Metal-Organic Framework-derived Au-ZnO Nanostructures for Ultrasensitive and Selective
H2S Gas Detection
(Donghyeon Kim)
1†
(Jimyeong Park)
2†
(Jiyeon Shin)
34
(Sunghun Hong)
1
(Minseo Kim)
1
(Chulwoong Han)
5
(Sung Cheol Park)
56
(Jae-Hyoung Lee)
4*
(Myung Sik Choi)
1234*
Copyright © 2025 The Korean Institute of Metals and Materials
Key words(Korean)
ZnO nanostructures, Au nanoparticles, Metal-organic framework, Gas sensor, Heterojunction
1. INTRODUCTION
H2S is a highly toxic, flammable gas that poses severe risks to the environment and
human health, even at trace concentrations. Occupational safety guidelines state that
exposure to H2S levels above 10 ppm can cause serious health issues [1-3], including respiratory paralysis, neurological disorders, and, in extreme cases,
sudden death. The presence of H2S in industries, such as petrochemical processing, sewage treatment, mining, and natural
gas extraction, necessitates continuous monitoring for accident prevention and regulatory
compliance [4-6]. Therefore, the development of highly sensitive, reliable H2S sensors exhibiting fast response and high selectivity has become a major priority.
Chemiresistive gas sensors, based on semiconducting metal oxides (SMOs), have garnered
significant attention owing to their low cost, simple device architecture, and suitability
for integration into portable, wireless monitoring systems [7-9]. Among various SMOs, ZnO, a wide bandgap (~3.2 eV) [10] n-type semiconductor, has been extensively explored owing to its tunable morphology,
chemical stability, and surface reactivity [11,12]. However, pristine ZnO exhibits inherent limitations, including moderate sensitivity
and poor selectivity toward H2S, as well as the need for high operating temperatures (> 200 °C), thus increasing
power consumption and decreasing long-term stability [13,14].
Noble metal functionalization has emerged as an effective strategy to overcome these
limitations and improve gas sensing performance. Incorporating Au nanoparticles onto
ZnO introduces abundant catalytic sites and generates Au-ZnO Schottky junctions that
facilitate charge transfer and modulate the surface depletion layer [15-17]. Additionally, Au decoration is known to promote the spillover of dissociated H2S species, accelerating reaction kinetics and amplifying sensor response [16,17]. Recent studies have reported substantially high H2S sensitivities of Au-functionalized ZnO [18], SnO2 [19], and WO3 [20] systems; however, challenges remain regarding selectivity against interfering gases
(e.g., NO2, NH3, and volatile organics) and reproducibility under varying temperature conditions
[21-23].
In addition to surface decoration, the structural design of ZnO influences sensor
performance. Conventional synthesis methods often yield dense morphologies with small
surface areas and a low density of active sites. Recently, the use of metal-organic
frameworks (MOFs) as sacrificial templates for metal oxide synthesis has attracted
considerable attention in gas sensing research [23-27]. MOFs, owing to their high surface area and well-defined porous structures, can
be thermally converted into nanostructured oxides containing abundant active sites
and enhanced charge transport pathways [28-30]. For instance, Montoro et al. [31] reported that MOF-derived CuO, NiO, and ZnO exhibit excellent selectivity toward
H2S, CO, and H2 gases, respectively, with substantially improved sensing capabilities relative to
the pristine MOF precursors. Furthermore, Ullah et al. [32] demonstrated that zeolitic imidazolate framework (ZIF)-67-derived Co3O4 coupled with g-C3N4 nanosheets afforded parts per billion-level NO2 detection at room temperature. In another notable study, Guo et al. [33] developed a hierarchical porous ZIF-67-derived Co3O4/In2O3 heterostructure, exhibiting ultra-sensitive and selective acetone sensing performance.
