Site‐Specific Molecular Engineering Redirects Radical Pathways for Selective Photocatalytic Coupling of Methane to C₂₊ Hydrocarbons

ABSTRACT

Toward carbon neutrality, multi‐carbon synthesis from photo‐driven oxidative coupling of CH₄ (POCM) remains a formidable challenge due to the activity‐selectivity trade‐off, originating from sluggish surface reaction kinetics and uncontrolled reactive radical reactions. Herein, alkyl thiols (C₉SH) were site‑specifically grafted onto Au sites supported on planar TiO₂, forming the archetypal C₉S‐Au^δ+/Au/TiO₂ system featuring covalent gold‐thiolate interplay. Combined experimental and theoretical analyses revealed that covalent Au‐thiolate interaction could modulate the interfacial electronic structure and upshift the d‑band center of Au sites, thereby strengthening *CH₃ adsorption and lowering the C─C coupling barrier, thus suppressing the ·O₂ ⁻‐driven overoxidation. Moreover, the reconstructed C₉S‐Au^δ+ sites acted as rapid electron extraction channels, drawing electrons from adjacent Au nanoparticles and preserving long‐lived photogenerated holes for C─H activation. Meanwhile, the alkyl chains served as “molecular fences”, effectively promoting local CH₄ enrichment and stabilizing *CH₃ intermediates. The optimized C₉S‐Au^δ+/Au/TiO₂ photocatalyst exhibited an excellent yield of 22.92 mmol g_cat ⁻¹ h⁻¹ for C₂₊ products with 93.9% selectivity, ranking it among the state‐of‐the‐art noble‐metal‐loaded photocatalysts for POCM. This work establishes site‐specific molecular engineering as an effective strategy to regulate interfacial charge redistribution and redirect radical coupling pathways, enabling CH₄ conversion to multi‐carbon products with enhanced activity and selectivity simultaneously.

Site‐specific self‐assembly of alkyl thiols on Au/TiO₂ enables precise modulation of interfacial electron distribution and the local microenvironment through covalent Au‐thiolate interactions. This site‐specific molecular engineering synergistically promotes surface kinetics and charge separation, suppressing ·O₂ ⁻‐driven overoxidation and thereby enhancing both activity and the C₂₊ selectivity of C₉S‐Au^δ+/Au/TiO₂ for photocatalytic oxidative coupling of CH₄.

1. Introduction

CH₄, abundantly present in natural gas, shale gas, and combustible ice, is a key feedstock for fine chemical synthesis, yet its upgrading is thermodynamically challenging due to low polarity and high C─H bond dissociation energy [1, 2, 3, 4]. Industrial routes like dry/steam reforming (>700°C) followed by Fischer–Tropsch synthesis are energy‐intensive and carbon‐emitting [5, 6]. Photocatalysis offers a sustainable alternative for C─H activation under mild conditions, with direct photocatalytic coupling of CH₄ to C₂₊ products (C₂H₆, C₂H₄, C₃H₈, etc.) being highly promising [7, 8, 9, 10, 11, 12]. Nonoxidative coupling of CH₄ (NOCM) can generate stoichiometric hydrogen and ethane with high selectivity, while the reactivity is restricted by the high free energy barrier (ΔG ⁰ _298K = 68.6 kJ·mol⁻¹) [13, 14]. In contrast, thermodynamically favorable oxidative coupling of CH₄ (OCM, ΔG ⁰ _298K = ‐320 kJ·mol⁻¹) can achieve high yield of C₂₊ products with O₂ as oxidant and electron sacrificial agent [15, 16]. Semiconductor oxides such as TiO₂, ZnO, CeO₂ integrated with diverse cocatalysts (Au, Pd, Ag et al.) exhibit great potential in the CH₄ coupling process but often confront with unsatisfied selectivity due to high activation barrier of C(sp³)─H bond and uncontrolled reactivity of reactive oxygen species (ROS, ·O₂ ⁻, ·OH, ·OOH etc.) (Figure 1a) [17, 18, 19, 20]. Therefore, it's been a long‐term endeavor to simultaneously enhance the activity and selectivity towards photocatalytic oxidative coupling of CH₄.

FIGURE 1.

Molecular engineering strategy and reaction mechanism of POCM. Schemes of (a) conventional metal/semiconductor oxide for POCM, (b) non‐site‐selective surface modification strategy, and (c) site‐specific self‐assembled molecular engineering used in this work.

