Zinc oxide (ZnO) is a typical n-type MOS with a wide bandgap of 3.37 eV, high exciton binding energy, high electron mobility, strong ultraviolet absorption, and good chemical stability^[75-77]. These features make ZnO one of the most widely studied sensing materials. At room temperature, the H₂S response of pristine ZnO is generally limited by insufficient surface reaction kinetics, therefore, nanostructure engineering is essential for increasing active adsorption sites and shortening diffusion pathways^[78,79]. Wang et al. fabricated ZnO nanorod thick films by a hydrothermal method and controlled the ZnAc₂/NaOH ratio to regulate morphology and selectivity^[12]. The resulting sensor showed preferential ethanol response at 350 °C but selective H₂S detection at room temperature, with R_a/R_g = 1.7^[12], whereas vapor-grown vertically aligned nanorods formed a porous flower-like network that increased exposed area^[78]. Other comb-like, prismatic, and belt-like ZnO architectures likewise demonstrate that anisotropic and porous structures can improve H₂S accessibility^[6,80,81].

Overall, ZnO enables room-temperature H₂S detection, but its performance remains limited by sluggish oxygen-mediated surface reactions, highlighting the importance of defect and nanostructure engineering.

Tungsten trioxide (WO₃) is a chemically stable n-type metal-oxide semiconductor that has been extensively investigated for H₂S sensing^[82-84]. Its H₂S response mainly originates from oxygen-mediated surface reactions that modulate carrier concentration and conductivity. Nanostructured WO₃ films and colloidal quantum dots have shown improved low-temperature H₂S sensing owing to increased surface exposure and enhanced surface-controlled charge transport^[11,85]. Hoel et al. further reported a 250-fold increase in conductivity toward 5 ppm H₂S at room temperature using WO₃ nanoparticles^[8]. These results indicate that particle sizes approaching the Debye length are beneficial.

Pristine WO₃ remains limited by sluggish room-temperature activation and incomplete recovery. Its sensing performance is strongly governed by the relationship between particle size and Debye length, together with W⁵⁺/W⁶⁺ redox states and oxygen-vacancy distributions.

Copper oxide (CuO) is a narrow bandgap p-type semiconductor and one of the most effective room-temperature H₂S sensing materials^[10,86]. In addition to oxygen-mediated reactions, H₂S can form conductive copper sulfide species and thus create additional conduction paths^[10,87,88]. Porous CuO nanosheets and biotemplated tubular structures have demonstrated that accessible surface sites, shortened diffusion pathways, and Cu/CuS-related conductive pathways can improve room-temperature H₂S sensing^[10,88]. More notably, sea-anemone-like CuO nanoarrays achieved a response of 24.08 toward 5 ppb H₂S with negligible responses to common interfering gases^[87]. Their high sensitivity was attributed to nanotentacle diameters approaching the Debye length and point-to-point contacts that amplified modulation of the hole-accumulation layer, while the preferential reaction of H₂S with chemisorbed oxygen contributed to selectivity. These results highlight the importance of coupling gas-accessible CuO architectures with Debye-length-scale conduction channels.

Despite its high H₂S sensitivity, pristine CuO often suffers from incomplete recovery, baseline drift, and poor reversibility because of H₂S-induced sulfide formation. Competitive water adsorption can further destabilize its response and baseline. Although MEMS and surface-acoustic-wave platforms improve power efficiency and portability^[89,90], the central challenge remains controlling reversible CuO/Cu_xS conversion while maintaining strong H₂S affinity and humidity resistance.

Indium oxide (In₂O₃) is a typical n-type MOS with high electrical conductivity, good chemical stability, and abundant oxygen-vacancy-related active sites. Early studies demonstrated the evolution of In₂O₃-based H₂S sensors from high-temperature nanocrystalline materials to porous nanotubes and nanowires capable of room-temperature operation^[91,92]. The improved performance of the latter was associated with their gas-accessible one-dimensional architecture and H₂S-induced surface sulfuration.

More importantly, ordered porous In₂O₃ ultrathin films exhibited an ultrahigh room-temperature response through a proposed humidity-assisted mechanism. Ambient water promoted H₂S hydrolysis, chemisorbed oxygen desorption, and the formation of a conductive surface water layer, thereby markedly increasing sensor conductance^[93]. Porous In₂O₃ nanoparticles achieved a response of 26,268.5 toward 1 ppm H₂S with a detection limit of 1 ppb at room temperature, while UV-generated holes were proposed to promote oxygen desorption and narrow the electron-depletion layer, thereby regulating surface conductivity^[59]. In₂O₃ nanocubes, colloidal quantum dots, and oxidized thin films have also demonstrated that crystal size, ligand-mediated surface states, and film structure strongly influence H₂S adsorption and carrier transport^[94-96]. These findings identify porous architecture, sulfuration, surface oxygen regulation, and humidity-assisted reactions as the principal factors governing room-temperature H₂S sensing by In₂O₃.

