojunction construction, and morphological engineering have been employed to enhance sensor performance⁸. Reduced graphene oxide (rGO) has an atomic-level thickness that endows the material with an extremely high specific surface area⁹. Meanwhile, the abundant surface defects and oxygen-containing functional groups enable it to form strong interactions with metal oxide semiconductor materials, thereby generating a synergistic effect within the composite material system¹⁰. This synergistic effect not only contributes to the improvement of the electrical properties of the material but also significantly enhances its adsorption and reaction activity towards target gases, thus effectively enhancing the gas sensing performance of the sensor. For instance, Jiang et al.¹¹ reported that SnO₂/NiO/rGO ternary composites exhibited a 21.11 response to 50 ppm acetone at 350 °C, with a response time of 5 s. Compared with SnO₂/NiO, SnO₂/NiO/rGO significant improvements in response speed and sensitivity are exhibited.

Gold nanoparticles (Au NPs), known for their catalytic activity and π-electron interactions with C₂H₄, further enhance sensitivity by lowering the activation energy for C₂H₄ oxidation¹². When Au exists in the form of nanoparticles, the coordination unsaturation of surface atoms increases, generating abundant active sites. These active sites can adsorb C₂H₄ molecules, inducing a certain degree of distortion and polarization of their carbon-carbon double bonds, thereby reducing the activation energy for C₂H₄ molecules to participate in redox reactions and promoting reaction progression¹³. Shiva et al.¹⁴ investigated the gas-sensing properties of 0.5 wt% Au-doped SnO₂ thick-film sensors toward n-butanol. The test results revealed that Au significantly altered the operating temperature and substantially enhanced the response level of the gas sensor. Compared with pure SnO₂, the Au/SnO₂ sensor exhibited a 70 °C reduction in operating temperature and a 4-fold increase in response. Studies indicate that upon metal doping, a depletion layer forms on the surface of metal oxide particles, altering the surface potential of SnO₂. The increased potential barrier ultimately improves the sensing performance.

Metal-organic frameworks (MOFs) are formed through the coordination-driven self-assembly of inorganic metal nodes and organic ligands. These architectures exhibit distinctive advantages including structural diversity, high hydrophobicity, exceptional porosity, and tunable pore dimensions^15,16. Such characteristics confer exceptional application potential in gas adsorption, separation, and sensing applications, establishing MOFs as prominent research foci in contemporary materials science. In a representative study, Homayoonnia Setareh¹⁷ synthesized a ZIF-8/MWCNT-coated sensor demonstrating exceptional acetone responsiveness. This sensor exhibited minimal humidity interference, high selectivity, excellent repeatability, complete reversibility and rapid response kinetics. Notably, it achieved a detection limit (LOD) of 1.7 ppb at 25 °C.

Against the aforementioned research background, this study innovatively proposes and fabricates the Au-rGO/ZIF8-SnO₂ composite material, aiming to integrate the high electrical conductivity and surface activity of rGO, the molecular sieve effect of ZIF-8, and the gas-sensing properties of SnO₂ to construct a novel high-performance C₂H₄ gas sensor. Using advanced material characterization techniques, including X-ray diffraction (XRD), field-emission scanning electron microscopy (FESEM), energy-dispersive X-ray spectroscopy (EDS), and nitrogen adsorption-desorption, a comprehensive and in-depth analysis and characterization of the crystal structure, microstructure, and pore size distribution of the Au-rGO/ZIF8-SnO₂ composite material was conducted.

Experimental

All reagents used were of analytical grade, purchased from Sinopharm Group, and used without further purification.

Synthesis of ZIF-8

ZIF-8 was synthesized using Zn(CH₃COO)₂·2H₂O as the zinc source and 2-methylimidazole as the ligand. Briefly, 0.594 g of Zn(CH₃COO)₂·2H₂O and 1.313 g of 2-methylimidazole were separately dissolved in 50 mL of methanol under continuous stirring. The two solutions were mixed and stirred for 30 min to form a white turbid suspension. Then transferred into a Teflon-lined autoclave and heated at 120 °C for 24 h. The product was centrifuged at 4000 rpm, washed three times with methanol, and dried at 60 °C under vacuum for 12 h to obtain ZIF-8 nanopowder.