To address the above issues, optimizing the surface reaction kinetics and charge separation efficiency of semiconductor photocatalysts provides effective strategies for steering the photocatalytic OCM process [21, 22]. Heteroatom engineering has been widely demonstrated to adjust the electronic band structure and surface‐active sites of semiconductor oxides, thus enhancing the C─H bond activation and subsequent C─C coupling step in photocatalytic CH₄ coupling [14, 23, 24, 25]. For instance, Ce doping in Au/ZnO regulates the reactivity of lattice oxygen by disrupting electronic asymmetry, thereby promoting *CH₃ desorption and *H elimination [25]. Besides, the incorporation of BiO _x species into TiO₂ enhances lattice oxygen activity and strengthens CH₄ chemisorption [26]. Despite these significant advances, current efforts primarily focus on promoting the surface reactivity by tuning the microenvironment of semiconductor substrates with high noble metal loading. Metal cocatalysts typically function as both charge transfer channels and C─C coupling sites in photocatalytic CH₄ coupling, forming a critical bridge between surface reaction kinetics and charge transfer dynamics [5, 26, 27]. Nevertheless, the influence of precisely modifying metal cocatalysts on surface reactivity and charge carrier dynamics is often overlooked and remains largely unexplored.

Molecular engineering has emerged as an effective approach to modulating catalytic performance by tailoring the geometric configuration, electronic structure, and surface wettability of active sites [28, 29, 30]. However, traditional molecular engineering through physical mixing or wet‐chemical approach often leads to the random distribution of modulators and weak adherence to catalysts, making the active sites blocked and catalytic stability decreased (Figure 1b) [31]. Recent studies have demonstrated that certain chelating molecule, such as amine‐based, thiol‐containing compounds, and N‐heterocyclic carbenes (NHCs), can specifically bind to active sites through covalent bonds, offering a promising strategy for site‐specific surface modification [32, 33, 34, 35, 36]. For example, alkyl thiol‑modified SnO₂ for CO₂ reduction achieved tailored interfacial wettability and local electronic reconstruction, thus stabilizing *OCHO for selective HCOOH production [37]. Similarly, axial NHC modification on Ag₁/TiO₂ enriched electron density at Ag sites, enhancing charge separation and O₂ activation for efficient H₂O₂ generation [38]. These results underscored that the site‐specific modification of organic modifiers can finely alter the electronic structure and interfacial microenvironment of active sites, giving a promising solution to precisely tailoring the metal cocatalyst of photocatalysts for efficient photocatalytic oxidative coupling of CH₄.

Herein, the archetypal C₉S‐Au^δ+/Au/TiO₂ photocatalyst was rationally constructed by site‐specifically grafting alkyl thiols onto Au sites of Au/TiO₂ through molecular engineering. In situ and time‐resolved spectroscopy measurements complemented with DFT calculations elucidated that the covalent grafting of alkyl thiols could modulate the local electronic distribution and upshift the d‐band center of Au sites, thereby enhancing the adsorption of *CH₃ intermediates and significantly lowering the activation barrier of C─C coupling for C₂₊ products, thus circumventing the undesired overoxidation by ·O₂ ⁻ radicals. More importantly, femtosecond transient absorption spectroscopy unveiled that the electronically modulated C₉S‐Au^δ+ sites could serve as rapid electron extraction channels to improve the overall electron transfer efficiency from TiO₂ to Au NPs. Besides, the grafted thiols with long alkyl chains on Au sites acted as “molecular fences” that concentrated CH₄ reactants and improved the surface coverage, thereby facilitating the C─H bond activation. Together, the archetypal C₉S‐Au^δ+/Au/TiO₂ photocatalyst delivered a C₂₊ yield of 22.92 mmol g_cat ⁻¹ h⁻¹ with 93.9% selectivity. In particular, the Au mass‐normalized activity of C₉S‐Au^δ+/Au/TiO₂ (3329.80 mmol g_Au ⁻¹ h⁻¹) was 2.43‐fold higher than that of pristine Au/TiO₂, simultaneously achieving superior selectivity under identical conditions. These results reveal that precisely manipulating the interfacial electronic structure via site‐specific molecular engineering can synergistically promote reactant adsorption/activation kinetics and charge separation dynamics, thus redirecting the radical chemistry and simultaneously enhancing the activity and selectivity for POCM (Figure 1c).