For practical room-temperature operation, In₂O₃-based H₂S sensors still require further improvement in reversibility, humidity calibration, and quantitative reliability. The sulfuration of In₂O₃ and humidity-induced H₂S hydrolysis can greatly amplify the sensing signal, but they may also cause baseline drift and recovery instability if not well controlled. Future studies should therefore clarify the coupled roles of surface oxygen species, hydroxyl groups, In₂O₃/In₂S₃ conversion, and surface-state-regulated charge transport. In addition, ligand engineering in In₂O₃ quantum dots and thickness/orientation control in In₂O₃ thin films may provide effective routes toward fast, stable, and low-power room-temperature H₂S detection.

As summarized in Table 1, single-material sensors achieve room-temperature H₂S detection mainly through increased surface accessibility and enhanced charge modulation. Porous architectures and dimensions approaching the Debye length increase active-site exposure and amplify surface-induced conductivity changes. However, their sensing pathways remain material-dependent. ZnO and WO₃ rely primarily on oxygen-mediated surface reactions and are limited by sluggish kinetics. CuO achieves high sensitivity through H₂S-induced CuO-to-Cu_xS conversion, but excessive sulfuration compromises recovery and repeatability. In₂O₃ combines oxygen-mediated reactions, humidity-assisted hydrolysis, and partial sulfuration, producing high responses but also pronounced humidity dependence and baseline instability. Future priorities should therefore focus on improving surface reactivity in ZnO and WO₃, balancing H₂S affinity with reversible sulfuration in CuO, and regulating humidity-related reactions in In₂O₃. These material-specific challenges must be addressed to balance response, recovery, selectivity, humidity tolerance, and long-term stability.

Functionalized graphene and rGO have been widely investigated as functional components in H₂S sensors because their high specific surface area and favorable charge-transport properties can promote gas adsorption and transduce adsorption-induced charge transfer into measurable electrical signals^[25,119,120]. Rather than acting as passive supports, they can serve as conductive scaffolds, nanocrystal-growth templates, and interfacial charge-transfer media. In particular, the residual oxygen-containing groups and structural defects of rGO provide adsorption and anchoring sites, while its conductive network facilitates carrier transport between semiconductor components. In oxide/rGO composites, these effects can be coupled with heterojunction-barrier modulation and changes in surface oxygen adsorption, thereby amplifying H₂S-induced resistance variations. The resulting performance depends strongly on the reduction degree, defect density, layer stacking, and interfacial contact of rGO.

Early studies showed that functionalized graphene could regulate the nucleation and dispersion of Cu₂O nanocrystals while providing conductive pathways for room-temperature H₂S sensing^[26]. Liu et al. subsequently constructed a well-defined Cu₂O/graphene heterojunction that achieved responses of 24,125.8% toward 500 ppb H₂S and 386% toward 5 ppb H₂S at 20 °C^[19]. Other graphene/oxide composites have further demonstrated that rGO can improve oxide dispersion, adsorption-site accessibility, and carrier transport, while dopants, hollow structures, and heterointerfaces provide additional pathways for H₂S activation and resistance modulation^[21,22,27]. More recently, Amith et al. integrated an rGO-Er₂O₃/Co₃O₄ composite with a low-cost disc electrode^[114]. The optimized sensor achieved a response of 75% toward 10 ppm H₂S, response and recovery times of 47 and 11 s, and stability exceeding 100 days^[114]. These studies show that graphene-based modification has evolved from simple conductivity enhancement toward coordinated control of oxide growth, charge transport, and heterointerfacial properties. However, reliable performance still requires precise regulation of rGO reduction, defect density, layer restacking, humidity sensitivity, and graphene/oxide interface quality.

Comparable improvements in response magnitude and response/recovery kinetics have also been reported for SnO₂ quantum-wire/rGO, CeO₂/graphene, and WO₃/rGO composites, further supporting the role of graphene in facilitating carrier transport and interfacial resistance modulation^[24,25,121].