Synthesis of Au-GO

Au nanoparticles (NPs) were synthesized by dissolving 20 mg of HAuCl₄·3H₂O in 50 mL of deionized water (DI) under reflux at 115 °C for 15 min. A solution of 147 mg trisodium citrate dihydrate (C₆H₅Na₃O₇·2H₂O) in 10 mL DI was slowly added to the refluxing solution, which turned from gray to purple and finally wine-red. After cooling, 200 mg of graphene oxide (GO) was dispersed in 140 mL DI via 90 W ultrasonication for 1 h. The Au colloid was mixed with the GO dispersion, stirred at 95 °C for 24 h, centrifuged at 15,000 rpm, washed with ethanol and DI, and dried at 75 °C to yield Au-rGO.

Synthesis of Au-GO/ZIF8-SnO₂

0.226 g SnCl₂·2H₂O and 0.3 g urea were dissolved in 50 mL ethylene glycol. The pre-prepared 0.5 g of ZIF-8 powder was added to the solution, followed by 3 h of ultrasonic dispersion to ensure complete dissolution of the ZIF-8 powder. A certain amount of Au-rGO powder was then added, followed by an additional 1 h of ultrasonication. The mixture was hydrothermally treated at 150 °C for 12 h, washed with DI and ethanol, dried at 60 °C, and calcined at 600 °C (2 °C/min) for 3 h. The final products, labeled ZIF-8/SnO₂, NG1 (1 at% Au-rGO), and NG2 (1.5 at% Au-rGO), were obtained.

The synthesis schematic is shown in Fig. 1.

Fig. 1

Experimental set-up for the nanocomposites

Gas sensor fabrication

The MEMS sensor is a cuboid with dimensions of 4 × 4 × 1 mm. There are pads on the bottom surface for soldering to the sensor signal acquisition circuit, the temperature-voltage curves presented in the Fig. 2 below. Material was prepared with ethanol and the adhesive aluminum silicate to form a slurry. The prepared slurry is applied to the interdigital electrodes of the micro hot plate to make a gas sensor. Finally, the gas sensor sample was maintained at an external heating voltage of 2.0 V for 72 h.

Fig. 2

temperature-voltage curves of micro hot plate

Synthetic air (21% O₂, 79% N₂) was used as the carrier gas, and different concentrations of C₂H₄ were configured by a gas dilution instrument, and the output port of the gas dilution meter was connected to the gas input port of our self-designed gas chamber. The gas flow continued to flow throughout the experiment.

Results and discussion

Material characterization

The XRD test was carried out using a Bruker D8 Advance instrument. Wide-angle diffraction tests were performed on the sample materials, and the test lasted for 10 minutes. The spectra are shown in Fig. 3.

Fig. 3

XRD spectra of ZIF8-SnO₂, NG2

The results in Fig. 3 indicate that all diffraction peaks of the prepared ZIF8-SnO₂ sample are indexed to ZnO (PDF#36-1451) and SnO₂ (PDF#41-1445), demonstrating the hexagonal polycrystalline structure of the as-prepared sample. The diffraction peaks at 2θ = 38.25° and 44.46° match well with the characteristic peaks of the (111) and (200) crystal planes of the Au standard card (PDF#99-0056). The XRD results demonstrate that the prepared nanomaterials contain ZIF-8 and SnO₂, with successful doping of Au.

In this study, FESEM characterization was performed using GeminSEM, with images of ZIF-8, ZIF8-SnO2, NG1, and NG2 shown in Fig. 4(a, b, c, d). Figure 4e, present the EDS images and quantitative data of NG1 and NG2 Fig. 4. FESEM images reveal that the prepared ZIF-8 exhibits a rhombic dodecahedral morphology with good symmetry, measuring approximately 0.8 μm in size. Figure 4b shows that ZIF8-SnO₂ consists of numerous SnO₂ nanoparticles attached to the ZIF-8 surface, forming gas diffusion channels between particles. The SnO₂ nanoparticles have a diameter of 10–20 nm. FESEM images of NG1 and NG2 Fig. 4c, d demonstrate that ZIF8-SnO₂ nanoparticles are anchored onto the surface of Au-GO nanosheets, with further reduced particle size.

Fig. 4: FESEM and EDS images.

a FESEM image of ZIF-8, b FESEM image of ZIF8- SnO₂, c FESEM image of NG1, d FESEM image of NG2, e EDS image of NG1 f EDS image of NG2

EDS elemental decomposition results confirm the presence of SnO₂, ZIF-8, and Au in the nanocomposites, with the atomic percentage of Au varying with doping concentration. These results verify that Au-GO has been successfully modified onto the nanoparticle surface, forming a heterostructural architecture.