2. Results and Discussion

2.1. Preparation and Structural Characterizations of Photocatalysts

As shown in Figure 2a, thiol modified Au/TiO₂ was prepared through the covalent interaction between the deprotonated sulfhydryl groups of thiols and Au NPs. The bond energy of generated Au‐thiolate bond was comparable to that of Au─Au bond, endowing this interaction with great potential to modify the microenvironment of the Au─Au bond around Au–sulfur interface [39, 40, 41]. PXRD patterns in Figure S3 revealed that the fabricated Au/TiO₂ photocatalysts exhibited the characteristic anatase structure (JCPDS No.21‐1272) combined with negligible diffraction peaks of cubic Au (JCPDS No.65‐8601). HRTEM demonstrated that the TiO₂ nanosheets predominantly exposed the (001) and (101) facets and Au NPs were mainly anchored on the edges of the anatase nanosheets (Figures 2b). The loading of Au species in Au/TiO₂ and C₉S‐Au^δ+/Au/TiO₂ measured by ICP‐OES was 0.728% and 0.723%, respectively. The mean size of Au NPs remained almost unchanged after thiol modification, with a slight decrease from 3.36 nm (Au/TiO₂) to 3.08 nm (C₉S‐Au^δ+/Au/TiO₂) (Figures S4 and S5). Besides, EDS element mapping in Figure 2c revealed that the Au NPs overlapped with S element on C₉S‐Au^δ+/Au/TiO₂, validating the site‐specific modification of thiol on Au NPs. Water contact angle test in Figure 2d revealed that alkyl thiol modification could enhance the hydrophobicity of Au/TiO₂ and the contact angle increased from the original 15.3° (Au/TiO₂) to 42.1° (C₉S‐Au^δ+/Au/TiO₂). Furthermore, the STEM and EELS in Figure 2e demonstrated that the signals of S L_2,3 edges emerged around Au NPs, while being almost negligible on the TiO₂ support, further confirming that thiol molecules could precisely modify Au NPs without influencing the physicochemical properties of TiO₂. As shown in Figure 2f, different from TiO₂ and Au/TiO₂, the C₉S‐Au^δ+/Au/TiO₂ exhibited characteristic peaks corresponding to C─H scissoring vibrations of ‐CH₂‐S (∼1415 cm⁻¹) and ‐CH₂‐ (∼1446 cm⁻¹), as well as symmetric/antisymmetric C─H stretching vibrations of ‐CH₂‐ (∼2854 and ∼2957 cm⁻¹, respectively) [42, 43, 44]. Raman spectroscopy in Figure S6 suggested that all of the photocatalysts exhibited typical anatase phase of TiO₂, consistent with the PXRD results. In addition, the emerging peaks for C₉S‐Au^δ+/Au/TiO₂ at 268, 326–350 cm⁻¹ were attributed to the radial and tangential vibrations of Au─S bond, which further verified the successful grafting of thiol ligands on Au NPs via the covalent Au‐thiolate interaction [45, 46]. Moreover, XPS was employed to characterize the variations of the electronic structure and chemical microenvironment on Au/TiO₂ photocatalyst after thiol modification. As shown in Figure 2g, the XPS peaks for Au/TiO₂ at 83.32 and 87.03 eV were assigned to the 4f _7/2 and 4f _5/2 orbitals of Au⁰, respectively [47]. In contrast, after thiol grafting onto Au sites, new peaks at 84.25 and 88.25 eV appeared, which were ascribed to Au^δ+ species and accounted for 25.9% of the total surface Au content [48]. The newly generated Au^δ+ species substantiated the strong interaction between 1‐nonanethiol and Au NPs.

FIGURE 2.

Experimental method and identification of site‐specific thiol modification. (a) Synthesis of Au/TiO₂ and C₉S‐Au^δ+/Au/TiO₂. (b) HRTEM and (c) EDS element mapping of C₉S‐Au^δ+/Au/TiO₂. (d) Contact angle of the photocatalysts (d‐i, d‐ii and d‐iii were TiO₂, Au/TiO₂, and C₉S‐Au^δ+/Au/TiO₂, respectively). (e) STEM image and EELS of S L_2,3 edge for C₉S‐Au^δ+/Au/TiO₂ in the marked region. (f) FTIR, (g) high‐resolution of Au 4f XPS and (j) CH₄ adsorption isotherms at 273.15 K of the prepared photocatalysts. (h) Differential charge density (yellow: electron accumulation, cyan: electron depletion) of C₆S‐Au₅/TiO₂ model. (i) PDOS of Au₅/TiO₂, C₆S‐Au₅/TiO₂ and C₆H₁₃SH.