TMDs, particularly MoS₂ and WS₂, have attracted widespread attention in the field of room-temperature H₂S sensing due to their semiconductor layered structures, which feature tunable bandgaps, abundant edge sites, and defect-sensitive surface chemistry^[30,122]. Kundu et al. demonstrated that the response of liquid-phase-exfoliated WS₂ nanosheets was strongly dependent on gas flow rate, highlighting the importance of mass transport and testing conditions for ultrathin sensing layers^[9]. MoS₂-GO hybrids further achieved a response of 39.16% toward 100 ppm H₂S at 28 °C with a response time of 6.13 s^[123]. Coupling TMDs with metal oxides provides additional interfacial amplification. Porous CuO/MoS₂ p-n heterostructures improved H₂S response and selectivity through increased surface accessibility and interfacial charge modulation^[18]. Huo et al. constructed In₂O₃@MoS₂ heterojunction fibers that exhibited a response of 460.61 toward 50 ppm H₂S and a detection limit of 3 ppb at room temperature^[28]. Beyond MoS₂, MOF-derived α-Fe₂O₃ hollow nanospheres decorated with MoSe₂ nanoflowers achieved a response of 57.7 toward 30 ppm H₂S with response and recovery times of 50 and 53 s^[124]. These studies demonstrate that TMD-based heterostructures combine edge- and defect-mediated adsorption with oxide-supported gas diffusion and junction-induced charge modulation. Further progress requires reproducible control of layer stacking, defect density, interfacial structure, humidity interference, and room-temperature desorption.

MXenes, particularly Ti₃C₂T_x, are attractive for low-power H₂S sensing because of their high electrical conductivity, chemically active surface terminations, and tunable interfacial properties. Recent studies have extended their role beyond conductive additives by integrating molecular enrichment, termination-dependent adsorption, and external activation^[125-127]. Ding et al. assembled Co-MOF nanoparticles on Ti₃C₂T_x nanosheets to form a 0D-2D heterostructure in which the porous MOF enriched H₂S molecules and the MXene provided rapid electron transport^[125]. The sensor achieved a response of 11.1 toward 400 ppb H₂S with an experimental detection limit of 50 ppb, demonstrating the value of coupling molecular preconcentration with conductive transduction^[125]. Ti₃C₂T_x-organic composites further preserved the inherent H₂S selectivity of MXene while producing a 30-fold enhancement in sensing response^[128]. DFT calculations indicated that mixed O, OH, and F terminations strongly influenced H₂S adsorption and charge transfer, emphasizing that termination composition is a critical determinant of MXene sensing behavior^[128]. Han et al. developed an near-infrared (NIR)-activated MXene/PbS heterostructure that achieved a response of 90% toward 10 ppm H₂S, response and recovery times of 2 and 89 s, and stable operation for up to 60 days [Figure 2A]^[126]. Its enhanced performance was attributed to improved NIR absorption and carrier separation within the heterostructure, together with photothermal acceleration of surface reaction and desorption kinetics. These studies show that MXene-based sensors are evolving into multifunctional platforms that integrate adsorption, charge transport, and external activation. Practical application, however, still requires improved control over oxidation, surface-termination stability, layer restacking, humidity interference, and batch-to-batch reproducibility.

2D COFs are crystalline porous polymers with molecularly tunable pores, recognition sites, π-conjugated backbones, and charge-transport properties. Meng et al. developed the intrinsically conductive nickel-phthalocyanine COF Ni-COF-DC-8 with a bulk conductivity of 2.51 × 10^-3 S m^-1[40]. The corresponding chemiresistive sensor exhibited a response of 62% toward 40 ppm H₂S and a theoretical detection limit of 204 ppb, with the signal attributed to charge transfer between H₂S and the nickel-phthalocyanine units^[40]. Benedetto et al. subsequently constructed an array of isostructural M-COF-DC-8 materials containing Fe, Co, Ni, or Cu centers. Cu-COF-DC-8 achieved the highest response of approximately 99% toward 80 ppm H₂S and a theoretical detection limit of 28 ppb in dry N₂, while the array enabled discrimination among H₂S, NO, CO, and NH₃ under both dry and humid conditions [Figure 2B]^[39]. Beyond chemiresistive sensing, a nanoporous COF optical-waveguide film achieved a response time below 2 s and a detection limit of 1.07 ppb through proton transfer at triazine-containing sites and the resulting refractive-index change^[41]. These studies demonstrate that COFs can regulate both molecular recognition and signal transduction through framework-level design. Their practical development, however, requires improved conductivity, crystallinity, film uniformity, reversibility, and humidity stability.

In summary, graphene/rGO, TMDs, MXenes, and 2D COFs share high surface accessibility and short gas-diffusion pathways, but differ substantially in their sensing functions. Graphene and rGO mainly provide conductive networks and heterointerfacial charge transport, while their performance is sensitive to reduction degree, layer restacking, and humidity-induced drift. TMDs contribute semiconducting channels, edge sites, and chalcogen vacancies for H₂S adsorption and interface modulation, although defect reproducibility and slow desorption remain concerns. MXenes combine high conductivity with termination-dependent adsorption and interfacial chemistry, but oxidation and batch-dependent surface composition limit stability. COFs offer molecularly designable pores and coordination sites for selective H₂S recognition, while insufficient conductivity and poor film uniformity constrain electrical transduction. Coupling these materials with metal oxides or other functional phases can integrate gas accessibility with interfacial charge modulation, but the resulting enhancement depends strongly on defect density, layer stacking, surface chemistry, and interface quality. Future studies should therefore prioritize reproducible material preparation, operando analysis of interface evolution, and standardized evaluation under humid and long-term operating conditions.