We conducted TEM and HRTEM characterization of the Au-GO sample using an FEI Tecnai G2 F30 field-emission transmission electron microscope. Figure 5a is TEM image of Au-GO, reveals that the synthesized GO exhibits a film-like multilayer structure. Gold nanoparticles with diameters of approximately 10–20 nm are attached to its surface, and most of these particles are uniformly distributed on the GO surface with minimal aggregation. Clear lattice fringes belonging to the Au nanoparticles were observed in the HRTEM image Fig. 5b.

Fig. 5: TEM and HRTEM images.

a TEM image of NG2, b HRTEM image of NG2

Raman spectroscopy characterization of GO and Au-GO nanomaterials was performed using a 514 nm laser as the excitation source, with the results shown in the Fig. 6. As can be observed, both GO and Au-GO exhibit three main characteristic peaks corresponding to the D band, G band, and G’ band of graphene. The ID/IG ratio for GO is 1.04. After doping with Au nanoparticles, the ID/IG ratio of Au-GO increases to 1.11, indicating that the incorporation of Au nanoparticles introduces a higher density of structural defects into the GO framework.

Fig. 6

Raman spectra of GO and Au-GO

We measured the UV–Vis absorption spectra and energy band gaps of ZIF8-SnO₂ and NG2 using UV - 3600 UV–Vis–NIR spectrophotometer. Figure 7a displays the UV-Vis absorption spectra of the two materials. The band gaps were calculated using the Tauc Plot method based on the absorption spectra. Results show that the band gaps of ZIF8-SnO₂ and NG2 are 3.19 eV and 2.79 eV, respectively. The narrowing of the band gap after Au-GO doping is attributed to the increased oxygen vacancy concentration. As oxygen vacancies act as donor defects, their presence shifts the Fermi level of the oxide upward, creating defect energy levels within the band gap and thus reducing the band gap. Additionally, the narrower band gap indicates that electrons in NG2 can more easily transition from the valence band to the conduction band, which enhances the sensitivity of the gas-sensing material toward target gases—a finding consistent with the gas-sensing performance test results of the sensor.

Fig. 7: Energy band gap test.

a UV vis diffuse reflectance spectra b Kubelka Munk function curves of ZIF8-SnO₂ and NG2

Specific surface area of metal oxide semiconductors significantly affects the adsorption and desorption capacities of target gases. So we investigated the pore size distribution and specific surface area of ZIF8- SnO₂ and NG2, we conducted nitrogen adsorption - desorption measurements using Micromeritics ASAP 2460. The result is shown in Fig. 8.

Fig. 8

N₂ adsorption desorption isotherms of ZIF8-SnO₂, and NG2 at 77 K

N₂ adsorption-desorption isotherms of ZIF8-SnO₂ and NG2 demonstrate that both samples exhibit Type IV isotherms with H3-type mesoporous hysteresis loops appearing at relative pressures (P/P₀) of 0.5–0.9, indicating that the pore structure of the synthesized materials consists of slit pores formed by the accumulation of numerous particles—a finding consistent with SEM characterization results. Using the BET adsorption isotherm equation, the BET specific surface areas of ZIF8-SnO₂ and NG2 were calculated to be 12.57 m²/g and 111.17 m²/g, respectively. These results reveal that Au-GO doping significantly enhances the specific surface area and pore size, which provides abundant active centers for oxygen adsorption and chemical reactions with analytes, thereby improving the gas-sensing properties of the sensor.

In situ XPS characterization of the material was performed, with the survey spectrum shown in Fig. 9a confirming the presence of five elements: Zn, Sn, C, O, and Au. Fig. 9b displays the Au 4f spectra of the NG2 material under clean air versus C₂H₄ gas exposure. A positive shift of approximately 0.3–0.4 eV in the Au 4f binding energy is observed upon C₂H₄ introduction, indicating altered electronic states at the gold surface. This observation indicates electron donation from Au d-orbitals to the π* antibonding orbitals of adsorbed C₂H₄ molecules, resulting in an increased effective oxidation state of gold that elevates the core-level binding energy. Figure 9c, d present the high-resolution O 1s spectra of ZIF-8/SnO₂ and NG2, respectively. Both spectra were deconvoluted into three Gaussian components corresponding to: lattice oxygen (O_L), oxygen vacancies (O_V) and adsorbed oxygen (O_C). The proportion of O_C in NG2 significantly increases, proving that after doping, the material's adsorption capacity for oxygen is enhanced. Additionally, the increased proportions of O_V and O_C demonstrate that NG2 possesses more defects and stronger capability to adsorb ionic oxygen, leading to band gap narrowing. This occurs because oxygen vacancies belong to donor defects, whose presence causes the Fermi level of the oxide to shift upward, introduces defect levels within the band gap, and consequently reduces the band gap.