These findings highlight that MOF-assisted synthesis strategies are a promising route
toward designing next-generation chemiresistive gas sensors with high sensitivity
and selectivity. In this study, we report on the rational design of MOF-derived Au-ZnO
nanostructures for ultrasensitive and selective H2S detection. ZIF-8 was employed as a precursor for synthesizing porous ZnO with a
large surface area, followed by Au decoration with controlled loading amounts. The
optimized Au-ZnO 2 sensor demonstrated an extraordinary response of ~419 at 10 ppm
H2S at 200 °C, which is nearly two orders of magnitude higher than that obtained using
pristine ZnO. Importantly, the sensor exhibited reliable sub-ppm-level detection down
to 0.2 ppm, excellent selectivity against common interfering gases, and fast response
and recovery dynamics. Moreover, comprehensive characterization via X-ray diffraction
(XRD), scanning electron microscopy (SEM), transmission electron microscopy (TEM),
and X-ray photoelectron spectroscopy (XPS) was used to elucidate the synergistic roles
of Au catalytic activity, spillover effect, oxygen vacancy modulation, and Schottky
barrier formation in imparting excellent sensing behavior.
2. EXPERIMENTAL
2.1 Materials
Zn(NO3)2·6H2O (≥99.0%), 2-methylimidazole (99%), and HAuCl4·3H2O (≥99.9%) were used as metal and ligand precursors. Methanol (≥99.9%, high-performance
liquid chromatography grade) was used as the solvent during synthesis, and hydrazine
hydrate (50–60%) served as the reducing agent for Au nanoparticle deposition. All
chemicals were purchased from Sigma-Aldrich (Germany) and were of analytical grade;
they were used without further purification. Moreover, deionized water was employed
as the solvent for all water-based processes.
2.2. Synthesizing MOF-derived ZnO
ZIF-8 was synthesized via a solution-based method. First, 0.03 mol of Zn(NO3)2·6H2O was dissolved in 250 mL of methanol and stirred for 30 min (solution A). Separately,
0.06 mol of 2-methylimidazole was dissolved in 250 mL of methanol and stirred for
30 min (solution B).
Solution B was added to solution A and stirred for an additional 30 min. The resulting
mixture was then left at room temperature for 24 h to crystallize. The obtained white
precipitate was washed three times with methanol and subsequently dried at 70 °C for
12 h. The dried ZIF-8 powder was collected and placed in an alumina boat and calcined
at 500 °C for 3 h in a tubular furnace in air at a heating rate of 3 °C/min. The final
product obtained was MOF-derived ZnO.
2.3. Synthesizing Au-ZnO
To prepare Au-ZnO, 0.1 g of ZnO powder was dispersed in 10 mL of deionized water via
sonication for 30 min. Subsequently, various weights of HAuCl4·3H2O (0.04–0.09 g) were added to the dispersion, and the mixture was stirred for 2 h
to facilitate the adsorption of Au precursors onto the ZnO surface. The resulting
solid was washed five times with deionized water to remove any unbound species. The
washed precipitate was subsequently redispersed in 20 mL of deionized water, followed
by the addition of hydrazine hydrate (0.2–0.4 mL). The mixture was stirred for an
additional 2 h to reduce Au3+. After the reaction, the purple-colored product obtained was washed twice with deionized
water and subsequently dried at 70 °C for 12 h to yield the Au-ZnO nanostructures.
To simplify notations, the undecorated sample is henceforth referred to as ZnO, and
the Au-decorated samples are named Au-ZnO 1, Au-ZnO 2, and Au-ZnO 3. To control the
Au loading, three samples were prepared using HAuCl4·3H2O precursor concentrations of 0.5, 1.0, and 2.0 mM, which were designated as Au-ZnO
1, Au-ZnO 2, and Au-ZnO 3, respectively. These precursor levels correspond to actual
Au contents of approximately 0.32, 0.97, and 1.84 wt%. The obtained powders were washed
and dried before further characterization.
2.4. Materials characterization
The morphologies of the nanostructures were examined using field-emission SEM (Regulus
8220, Hitach Ltd., Tokyo, Japan) and TEM (Talos F200X, Thermo Fisher Scientific, USA).
Elemental compositions and distributions were analyzed via energy-dispersive X-ray
spectroscopy (EDS) and high-angle annular dark-field scanning TEM. The crystal structures
and phases of the samples were identified via XRD (Empyrean, Panalytical, Netherlands),
and the surface chemical states were investigated via XPS (NEXSA, Thermo Fisher Scientific,
USA).
2.5. Gas sensing measurements
The synthesized nanostructures were dispersed in deionized water and drop-cast onto
an alumina substrate with Au electrodes. After drying, the sensor was placed in a
test chamber, in which target gases were introduced by mixing high-purity gases with
synthetic air using mass flow controllers. The total flow rate was fixed at 500 sccm.