Pristine and framework-retaining MOFs provide permanent porosity, tunable pore environments, and accessible metal sites for H₂S enrichment and molecular recognition. Luo et al. compared MIL-100(Fe), HKUST-1(Cu), and bimetallic MOF-919(Fe-Cu), among which MOF-919 exhibited the highest response of 931.6% toward 10 ppm H₂S and a detection limit of 0.31 ppm at room temperature^[44]. The improvement was attributed to suitable pore dimensions and cooperative Fe/Cu adsorption sites, demonstrating that metal-site chemistry is as important as surface area in determining MOF sensing behavior. A porous Cu-MOF was also employed as the sensing electrode in a K₂Fe₄O₇-based mixed-potential sensor, achieving a detection limit below 5 ppb at room temperature. Electrode passivation extended stable operation to 66 cycles by suppressing the accumulation of sulfur-containing products^[43]. In a different transduction mode, Cu-modified MOF-808 enabled reversible colorimetric detection of 100 ppm H₂S through Cu²⁺/Cu⁺ redox conversion, with response times of 1.3-2.2 min and thermal regeneration under moist air^[45]. These studies demonstrate that intact MOFs can combine pore-confined enrichment and metal-site chemistry with electrical, mixed-potential, or optical readout. Their practical use remains constrained by low intrinsic conductivity, humidity-dependent adsorption, slow desorption, and sulfur-induced deactivation of active metal sites.

Unlike conventional MOFs with limited intrinsic conductivity, conductive metal-organic frameworks (cMOFs) combine permanent porosity and accessible coordination sites with framework-mediated charge-transport pathways, enabling direct chemiresistive transduction of gas-framework interactions^[42,130,131]. Two-dimensional cMOFs based on square-planar metal nodes and π-conjugated ligands are particularly attractive because charge transport can proceed through in-plane π-d conjugation and interlayer π-π stacking. Jeon et al. compared six 2,3,6,7,10,11-hexasubstituted triphenylene-based metal-organic frameworks (M-HXTP; M = Co, Ni, or Cu) and found that the Cu-containing cMOFs exhibited the strongest H₂S responses^[42]. Spectroscopic characterization and DFT calculations linked the resistance modulation to H₂S oxidation, Cu reduction, and distortion of the conductive CuX₄ units. Zhao et al. further tuned the metal-node composition by constructing bimetallic Co_1.8Ni_1.2(HITP)₂, which showed a 7.5-fold higher response toward 5 ppm H₂S than Ni₃(HITP)₂^[47]. The sensor maintained a signal drift below 4.13% over 60 days in oxygen-deficient SF₆, indicating that metal-node engineering can regulate H₂S affinity, oxygen-independent charge transfer, and operational stability.

Beyond metal-node and ligand selection, 2D cMOF sensing can be regulated through morphology, layer stacking, and film processing. Solution-processable Ni₃(HITP)₂/NUS-8 MOF-on-MOF nanosheets (HITP: 2,3,6,7,10,11-hexaiminotriphenylene; NUS: National University of Singapore) enabled the fabrication of large-area, thickness-controlled films, as confirmed by the cross-sectional scanning electron microscope (SEM) images in Figure 3A₁-A₄. Kelvin probe force microscopy (KPFM) revealed different surface-potential contrasts for NUS-8 and Ni₃(HITP)₂/NUS-8 [Figure 3A₅ and A₆], consistent with interfacial charge redistribution. Following exposure to 10 ppm H₂S, the S 2p and N 1s XPS changes and the electron paramagnetic resonance (EPR) signal at g ≈ 2.0 supported the formation of sulfur-containing species, modification of the N coordination environment, and generation of paramagnetic centers; the corresponding redox sensing mechanism is illustrated in Figure 3A₇-A₁₀. The resulting sensor achieved a room-temperature detection limit of approximately 6 ppb^[46]. Zeolitic imidazolate framework-67 (ZIF-67)-templated cobalt–2,3,6,7,10,11-hexahydroxytriphenylene (Co-HHTP) architectures improved electrical conduction, Co-site accessibility, and gas diffusion, achieving detection limits of 44 and 34 ppb for nanoparticles and nanosheets, respectively^[132]. Their response mainly arose from H₂S binding at exposed Co sites, where Co-S interactions reduced the effective hole concentration and increased resistance. 1,4,5,8,9,12-hexaazatriphenylenehexacarbonitrile (HATCN) intercalation between Cu₃(HHTP)₂ layers further introduced out-of-plane transport pathways and electron-deficient adsorption environments, enabling selective H₂S detection down to 120 ppb^[133]. These studies demonstrate that morphology, active-site accessibility, interlayer coupling, and scalable film formation are critical to the sensing performance of 2D cMOFs.