Fig. 9: XPS characterization.

a XPS spectra of ZIF8-SnO₂ and NG2, b XPS characterization of Au 4f; c O1s spectra of pure ZIF8-SnO₂; d O1s spectra of pure NG2

Gas-sensing properties

We test the gas sensing performance based on the prepared materials toward C₂H₄. Prior to conducting other gas sensing experiments, the optimal operating temperature of the sensor was prioritized. Therefore, the response curves of gas sensors based on three nanomaterials (ZIF8-SnO₂, NG1, and NG2) to 50 ppm C₂H₄ gas were first measured within the temperature range of 160 °C to 360 °C, with results shown in the Fig. 10.

Fig. 10

Sensor response value vs. operating temperature for 50 ppm C₂H₄

In terms of response sensitivity, at the same operating temperature and for the same concentration of C₂H₄, the NG2 sensor exhibits a higher response value, while the undoped ZIF8-SnO₂ sensor shows the lowest response and does not reach its optimal operating temperature within this temperature range. From the trend of temperature characteristic curves, the response curves of NG1 and NG2 sensors to C₂H₄ gas both exhibit a mountain-shaped profile: the response value increases with temperature rise, reaches a peak at the optimal operating temperature, and then decreases with further temperature increase. Notably, the optimal operating temperature of the high-doping NG2 sensor is lower than that of the low-doping NG1 sensor. Considering power consumption, subsequent tests were conducted at 280 °C with NG2 sensor.

Figure 11a shows the dynamic response-recovery transient curves of ZIF8-SnO₂, NG1, and NG2 sensors to 1–50 ppm C₂H₄ gas at an operating temperature of 280 °C. Figure 11b presents the linear relationship between the response value (S) of the NG2 sensor and lgC. The response value was calculated as R_a/R_g, where R_a is the sensor resistance in air and R_g is the resistance in the test gas, with C representing the gas concentration.

Fig. 11: Linearity test of sensors.

a Continuous dynamic response curves of ZIF8-SnO₂, NG1 and NG2 sensor to different concentrations of C₂H₄, b linear fitting curves between sensor response and C₂H₄ concentration

As observed, the responses of NG1 and NG2 gas sensors increase with rising C₂H₄ concentration, while the undoped ZIF8-SnO₂ sensor exhibits insignificant response due to not reaching its optimal operating temperature. Notably, Au-GO modification significantly improves sensor performance, with NG2 demonstrating the highest response value. Fitting the dynamic response curve of NG2 to C₂H₄ concentration yields a linear relationship: S = 4.25359lgC + 0.26533 (R²= 0.94006). A favorable linear correlation is observed between the sensor response values and C₂H₄ concentrations.

To investigate the response/recovery time characteristics of the NG2 sensor, its response and recovery to 20 ppm C₂H₄ were measured at an operating temperature of 280 °C, with results shown in Fig. 12. The response time and recovery time are defined as the time required for the sensor response to change from the initial state to 90% of the steady-state value. Results demonstrate that the NG2 sensor exhibits extremely rapid response and recovery times of 14 s and 17 s, respectively.

Fig. 12

Response/recovery curves to 20 ppm C₂H₄ at 280 °C

Figure 13 presents the response of the NG2 sensor to six gases at the same concentration (50 ppm): C₂H₄, methane (CH₄), ethane (C₂H₆), formaldehyde (HCHO), carbon monoxide (CO), and hydrogen (H₂). The sensor exhibits superior selectivity toward C₂H₄, with a response value 3.26 times higher than that of the second-highest responding gas (H₂).

Fig. 13

Selective testing of NG2 sensor

Response of the sensor to 20 ppm C₂H₄ under different humidity was tested. The sensor resistance value and the response value are shown in the figure. Elevated relative humidity induces a decrease in sensor resistance, attributed to the formation of a conductive layer by adsorbed H₂O molecules on the material surface. Furthermore, competitive adsorption occurs at active sites on the sensing material: H₂O molecules compete with target gas molecules for adsorption sites, reducing the probability of target gas adsorption. Water molecules preferentially adsorb and form hydroxyl groups, occupying surface active centers. This competitive adsorption mechanism, consistent with the Langmuir adsorption model, results in reduced material responsiveness Fig. 14.

Fig. 14: Sensor moisture resistance test.

a Resistance curve vs. humidity, b Sensor response vs. humidity

In terms of repeatability, the sensor was exposed to 50 ppm C₂H₄ gas for 7 consecutive cyclic tests as Fig. 15. The resistance of the NG2 sensor recovered to its initial state in each cycle, and the response-recovery characteristic curves of the 7 tests remained essentially consistent—demonstrating the excellent reproducibility of the NG2 sensor.