To unravel the nature of Au‐thiol interfacial interaction, detailed electron transfer and orbital interplay were analyzed based on DFT calculations. The Au/TiO₂ and C₉S‐Au^δ+/Au/TiO₂ were modeled by loading Au₅ cluster and the C₆H₁₃SH modified Au₅ cluster on TiO₂(101) slab (denoted as Au₅/TiO₂ and C₆S‐Au₅/TiO₂, respectively) (Figure S7). Bader charge analysis in Figure 2h revealed that 0.1 electron transferred from Au cluster to S sites, suggesting that thiol modification could induce the electronic reconstruction of Au species, well aligning with the corresponding XPS results. Besides, the projected density of states (PDOS) analysis in Figure 2i revealed that the S‐p orbitals of thiols were largely overlapped with the s and d orbitals of Au. The p‐d and s‐p orbital hybridization emerged after the alkyl thiols grafted on Au sites, further confirming the covalent nature of the Au─S interaction and indicating electron redistribution across the organic‐metal interface. Moreover, to deduce the influence of thiol modification on CH₄ adsorption capacity of the obtained C₉S‐Au^δ+/Au/TiO₂ photocatalysts, CH₄ adsorption isotherms were collected at 273.15 K. As shown in Figure 2j, the CH₄ adsorption capacity of C₉S‐Au^δ+/Au/TiO₂ was 17.65 cm³ g⁻¹ at 273.15 K and 0.92 bar, which was much higher than that of TiO₂ (7.85 cm³ g⁻¹) and Au/TiO₂ (4.92 cm³ g⁻¹), suggesting that thiols grafted on Au cocatalysts could serve as “molecular fences,” effectively confining CH₄ reactants and increasing the surface coverage. As mentioned above, these results clearly demonstrated that the alkyl thiols could precisely modify Au NPs deposited on TiO₂ via the strong covalent Au─S interaction. The site‑specific modification of thiols with long alkyl chains could induce local electron distribution of Au sites and enhance the surface coverage of CH₄, thereby holding great potential for boosting the photocatalytic coupling of CH₄.

2.2. Performance Evaluation of POCM

To investigate the effect of thiol grafting on the photocatalytic performance of Au/TiO₂, photocatalytic coupling of CH₄ with oxygen (POCM) was conducted in a continuous‐flow reactor under ambient pressure (Figure S1a). Figure 3a suggested that the POCM performance was significantly enhanced after grafting thiol onto Au/TiO₂ photocatalyst. The yield and selectivity of C₂₊ products (including C₂H₆ and a small amount of C₃H₈ and C₄H₁₀) both exhibited volcano‐type dependence on the carbon chain length of thiol molecules grafted on Au/TiO₂, with the nonanethiol‐modified photocatalyst (C₉S‐Au^δ+/Au/TiO₂) exhibiting the highest activity and selectivity. Specifically, the yield of C₂₊ over C₉S‐Au^δ+/Au/TiO₂ could achieve 120.37 µmol h⁻¹ with the selectivity up to 91.7%. Figure 3b indicated that the Au mass‐normalized activity of C₂₊ products for C₉S‐Au^δ+/Au/TiO₂ could reach 3329.80 mmol g_Au ⁻¹ h⁻¹, 2.43‐fold higher than that of unmodified Au/TiO₂ (1370.85 mmol g_Au ⁻¹ h⁻¹) with enhanced selectivity, highlighting the critical role of precisely modified electronic structure and local microenvironment of Au cocatalysts. Subsequently, the modification amount of thiol was further investigated. As shown in Figure 3c, the yield of C₂₊ products increased from 76.15 to 149.17 µmol h⁻¹ when the adding amount of thiol ranged from 50 to 200 µL. However, the reactivity of POCM exhibited severe decrease when further increasing the modification amount of thiol above 200 µL, indicating the poisoning and the blocking of active sites by the excess thiol modification. The mass of C₉S‐Au^δ+/Au/TiO₂ was optimized to achieve higher yield of C₂₊ products. As shown in Figure 3d, the C₂₊ yield was observed to improve with higher photocatalyst mass, reaching 229.20 µmol h⁻¹ (3170.12 mmol g_Au ⁻¹ h⁻¹) with an excellent selectivity of 93.9% at the mass loading of 10 mg. A further increase in photocatalyst mass led to negligible change in either C₂₊ yield or selectivity, likely because excessive stacking in the photocatalyst bed limited light penetration.

FIGURE 3.

Photocatalytic coupling of CH₄ with oxygen. (a) POCM performance of Au/TiO₂ and Au/TiO₂ modified by thiols with different alkyl length. (b) The yield of Au/TiO₂ and C₉S‐Au^δ+/Au/TiO₂ normalized by the mass of Au. The influence of (c) different adding amount of nonanethiol and (d) photocatalyst mass loading of C₉S‐Au^δ+/Au/TiO₂. (e) Online mass spectrometry under dark and light irradiation. (f) Comparation of the POCM performance with reported results (more details in Table S2). (g) Evaluation of long‐term stability of C₉S‐Au^δ+/Au/TiO₂. Reaction conditions: 72 mL min⁻¹ CH₄ + 1 mL min⁻¹ air, 303 K, 365 nm LED, 86 mW cm⁻², 10 mg photocatalyst.