Conductive-MOF heterostructures couple framework-mediated transport with the adsorption and electronic properties of secondary semiconductors. Sun et al. decorated Co₃(HITP)₂ with SnO₂ nanoparticles, producing a flexible sensor with a response of 3.3 toward 50 ppm H₂S and a detection limit of 160 ppb at room temperature under 25% RH^[134]. DFT and band-structure analyses attributed the enhancement to stronger H₂S adsorption, interfacial charge transfer, and heterojunction-mediated carrier modulation. Key limitations include anisotropic transport, layer stacking, humidity-dependent conductivity, limited film processability, and partially irreversible reactions at redox-active metal sites.

MOF-derived sensing materials are produced by converting MOF precursors into metal oxides, carbon-containing phases, or multicomponent heterostructures. Although the coordination framework is destroyed, the derivatives can retain the precursor morphology, porosity, and homogeneous metal distribution while generating small crystallites, defects, and accessible diffusion pathways. Jiang et al. converted CuBDC into hollow rod-shaped CuO, which exhibited a response of 330 toward 1 ppm H₂S at 20 °C and a theoretically estimated detection limit in the ppt range^[135]. The enhancement was attributed to the porous hollow architecture, optimized oxygen-vacancy concentration, and strong CuO-H₂S interaction. This example demonstrates that MOF templating can suppress oxide aggregation and preserve gas-accessible internal structures.

MOF precursors enable compositionally uniform heterojunctions by preserving metal distribution and porous architectures during conversion. Ding et al. derived porous NiO-CuO structures containing multiple p-p junctions from a bimetallic MOF^[136]. The sensor exhibited responses of 3.99 and 11.0 toward 1 ppm H₂S in dry air and at 40% RH, respectively, with a detection limit of 50 ppb^[136]. The enhancement was attributed to inherited porosity, increased adsorbed oxygen, and NiO/CuO interfacial regulation. Cai et al. further derived a CeO₂/N-doped carbon p-n heterojunction that achieved a detection limit of 16 ppb and stable sensing at 80% RH through the combined effects of a built-in electric field, oxygen vacancies, and a hydrophobic carbon matrix^[129].

Coupling MOF-derived oxides with two-dimensional materials can improve carrier transport and interface utilization. Zhang et al. dispersed MIL-88-derived γ-Fe₂O₃ octahedra on rGO, producing a response of 520.73 toward 97 ppm H₂S at room temperature^[137]. The rGO network promoted charge transport and limited oxide aggregation, while porous γ-Fe₂O₃ provided accessible reaction sites. MOF-derived α-Fe₂O₃ hollow nanospheres decorated with MoSe₂ nanoflowers exhibited response and recovery times of 50 and 53 s toward 30 ppm H₂S through improved gas diffusion, additional adsorption sites, and interfacial charge modulation^[124]. These studies show that MOF derivation can integrate porous structures with conductive networks and heterointerfaces. Remaining challenges include calcination-induced shrinkage, phase variability, uncontrolled defect concentrations, slow recovery, and humidity-dependent reactions.

MOF-based H₂S sensors follow three distinct design routes with different advantages and limitations. Pristine MOFs exploit permanent pores and open metal sites for H₂S enrichment and molecular recognition, but their low conductivity, humidity-sensitive adsorption, slow desorption, and sulfur poisoning restrict reusable electrical sensing. Conductive MOFs introduce framework-mediated charge transport for direct chemiresistive transduction, although anisotropic conductivity, dense layer stacking, limited film processability, and redox-induced signal irreversibility remain problematic. MOF-derived materials sacrifice the original coordination framework but retain precursor morphology and metal distribution, enabling porous oxides, carbon-containing phases, and heterostructures with abundant defects and interfaces. Their performance is instead limited by calcination-induced shrinkage, phase variability, uncontrolled defect populations, and slow room-temperature recovery. Future priorities should therefore focus on protecting and regenerating active sites in pristine MOFs, controlling layer orientation and film assembly in conductive MOFs, and achieving reproducible phase and defect regulation during MOF conversion. Operando studies under humid H₂S exposure are particularly important for resolving framework reactions, sulfur accumulation, and recovery pathways across these material classes.