Fig. 15

Gas sensitivity test results of repeatability

The long-term stability of the sensor is another indispensable factor of practical application, so a study on the long-term stability of the NG2 sensor was conducted, as plotted in Fig. 16. The response of the sensor to 50 ppm C₂H₄ was measured over 20 days. Keep the sensor powered on for 20 days, testing at the same time every day. As shown by the results, the sensor was stable and exhibited only a very small decrease in response within 20 days.

Fig. 16

Gas sensitivity test results of long-term stability

We summarized the optimal gas sensing performance of the NG2 nanocomposite-based sensor and compared it with the reported MEMS C₂H₄ gas sensors in recent literature, as shown in the table. The table indicates that the gas sensor fabricated in this study exhibits favorable performance in terms of operating temperature, sensitivity, detection limit, and response/recovery time. The synthesized gas sensor holds potential application value for low-concentration C₂H₄ detection Table 1.

Table 1 Performance comparison of C₂H₄ gas sensors based on SnO₂

Sensing mechanism of the nanocomposites of Au-GO/Co-ZnO

The Au-GO/ZIF8-SnO₂ composite exhibits n-type semiconductor gas-sensing characteristics. Consistent with the general gas-sensing mechanism of n-type semiconductors, its widely accepted gas-sensing mechanism follows the gas adsorption-desorption models chematically depicted in Fig. 12. At the operating temperature of 280 °C, O₂ molecules act as electron acceptors, capturing free electrons from the conduction band of SnO₂ with the promotion of Au nanoparticles to form chemisorbed oxygen species (O₂⁻ and O⁻). During this process, the reduction of electrons in the material's conduction band leads to an increase in material resistance. Upon exposed to C₂H₄ gas, as shown in Eq. 1, C₂H₄ molecules react with the adsorbed oxygen species on the material surface, oxidizing into CO₂ and releasing abundant electrons. These electrons return to the conduction band, manifesting macroscopically as decreased resistance Fig. 17

$${{\rm{C}}}_{2}{{\rm{H}}}_{4}+{{\rm{O}}}^{-}\to {{\rm{CO}}}_{2}+{{\rm{e}}}^{-}$$

(1)

Fig. 17

Schematic diagram of sensing mechanism

The introduction of Au-GO significantly enhances the gas-sensing performance of the sensor, play the critical role in sensitizing the sensing material. The surface catalytic activity of Au NPs notably reduces the activation energy barrier for C₂H₄ oxidation reaction, releasing abundant electrons and lowering the operating temperature required for C₂H₄ gas reactions on the material surface—thereby improving sensor sensitivity. Furthermore, the d-band center of Au (≈−2.8 eV) matches the energy level of the C₂H₄ π* orbitals (LUMO≈−3.1 eV), forming strong hybridization that selectively adsorbs C₂H₄ and enhances sensor selectivity. Furthermore, BET calculations reveal that NG2 exhibits a significantly larger specific surface area compared to ZIF8-SnO₂. The excellent electrical conductivity of graphene accelerates gas adsorption/desorption and electron transport rates, further enhancing sensor sensitivity. Finally, the introduction of Au-GO forms a heterojunction, strengthening the electronic sensitization of the nanomaterial. As a p-type semiconductor (Au-GO) and n-type semiconductor (SnO₂) reach equilibrium, the electron depletion layer at the heterojunction interface expands, facilitating additional modulation of sensor resistance and improving gas-sensing performance. Moreover, Au-GO doping narrows the band gap of SnO₂, enabling easier electron transitions from the valence band to the conduction band and boosting the sensor’s gas-sensing performance.

Conclusion

In this study, Au-GO-doped ZIF8-SnO₂ nanomaterials NG1 and NG2 were successfully prepared and applied as sensitive materials for C₂H₄ sensors. Their physical and electrical properties are compared in Table 2.

Table 2 comparison of NG1 and NG2

The 1.5 at% Au-GO-doped sample (NG2) demonstrated a response value of 8.24 toward 50 ppm C₂H₄ at 280 °C, showcasing excellent linearity within the 1–50 ppm testing concentration range. Notably, the sensor exhibited remarkable stability, outstanding reproducibility, short response/recovery times, and superior selectivity for C₂H₄ detection. This study provides a novel approach and new insights for highly sensitive and low-detection-limit detection of C₂H₄.