The electrical resistances of the sensors were recorded under ambient air (Ra) and target gas (Rg) conditions. The sensor responses were defined as Rg/Ra and Ra/Rg corresponding to oxidizing and reducing gases, respectively. Each sensing cycle involved
500 s of exposure to the target gas, followed by 1000 s of recovery in air. Response
and recovery times were defined as the times required for the resistance to reach
90% of the final value after gas exposure and for the resistance to return to 90%
of the baseline value after gas removal, respectively.
3. RESULTS AND DISCUSSION
3.1. Morphological and chemical studies
The XRD patterns of the pristine ZnO and Au-ZnO nanostructures are shown in Figure 1. All samples yielded sharp diffraction peaks at 2θ ≈ 31.7°, 34.4°, and 36.2°, corresponding
to the (100), (002), and (101) planes of wurtzite ZnO (JCPDS No. 36-1451), respectively,
confirming that the samples were highly crystalline. Additional peaks were detected
at 38.2° and 44.4° in the XRD patterns of the Au-decorated samples, assignable to
the Au (111) and Au (200) planes, respectively, indicating the successful incorporation
of metallic Au nanoparticles. Interestingly, the relative intensity of the Au (111)
to ZnO (101) reflections increased with the increase in Au loading, suggesting that
Au decoration slightly perturbed the ZnO crystallite orientation.
SEM analysis (Figure 2) revealed that pristine ZnO contained porous aggregates with a uniform morphology,
a characteristic inherited from the ZIF-8 precursor. Au decoration increased surface
roughness and introduced nanoscale particulates owing to the nucleation of Au nanoparticles.
Figure 3a–c compares the surface morphologies of ZnO and Au-ZnO composites.
While the pristine ZnO (Figure 3a) exhibits smooth hexagonal facets, Au decoration induces a progressive increase in
surface roughness and particle density (Figures 3b–c). In particular, the Au-ZnO 2 sample (Figure 3c) shows densely decorated nanoparticles on the ZnO surface, indicating that moderate
Au loading promotes uniform dispersion without agglomeration. These morphological
changes are consistent with the TEM and elemental mapping results (Figures 3d–k), confirming that Au nanoparticles are anchored on the ZnO surface and contribute
to enhanced surface area and reactivity. All SEM and TEM micrographs presented in
this work include clearly visible scale bars.
Figure 4a shows the XPS survey spectra of pristine ZnO and Au-ZnO composites, confirming the
presence of Zn, O, and Au without any extraneous elements. Figure 4b–d present the Zn 2p spectra, in which the 2p3/2 and 2p1/2 peaks located at 1021.6 eV and 1044.7 eV correspond to the Zn²⁺ oxidation state of
wurtzite ZnO. Figure 4e–g displays the O 1s spectra deconvoluted into lattice oxygen (O_latt ≈ 530.1 eV) and
chemisorbed/vacancy-related components (O_ads ≈ 531.5 eV). The O_latt fraction shows
the most pronounced increase after Au decoration, whereas changes in O_ads and O_v
are relatively minor, indicating slight surface re-oxidation while retaining sufficient
reactive sites for gas adsorption. Figure 4h shows the Au 4f spectrum of the Au-ZnO 2 sample, exhibiting two main peaks at ~84.0
eV (Au 4f7/2) and ~87.7 eV (Au 4f5/2), characteristic of metallic Au⁰ [34]. A weak shoulder near 84.5 eV indicates minor Zn–Au interfacial bonding and partial
charge transfer. Figure 4i summarizes the quantitative O_ads/O_latt ratios and Au atomic fractions from XPS
fitting, confirming successful Au reduction and surface modification after Au incorporation.