Ag is attractive for H₂S sensing because of its electrical conductivity, catalytic activity, and strong affinity for sulfur-containing species. Ag nanowire/hollow polypyrrole nanotube composites achieved a detection limit of 10 ppb at room temperature^[55]. The proposed mechanism involved preferential H₂S adsorption and partial dissociation on Ag, followed by proton transfer to polypyrrole that increased its carrier concentration and conductivity. Ag-In₂O₃ nanorod composites further exhibited an ultrahigh response of 93,719 toward 20 ppm H₂S with a detection limit of 5 ppb^[58], while Ag incorporation into CeO₂/porous-Si films improved sensing stability and humidity tolerance^[140]. Lee et al. decorated TiO₂ nanosheets with Ag nanoparticles, increasing the response from 1.12 to 2.55 toward 20 ppm H₂S at 25 °C and improving selectivity over H₂, NH₃, and acetone^[138]. XPS analysis showed that the oxygen-vacancy-related and adsorbed-oxygen-related O 1s contributions increased from 16.18% to 26.14% and from 9.54% to 13.18%, respectively. Negative shifts of the Ag 3d peaks relative to bulk Ag further supported electron transfer from Ag to TiO₂. The proposed interfacial mechanism, including narrowing of the TiO₂ electron-depletion layer, defect-assisted electron transport, and modulation of inter-nanosheet potential barriers, is illustrated in Figure 3B^[138]. Ag-catalyzed H₂S dissociation, oxygen spillover, and possible silver-sulfide formation were additionally proposed to contribute to the enhanced response. These studies indicate that Ag sensitization can enhance H₂S sensing through sulfur-affinitive adsorption, catalytic oxygen activation, and support-dependent charge-transfer pathways.

Ag sensitization operates through three material-dependent pathways. Ag promotes O₂ and H₂S adsorption and activation, thereby increasing the reactivity of surface oxygen species^{[138,139,148]}. Ag/support charge redistribution also modifies carrier concentration and interfacial barriers, amplifying resistance changes^{[56,138,139,148]}. In addition, Ag-S affinity favors H₂S capture and the formation of Ag-S-related species in Ag/polymer and selected Ag/oxide systems^[55,138,148]. Their relative contributions depend on the Ag state, support, surface oxygen chemistry, and operating conditions. Although Ag-S interactions enhance affinity and selectivity, excessive sulfidation may impede desorption, baseline recovery, and long-term stability.

Au sensitization enhances room-temperature H₂S sensing through coupled catalytic and electronic effects. Au nanoparticles promote oxygen adsorption and H₂S surface reactions, while their high work function relative to many n-type oxides induces interfacial electron transfer and Schottky-type barriers, thereby amplifying resistance modulation under limited thermal activation.

Au decoration markedly improves room-temperature ZnO sensing. Flower-like Au/ZnO nanorods achieved a response of 1,270 toward 6 ppm H₂S, 3.7 times that of pristine ZnO, although their response and recovery remained slow^[50]. Au-modified ZnO nanowires reached a response of 79.4 toward 5 ppm H₂S, representing a 16-fold enhancement, and shortened the recovery time from 860 to 170 s^[54]. These results demonstrate strong Au sensitization, although recovery remains dependent on morphology and surface-reaction reversibility.

Deb et al. developed a flower-petal-like Au/SnO₂ sensor for ppb-level H₂S detection at room temperature. At 0.5 V and 24 ± 1 °C, the sensor achieved an experimental detection limit of 2 ppb^[48]. Upon exposure to 500 ppb H₂S, the response increased from 35% ± 7% for pristine SnO₂ to 250% ± 30% for Au/SnO₂, while the recovery time decreased from 510 to 126 s^[48]. Figure 3C₁ illustrates the proposed low-bias interfacial dipole effect. The O 1s XPS spectra show an increase in the oxygen-vacancy-related contribution from 26.49% to 37.49% after Au incorporation [Figure 3C₂ and C₃], while Figure 3C₄ illustrates the overall gas-sensing mechanism. However, the contribution of localized interfacial dipoles remains a plausible interpretation that requires further experimental verification^[48].

Overall, Au sensitization enhances H₂S sensing through catalytic oxygen activation and Au/oxide interfacial electronic modulation. Unlike Ag, whose enhancement often involves strong Ag-S interactions and surface sulfidation, Au systems more commonly rely on Schottky-barrier regulation, morphology-dependent surface exposure, and low-bias interfacial polarization. Future studies should clarify the evolution of Au/oxide interfaces during repeated H₂S exposure and improve recovery, humidity tolerance, and scalable low-voltage device fabrication without sacrificing ultralow detection limits.

Pt sensitization of room-temperature H₂S sensors primarily combines catalytic oxygen activation, electron extraction, and interfacial barrier modulation^{[52,57,142,143]}. Xuan et al. integrated Pt with etched W-doped ZnO nanotubes grown in situ on FTO electrodes. The optimized Pt(1.5%)-ZnO sensor exhibited a response of 16.59 toward 5 ppm H₂S and detected concentrations down to 100 ppb^[57]. Pt 4f binding-energy shifts and increased chemisorbed oxygen supported electron transfer from ZnO to Pt and enhanced oxygen activation. Based on this evidence and the Pt/ZnO work-function difference, the enhancement was attributed to oxygen-spillover-mediated chemical sensitization and nano-Schottky-barrier-mediated electronic sensitization. This study demonstrates the benefit of coupling Pt sensitization with defect regulation and integrated device architecture^[57].