The evolution of main products during POCM was investigated by online mass spectrometry under reaction conditions. As shown in Figure 3e, no product was detected under dark conditions. Upon illumination, O₂ (m/z = 32) was gradually consumed, while C₂H₆ (m/z = 30), H₂O (m/z = 18), and CO₂ (m/z = 44) were progressively generated. In addition, light‐pulse confirmed the photo‐driven nature of POCM, as evidenced by the strict dependence of product evolution on illumination. Besides, the apparent quantum yield (AQY) for POCM over C₉S‐Au^δ+/Au/TiO₂ exhibited a wavelength‐dependent with its UV–vis diffuse reflectance spectrum, achieving a peak of 4.03% at 365 nm (Figure S8 and Table S1). The AQY dropped sharply to 1.08% at 400 nm, with no detectable conversion at 550 nm, indicating that the plasmonic resonance effect of Au NPs was insufficient to promote CH₄ coupling. As confirmed above, the site‐specific thiol modification on Au/TiO₂ significantly improved the activity and selectivity for C₂₊ products in POCM. The high yield of C₂₊ for C₉S‐Au^δ+/Au/TiO₂ could surpass most of the state‐of‐the‐art noble‐metal‐loaded photocatalysts reported in recent years (Figure 3f, Table S2).

Beyond catalytic selectivity and activity, stability is an essential metric for determining its industrial viability. The long‐term stability evaluation was conducted under optimal conditions (Figure 3g). The yield of C₂₊ products maintained at 203.69 µmol h⁻¹ with a selectivity nearly 94.4% after 24 h of continuous testing. HRTEM in Figure S9a suggested that the mean size of Au NPs was merely increased after the long‐term stability test, rising from the original 3.08 nm to 3.45 nm. PXRD patterns in Figure S9b evidenced that the crystal structure of the post‐reacted C₉S‐Au^δ+/Au/TiO₂ was unchanged, indicating the high structural stability of prepared TiO₂‐based photocatalysts. More importantly, Figure S9c revealed that the content of surface oxygen vacancies of the spent C₉S‐Au^δ+/Au/TiO₂ remained nearly unchanged, implying that O₂ could effectively replenish the oxygen vacancies formed during the POCM process via lattice oxygen looping, thereby ensuring high structural stability of the C₉S‐Au^δ+/Au/TiO₂ [25, 26, 27]. Besides, as evidenced in Figure S9d, partial oxidation occurred at Au sites after the long‐term stability test, which was likely attributed to the strong oxidative conditions caused by photogenerated ROS under UV radiation.

2.3. Investigation of Photophysical Properties

The photophysical properties were investigated to elucidate the structure‐activity relationship of the as‐synthesized photocatalysts. UV–visible diffuse reflectance spectra (UV–vis DRS) in Figure S10a revealed that strong localized surface plasmon resonance (LSPR) at ∼550 nm was observed on both Au/TiO₂ and C₉S‐Au^δ+/Au/TiO₂. It's noted that the band structure of C₉S‐Au^δ+/Au/TiO₂ was nearly unchanged relative to Au/TiO₂ and TiO₂, but the LSPR peak was obviously broadened. The broader LSPR peak of C₉S‐Au^δ+/Au/TiO₂ suggested that thiol‑induced electronic reconstruction could create additional electron transfer channels, accelerating plasmon dephasing and electron transfer dynamics [49, 50]. The derived Tauc plots based on UV–vis DRS combined with the XPS valence spectra in Figures S10b–e determined the band structures of the as‐synthesized photocatalysts. As shown in Figure S10f, the valence band potential of the photocatalysts satisfied the thermodynamic requirement for the oxidation of CH₄ to ·CH₃, while its conduction band potential enabled the reduction of O₂ to ·O₂ ⁻, reflecting that the prepared photocatalysts exhibited promising activity for CH₄ oxidation [51, 52]. In parallel, the photoluminescence (PL) spectra were evaluated to investigate the charge separation efficiency and charge carrier lifetime of the prepared photocatalysts. As illustrated in Figure S11, the PL intensity of pristine TiO₂ decreased significantly after Au NPs doping, indicating that Au NPs effectively suppressed photogenerated carrier recombination. Moreover, the C₉S‐Au^δ+/Au/TiO₂ photocatalyst exhibited even lower PL intensity than Au/TiO₂, reflecting that thiol modification further enhanced charge separation efficiency.