3.2. Gas sensing studies
The chemiresistive sensing performances of the pristine ZnO and Au-ZnO nanostructures
were systematically evaluated under H2S exposure (Figure 5). As listed in Table 1, pristine ZnO exhibits only a moderate response of 4.84 (Ra/Rg) at 10 ppm H2S and 200 °C, reflecting the low density of active oxygen species and the sluggish
reaction kinetics. In contrast, Au decoration dramatically increased the responses
of the samples. Specifically, Au-ZnO 2 achieved an extraordinary response of 419.1
under the same conditions; this response is nearly two orders of magnitude higher
than that obtained from pristine ZnO and far exceeding those obtained from Au-ZnO
1 (50.6) and Au-ZnO 3 (130.4). This observation clearly indicates that an optimum
Au loading exists that maximizes catalytic activity and interfacial electronic sensitization
while avoiding excessive particle aggregation or site blocking that can occur at high
Au concentrations. The data summarized in Table 1 show that the sensing responses of all samples increase with operating temperature
and reach a maximum at 200 °C, where Au-ZnO 2 exhibits the highest response among
the series, confirming its optimal Au loading and catalytic activation.
The concentration-dependent behavior of Au-ZnO 2 revealed a unique dual-regime trend.
At low H2S levels (< 6 ppm), the response increased exponentially with the increase in H2S concentration, consistent with catalyst-dominated amplification, in which the presence
of abundant Au active sites and oxygen vacancies enables efficient adsorption and
rapid spillover reactions. At high H2S levels (> 6 ppm), the response transitioned into a quasi-linear regime, indicative
of Langmuir-type adsorption kinetics, in which surface saturation began to dominate.
This dual behavior clarifies the underlying surface chemistry and demonstrates the
sensor’s ability to maintain proportionality in its response across both trace and
high concentrations of the target gas, a feature highly desirable in real-world monitoring
systems.
The concentration-dependent behavior listed in Table 2 shows a clear transition from an exponential increase at low H2S levels to a quasi-linear regime at higher concentrations, highlighting the dual-regime
adsorption kinetics that govern the sensing process.
The dynamic resistance profiles of Au-ZnO 2 further confirmed its superior performance.
At 1 ppm H2S and 200 °C, the device displayed response and recovery times of 472 s and 345 s,
respectively (Table 3), while still delivering reproducible signals at H2S concentrations as low as 0.2 ppm. Such sub-ppm detection capabilities highlight
the utility of the MOF-derived architecture and noble-metal-assisted surface reactions
in achieving ultrasensitive gas detection. Importantly, repeated sensing cycles yielded
negligible signal degradation and baseline drift, confirming the robustness of the
heterostructure under prolonged operation. The results presented in Table 3 further show that both the response and recovery times of Au-ZnO 2 decrease gradually
with increasing H2S concentration, indicating accelerated surface reaction dynamics and efficient charge
transfer under higher analyte exposure.
Selectivity is another critical requirement for practical gas sensors, as cross-sensitivity
to interfering gases often reduces reliability in complex environments. To address
this, we examined the response of Au-ZnO 2 to a range of common interfering gases,
including NO2, H2, NH3, C2H5OH, C6H6, and C7H8 (each at 10 ppm and 200 °C). The results (Figure 6) revealed that the sensor exhibited considerably higher response to H2S than to all other analytes, for which negligible responses were observed. This exceptional
selectivity is attributable to the strong affinity of H2S for Au catalytic sites, coupled with its facile dissociation pathway, unlike in
the case of other gases [16,34]. The unique interaction of H2S with surface oxygen species in the presence of Au further amplified this selective
behavior [17,35].
Overall, these results demonstrated that Au-ZnO 2 exhibited ultrahigh sensitivity,
with a record response of 419.1 at 10 ppm of H2S and achieved reliable sub-ppm detection, fast dynamics, and excellent selectivity.
These characteristics suggest that MOF-derived Au-ZnO can be used as a promising next-generation
sensing platform, capable of satisfying stringent industrial and environmental monitoring
requirements.
3.3. Gas sensing mechanism
3.3.1 Catalytic effect of Au nanoparticles
Before introducing the catalytic mechanism of Au, it should be noted that pristine
ZnO itself exhibits poor sensitivity toward H2S owing to its limited surface reactivity and slow redox kinetics [13,14]. The enhanced H2S sensing performance of the Au-ZnO composites mainly originates from the catalytic
activation of Au nanoparticles [36,37]. As illustrated in Figure 7a, Au acts as an active site for the dissociative adsorption of H2S into HS⁻ and S species [19,35]. These intermediates form Au–S bonds and subsequently migrate to the ZnO surface,
where they react with pre-adsorbed oxygen ions (O_ads) to release electrons into the
conduction band [38-40]. This spillover-driven catalytic process lowers the activation barrier and accelerates
surface redox kinetics, yielding a pronounced response enhancement at 200 °C. Figure 7b schematically depicts this Au-assisted dissociative adsorption and spillover reaction
on the ZnO framework.