Pt size, dispersion, and electronic state critically influence H₂S sensing. Kou et al. decorated TiO₂@ZnFe₂O₄ nanotube arrays with Pt nanoclusters, achieving a response of 16.6 toward 10 ppm H₂S and a detection limit of 0.43 ppb at room temperature^[52]. The highly dispersed Pt generated oxygen vacancies and Pt^δ+ species, enabling H₂S activation without heating or UV irradiation. Similarly, 0.3 wt.% Pt-SnO₂ mesoporous nanoflowers exhibited a response of 68 toward 1 ppm H₂S at 30 °C with a detection limit of 100 ppb^[143]. Their enhancement arose from efficient gas diffusion through the mesoporous structure and Pt-mediated catalytic sensitization and oxygen spillover.

Pt sensitization is progressing from simple nanoparticle decoration toward precise regulation of dispersion, cluster size, and valence state. Highly dispersed Pt species improve atomic utilization and electron extraction, enabling H₂S reactions under limited thermal activation. The central challenge is to stabilize these active states at low Pt loading while preventing aggregation and preserving catalytic and electronic sensitization during humid operation and repeated H₂S exposure.

Noble-metal functionalization has expanded from metal oxides to two-dimensional MXenes, TMDs, and graphene-based sensing layers. Their exposed surfaces, defects or terminations, and in-plane conductive pathways favor metal dispersion and amplify adsorption-induced electrical changes. Noble metals anchored at surface terminations, chalcogen vacancies, or other defect sites can provide catalytic and sulfur-affinitive centers, redistribute local carriers, and modulate contact resistance, thereby coupling H₂S adsorption more effectively to room-temperature electrical transduction.

MXenes are attractive noble-metal supports because their high conductivity and active surface terminations enable nanoparticle anchoring and rapid signal transduction. Xu et al. functionalized monolayer Ti₃C₂T_x with Ag nanoparticles to construct a room-temperature field-effect-transistor sensor^[56]. The device achieved a detection limit of 35 ppb over 0.05-10 ppm H₂S, response and recovery times of 34 and 58 s, and a response to 1 ppm H₂S more than 4.5 times that of pristine Ti₃C₂T_x^[56]. Ag-mediated chemical and electronic sensitization enhanced the response, while calibration curves obtained at 5%-80% RH enabled humidity-compensated quantification. In a fuel-cell-type sensor, Pt/Ti₃C₂ achieved a sensitivity of 0.162 μA ppm^-1 and a detection limit of 10 ppb. Its response decreased by only approximately 2% after 90 days, compared with 22.9% for Pt/C^[149]. Experiments and DFT calculations attributed this durability to Pt-O-Ti bonding and strong metal-support interactions, while Pt lowered the H₂S dissociation barrier^[149]. These studies establish MXenes as both conductive sensing channels and stable catalyst supports, although oxidation and termination-dependent behavior remain concerns.

Noble-metal functionalization creates active adsorption sites and tunes the electronic structure of TMDs. Verma et al. developed 5 at% Pd-doped MoS₂ films that exhibited a response of 276% toward 100 ppm H₂S, response and recovery times of 45 and 65.8 s, and a reported detection limit of 0.3 ppb at room temperature^[116]. Experiments and DFT calculations indicated that Pd altered the electronic structure of MoS₂ and strengthened H₂S adsorption and charge transfer. An oxygen-vacancy-rich MoS₂/Au@Cu₂O multi-heterostructure achieved a response of 201.6% toward 0.5 ppm H2S after 5 s of exposure. Its enhancement arose from the combined contributions of defect-rich MoS₂, Au sensitization, Cu₂O-H₂S affinity, and multiple heterointerfaces^[150]. DFT calculations further predicted that Au, Ag, and Pt incorporation could strengthen H₂S adsorption and electronic modulation on monolayer WS₂, identifying Ag-WS₂ as a promising but experimentally unverified candidate^[151].

Graphene provides an ultrathin conductive channel highly sensitive to adsorption-induced charge transfer. Ag-decorated chemical vapor deposition (CVD) graphene enabled selective and repeatable room-temperature H₂S sensing with a reported detection limit below 100 ppb^[120]. The enhancement was attributed to Ag-assisted H₂S adsorption and dissociation, followed by charge transfer that modulated the graphene carrier density.