2.4. Charge Transfer and O₂ Activation Mechanism

To unveil the charge transfer mechanism, in situ XPS was evaluated under UV illumination. The Au 4f and Ti 2p orbitals shifted toward lower binding energies (Figures 4a,b), declaring that photogenerated electrons could accumulate at the surface Ti and Au sites. Moreover, in situ electron paramagnetic resonance (EPR) was evaluated at 77 K under Ar atmosphere to monitor the evolution of paramagnetic species over the synthesized photocatalysts under dark and light illumination. As shown in Figure 4c, two minor EPR signals with g = 2.001 and 1.987 could be assigned to O_v and Ti³⁺ under dark conditions, respectively [53, 54]. The signals of Ti³⁺ were obviously improved after continuous illumination, underlining the trapping of photogenerated electrons on Ti⁴⁺ sites, well associated with the corresponding in situ XPS analysis. In parallel, the weak signals of O⁻ species (denoted as O⁻(L)) with the g values at 2.011, 2.003 and the overlapping region with O_v were emerging, which originated from photogenerated holes trapped by lattice oxygen [55, 56]. As mentioned above, the Ti³⁺‐O⁻(L) pairs could be formed after light excitation. In particular, C₉S‐Au^δ+/Au/TiO₂ exhibited more enhanced intensity of these photogenerated paramagnetic species, further confirming that the separation efficiency of charge carriers could be improved by the thiol grafted on Au NPs. Subsequently, the signals of Ti³⁺ and O⁻(L) were obviously decreased when CH₄ was introduced (Figure 4d), suggesting that the photogenerated Ti³⁺‐O⁻(L) pairs played significant roles in C─H bond activation of CH₄.

FIGURE 4.

Charge transfer and lattice oxygen looping mechanism in POCM. In situ XPS of (a) Au 4f and (b) Ti 2p for C₉S‐Au^δ+/Au/TiO₂. (c) In situ EPR of Au/TiO₂ and C₉S‐Au^δ+/Au/TiO₂ under Ar at 77 K. (d) In situ EPR of C₉S‐Au^δ+/Au/TiO₂ under different conditions at 77 K. (e) In situ EPR of C₉S‐Au^δ+/Au/TiO₂ under dark or different light irradiation time at 298 K. Intensity evolution of (f) O_v (g = 2.001) and (g) ·O₂ ⁻ (g = 2.024). (h) Online mass spectrometry of C₉S‐Au^δ+/Au/TiO₂ under light irradiation.

To illustrate the underlying reaction mechanism of POCM, in situ EPR experiments were examined at 298 K under different conditions. As shown in Figure 4e, the EPR signal of O_v at g = 2.001 progressively increased with continuous light illumination in CH₄ and oxygen‐deficient atmosphere. With prolonged light illumination, paramagnetic signals at g = 2.009, 2.024 and those overlapping with the O_v region gradually appeared, which were assigned to ·O₂ ⁻ formed by O₂ photoreduction [20, 57]. The signals of ·O₂ ⁻ disappeared while the signal intensity of O_v was higher than its initial level after light‐off for 10 min, indicating the involvement of lattice oxygen in POCM process under O₂‐deficient atmosphere. Upon subsequent introduction of excess O₂ under light, the signals of O_v and ·O₂ ⁻ were all emerged rapidly, suggesting that O₂ could act as electron sacrificial agents to accelerate the separation of photocarriers. Besides, the integrated intensity of O_v and ·O₂ ⁻ in Figures 4f,g demonstrated that these paramagnetic species were highly dependent on illumination and their intensity decayed rapidly or even became undetectable once the illumination was turned off. In particular, as shown in Figure 4f, the integrated intensity of O_v under the oxygen‐enriched conditions after light‐off for 10 min was obviously weaker than its initial level. The replenishment of O_v further suggested that oxygen could participate in the looping of lattice oxygen in POCM to maintain the structural stability of Au/TiO₂ photocatalysts. Moreover, online mass spectrometry was implemented to trace the oxygen cycle using isotope labeling ¹⁸O₂ and CH₄ as reactants (Figure 4h). The gradual emergence of the H₂O¹⁶ signal over reaction time directly confirmed the participation of lattice oxygen in POCM, in good agreement with the corresponding in situ EPR analysis. Besides, the optimal O₂ concentration in POCM represents a kinetic balance between efficient photogenerated charge separation and controlled generation of reactive oxygen species. The rapidly emerging signals for C¹⁸O₂ upon continuous illumination revealed that excessive oxygen could result in severe overoxidation of the value‐added C₂₊ products, which was mainly due to the generation of excessive ·O₂ ⁻ species as convinced by in situ EPR results under oxygen‐enriched conditions. On this basis, in situ EPR with DMPO in CH₃OH further revealed weaker DMPO‐OOH signals for C₉S‐Au^δ+/Au/TiO₂ than Au/TiO₂ under light illumination (Figure S12), confirming that thiol modification could effectively mitigate O₂ photoreduction, thus avoiding the excessive formation of ·O₂ ⁻ radicals.