3.3.2 Schottky barrier and charge transfer
In addition to catalytic effects, the work-function difference between ZnO (φ ≈ 4.3
eV) and Au (φ ≈ 5.1 eV) induces a Schottky junction at their interface [18,41,42]. In air, electrons transfer from ZnO to Au until Fermi-level equilibrium is achieved,
leading to upward band bending and formation of an electron-depletion layer. During
H2S exposure, reaction-released electrons lower the barrier height and consequently
reduce the sensor resistance [43,44]. This modulation of interfacial charge transport accounts for the large and reversible
resistance variation observed for Au-ZnO 2. The corresponding band-structure evolution
during gas exposure is illustrated in Figure 7c, showing how electron transfer and barrier modulation drive the high sensitivity.
3.3.3 Role of oxygen vacancies and adsorbed oxygen
Oxygen vacancies (O_v) and chemisorbed oxygen species (O_ads) also play essential
roles in sustaining surface reactivity and charge equilibrium. A moderate O_v concentration
facilitates adsorption of reactive oxygen without excessive carrier trapping, while
Au decoration stabilizes these oxygen-related states, as supported by the O 1s XPS
spectra (Figure 4c). These surface oxygen defects promote efficient electron exchange between ZnO and
Au, synergistically reinforcing catalytic activation and electronic sensitization.
Figure 7d schematically summarizes this combined mechanism, in which Au catalysis, Schottky-barrier
modulation, and oxygen-defect chemistry collectively yield the enhanced response,
selectivity, and reversibility of the Au-ZnO sensors.
5. CONCLUSIONS
In this study, MOF-derived Au-ZnO nanostructures were successfully synthesized using
ZIF-8 as a sacrificial template, and the samples were subsequently investigated as
highly efficient chemiresistive gas sensors for H2S detection. The combination of a porous ZnO framework and optimally loaded Au nanoparticles
remarkably enhanced sensing performance relative to pristine ZnO. In particular, Au-ZnO
2 exhibited an extraordinary response of ~419.1 to 10 ppm of H2S at 200 °C, nearly two orders of magnitude higher than that obtained using bare ZnO.
Furthermore, the sensor reliably detected sub-ppm H2S concentrations down to 0.2 ppm and exhibited reproducible response-recovery cycles
and negligible signal drift.
Comprehensive structural and surface analyses confirmed that Au nanoparticles were
uniformly anchored on the ZnO surface and established strong interfacial coupling.
This dual modification, i.e., structural porosity obtained from the MOF precursor
and electronic sensitization obtained from Au decoration, was demonstrated to enrich
oxygen vacancies, increase chemisorbed oxygen content, and generate Schottky junctions
at the Au-ZnO interfaces. These features collectively enabled the rapid catalytic
dissociation of H2S, accelerated spillover reactions, and amplified charge transfer across the heterointerface.
The observed exponential-to-linear transition in the concentration-dependent response
further highlights the dual-regime adsorption kinetics governed by catalytic amplification
at trace H2S levels and Langmuir-type site saturation at high H2S concentrations.
Importantly, the Au-ZnO 2 sensor displayed excellent selectivity against common interfering
gases, such as NO2, NH3, H2, and volatile organics, underscoring the reliability of the sensor for operation
in complex environments. Benchmarking with recent high-impact studies revealed that
the performance of Au-ZnO 2 is competitive, surpassing many state-of-the-art H2S sensors in terms of response magnitude, sub-ppm sensitivity, and operational stability.
Overall, this study demonstrates a rational design strategy that integrates MOF-derived
pore formation with noble-metal functionalization to produce ultrasensitive, selective,
and stable gas sensors. The combined effect of catalytic activity, interfacial charge
modulation, and defect chemistry provides clear guidelines for advancing next generation
chemiresistive sensors. In addition to H2S detection, the proposed approach can be used to detect other hazardous gases and
scaled for application in industrial monitoring systems, facilitating practical applications
in environmental safety and occupational health.