Overall, noble-metal/oxide sensors benefit from mature synthesis, tunable defect chemistry, and well-established Schottky- or depletion-layer modulation, while catalytic oxygen spillover accelerates surface reactions. Their performance may nevertheless be limited by metal aggregation, insufficient surface accessibility, and buried interfaces. Noble-metal-functionalized two-dimensional materials combine Ag, Au, Pt, or Pd sensitization with exposed surfaces and rapid in-plane transport. MXenes enable termination-mediated anchoring and strong metal-support interactions, TMDs provide defect-rich semiconducting channels, and graphene improves metal dispersion and carrier transport. Shared limitations include aggregation, sulfur-induced deactivation or overly strong H₂S binding, humidity-dependent transport, MXene oxidation, and poorly controlled metal loading. Operando characterization combined with site-controlled synthesis is needed to distinguish catalytic activation, metal-sulfur interactions, defect-assisted adsorption, and interfacial charge modulation under realistic conditions.

Device design determines whether a material response can be converted into a stable monitoring signal, requiring coordinated control of electrode geometry, film deposition, gas transport, packaging, power supply, readout, and communication. Huang et al. integrated a ZnO:Ga nanowire/nanosheet sensor with a suspended membrane, interdigitated electrodes, and a microheater on a silicon MEMS platform^[73]. The device detected 0.1-0.8 ppm H₂S and exhibited response and recovery times of 22.4 and 16.8 s toward 0.4 ppm^[73]. A complementary flexible and self-powered system combined an Ag-Zn galvanic cell with a leather-hydrogel electrolyte, open-circuit-voltage readout, Bluetooth, and cloud communication for monitoring periodontitis, meat spoilage, and H₂S leakage [Figure 5A]^[74]. These examples show that practical devices require not only miniaturization and low-power operation but also reproducible film-electrode interfaces, corrosion-resistant packaging, low-noise readout, and reliable wireless transmission

In complex atmospheres, the nominal selectivity of a single sensor may be insufficient for distinguishing H₂S from coexisting gases. Sensor arrays instead combine partially selective elements to generate differentiated response patterns, while an electronic nose integrates gas sampling, signal conditioning, data acquisition, and pattern-recognition algorithms. Shahid et al. developed a room-temperature single-chip electronic nose based on silicon field-effect transistors (Si-FETs) functionalized with VO₂, TiO₂, Au, and Pd^[160]. Combined with a polydimethylsiloxane (PDMS) microfluidic channel and a PCB-based readout system, the array produced distinct patterns for H₂S, NH₃, NO₂, and air, and linear discriminant analysis achieved an overall classification accuracy of 89.66%. Zhou et al. further constructed an eight-channel MEMS electronic nose incorporating temperature/humidity acquisition, pump control, signal conditioning, 4G transmission, and cloud connectivity^[161]. K-nearest-neighbor and random-forest models achieved 97% test-set accuracy for H₂S, NH₃, and their mixtures, while a neural network enabled quantitative prediction of both components^[161]. Long-term reliability was addressed by Gawande et al., who used robust regression to correct the response drift of a six-sensor MEMS electronic nose after one year of operation^[71]. These approaches convert multidimensional signals into gas identity, concentration, and monitoring decisions; however, reliable evaluation requires independent train-test separation by experiment, device, or measurement date, together with external validation under variable humidity and mixed-gas conditions.

Practical validation should reproduce the sample matrix, environmental conditions, sampling procedure, and decision criteria of the intended application. For on-site periodontitis screening, a portable analyzer based on Yb-doped Bi₂S₃ nanoribbons integrated oral-exhalation sampling, real-time H₂S detection, and smartphone-based readout, enabling the differentiation of healthy individuals from patients with periodontitis [Figure 5B]^[65]. For oral-health monitoring, a flexible Mo₂CT_x/MoSe₂ sensor coupled with molecular-sieve and silica-gel dehumidification distinguished breath samples from 15 patients with periodontitis and 15 healthy controls. Measurements from six patients after treatment further indicated its potential for longitudinal monitoring^[72]. For food storage, an In₂O₃/SnO₂ H₂S sensor was combined with a commercial ethanol channel in a refrigerator-monitoring system, enabling real-time freshness assessment across multiple spoilage products^[159]. At the industrial scale, an electronic nose containing CuO-, SnO₂-, and WO₃-based sensors was trained using dynamic olfactometry as a reference and validated with samples from a municipal solid-waste treatment plant and two full-scale biofilters^[162]. Broader deployment requires larger clinical or field datasets, standardized sampling, application-specific alarm thresholds, continuous-operation testing, and long-term calibration and maintenance strategies.

Overall, practical H₂S monitoring requires coordinated development of application-specific devices, cross-responsive arrays, data-processing methods, and field validation. MEMS platforms support miniaturization but require stable packaging, while flexible and self-powered sensors improve portability at the cost of environmental robustness. Electronic noses address limited single-sensor selectivity through multidimensional response patterns but remain vulnerable to device variability, baseline drift, and insufficient training data. Future work should standardize device-level evaluation, validate systems using real samples, and develop data-processing frameworks that maintain accuracy under variable humidity, temperature, and background gases.