2.5. Exploration of Ultrafast Charge Separation Dynamics

Based on the above analysis, femtosecond transient absorption spectroscopy (fs‐TAS) was further conducted to unveil the influence of thiol modification on charge transfer dynamics of Au/TiO₂. The TA spectra of photogenerated holes and electrons overlap for TiO₂ in the UV–visible region, complicating charge transfer kinetics analysis, whereas the infrared region can reveal the ultrafast dynamics of photoelectrons more clearly [58, 59, 60]. Therefore, the TA spectra in the near‐infrared region (900‐1300 nm) were first used to probe photogenerated electron dynamics under 350 nm excitation (Figures 5a–f). All of the TiO₂‐based photocatalysts exhibited the wide absorption peaks of photoelectrons, in good agreement with related works [60, 61]. The normalized decay kinetics of photoelectrons at 1000 nm was fitted within 3000 ps to track the photoelectron transfer dynamics (Figure 5g). The corresponding fitting results were listed in Table S3 and the average photoelectron lifetime of TiO₂ was decreased from the original 494.56 ps to 146.94 ps after the photodeposition of Au NPs, which was further decreased to 67.03 ps for C₉S‐Au^δ+/Au/TiO₂. Notably, the relaxation dynamic of Au/TiO₂ was 3.10 times faster than that of pristine TiO₂, while C₉S‐Au^δ+/Au/TiO₂ exhibited a 2.19‑fold acceleration relative to Au/TiO₂, demonstrating that thiol modification substantially facilitated charge transfer from TiO₂ to Au, consistent with in situ XPS analysis and relevant reports [17, 25].

FIGURE 5.

Ultrafast charge transfer dynamics of the photocatalysts. Pseudocolor plots and fs‐TAS recorded at different delay times for TiO₂ (a and d), Au/TiO₂ (b and e) and C₉S‐Au^δ+/Au/TiO₂ (c and f) under 350 nm excitation in the range of near‐infrared region (900–1300 nm) at Ar atmosphere. (g) Decay dynamics of photoelectron probing at 1000 nm for the photocatalysts. Schemes for the photoelectron decay pathways of (h) Au/TiO₂ and (i) C₉S‐Au^δ+/Au/TiO₂.

Furthermore, the fs‐TAS in the ultraviolet‐visible region was employed to decipher charge transfer dynamics associated with Au species after thiol modification. As shown in Figure S13, both Au/TiO₂ and C₉S‐Au^δ+/Au/TiO₂ exhibited an obvious bleach feature centered at approximately 550 nm, which could be assigned to the localized surface plasmon resonance (LSPR) of Au nanoparticles and was consistent with the corresponding UV–vis spectra. The strong spectral overlap between the plasmon bleach signals of Au NPs and relatively weak photogenerated carrier absorption in TiO₂ hinders the precise assessment of the carrier dynamics for Au NPs. Nevertheless, compared with TiO₂ and Au/TiO₂, C₉S‐Au^δ+/Au/TiO₂ displayed a markedly broadened bleach profile and faster spectral evolution around 550 nm during the decay process. This behavior suggested that thiol grafting electronically modulated the Au nanoparticles and introduced additional relaxation pathways at the Au‐thiolate interface. Specifically, the formation of interfacial C₉S‐Au^δ+ species could accelerate electron extraction from adjacent Au nanoparticles, thereby shortening the plasmon coherence lifetime. Such accelerated plasmon relaxation was consistent with the broadened LSPR peak of C₉S‐Au^δ+/Au/TiO₂ observed in the UV–vis spectra and agreed well with previous reports [48, 50, 62]. Additionally, DFT‐calculated work functions based on electrostatic potentials were 4.82 eV for TiO₂ and 5.10 eV for Au (Figure S14), suggesting a higher Fermi level for TiO₂ that could drive electron transfer to Au until equilibrium. Based on the aforementioned analysis, the photoelectron kinetics of Au/TiO₂ could be deconvoluted into three processes: the short‐lived component τ₁ corresponded to electron trapping by shallow trap states; the second electron decay pathway (τ₂) was assigned to the electron transfer between TiO₂ and Au; the long‐lived component τ₃ represented electron‐hole recombination (Figures 5h,i). It's noted that the τ₂ process of C₉S‐Au^δ+/Au/TiO₂ (1.19 ps) was much faster than that of Au/TiO₂ (17.88 ps), revealing that the newly generated C₉S‐Au^δ+ channels could effectively extract photoelectrons from the neighboring Au NPs, thus accelerating the overall interfacial charge transfer between TiO₂ and Au NPs. The enhanced interfacial electron transfer reduced the availability of photoelectrons for hole recombination, thereby preserving long‐lived holes for the activation of C(sp³)─H bonds in CH₄.

2.6. Tracking Reaction Pathway of POCM

To unlock the reaction pathway of POCM, in situ attenuated total reflectance‐Fourier transform infrared (ATR‐FTIR) spectroscopy was conducted under reaction conditions. As shown in Figures 6a and S15, the peaks at 3017 cm⁻¹ and 1305 cm⁻¹ for both C₉S‐Au^δ+/Au/TiO₂ and Au/TiO₂ emerged after the introduction of CH₄ and O₂ in the dark, which were attributed to the antisymmetric stretching and deformation vibrations of C─H bond in gaseous CH₄, respectively [5]. In particular, the slight peak located at ∼1540 cm⁻¹ was ascribed to the symmetric deformation vibrational mode of C─H bond in chemisorbed *CH₄, which was regarded as a key species for C─H bond activation [15, 60]. Besides, the characteristic peak at 2828 cm⁻¹ was attributed to the redshift of symmetric stretching mode (v ₁) of C─H bond in CH₄, further suggesting the stretching and activation of C─H bond [63]. As shown in Figures6b,c, those new peaks gradually appeared at 1340, 1488, and 1645 cm⁻¹ after light illumination, which were assigned to the bending vibration of C─H bond in *CH₃, the alkyl deformation vibration in *C₂H₆ and the symmetrical stretching vibration of O—H bonds in H₂O, respectively [25, 27, 64]. These emerged peaks confirmed that the C─H bond of CH₄ was successfully activated and subsequent C─C coupling occurred after photoexcitation. Furthermore, the asymmetric stretching vibration of CO₂ at 2315–2375 cm⁻¹ progressively appeared under continuous illumination [65]. In detail, the integrated intensity of CO₂ on C₉S‐Au^δ+/Au/TiO₂ was markedly weaker than that of Au/TiO₂ under identical conditions (Figure 6d), further substantiating that thiol grafted on Au sites could efficiently alleviate ·O₂ ⁻‐driven overoxidation pathway in the POCM process, in good agreement with the relevant in situ EPR, online mass spectra and experimental results.

FIGURE 6.

Investigation of reaction mechanism of POCM. In situ ATR‐FTIR of (a and b) C₉S‑Au ^{δ +}/Au/TiO₂ and (c) Au/TiO₂. (d) CO₂ intensity evolution and (e) ratio of *CH₃/*CH₄ integrated from the corresponding in situ ATR FTIR results of the prepared photocatalysts at different irradiation times. (f) Relative reaction energy for the C–H activation of CH₄ on the C₆S‐Au₅/TiO₂ models. (g) Online mass spectra of C₉S‑Au ^{δ +}/Au/TiO₂ under gradient increasing of light intensity. (h) PDOS and d band center of C₆S‐Au₅/TiO₂ and Au₅/TiO₂. (i) The coupling energy barrier of *CH₃ over Au₅/TiO₂ and C₆S‐Au₅/TiO₂. (j and k) POCM mechanism on the C₉S‐Au^δ+/Au/TiO₂ photocatalyst.

More importantly, the local concentrations of *CH₄ and *CH₃ were evaluated by integrating the corresponding peak areas obtained in Figures 6b,c. Both the integrated intensity of *CH₄ and *CH₃ on C₉S‐Au^δ+/Au/TiO₂ were higher than those of Au/TiO₂ (Figure S16), in line with the corresponding CH₄ adsorption isotherms, suggesting that the confined microenvironment was created by the alkyl chains of thiols grafted onto Au sites. Additionally, Figure 6e revealed a higher ratio of Area_*CH3/Area_*CH4 for C₉S‐Au^δ+/Au/TiO₂ than that for Au/TiO₂, demonstrating that the confinement effect of alkyl thiol modification sustained an enhanced local CH₄ concentration, thereby promoting *CH₃ formation for subsequent C─C coupling. Furthermore, online mass spectra in Figure 6g revealed that the POCM products exhibited gradual increase with progressively rising light intensity. Notably, C₉S‐Au^δ+/Au/TiO₂ exhibited higher C₂H₆ but lower CO₂ signals than Au/TiO₂ across all light intensities, further confirming that thiol modification effectively facilitated CH₄ activation while suppressing overoxidation of the C₂₊ products. With increasing light intensity, C₉S‐Au^δ+/Au/TiO₂ exhibited stronger H₂O signals than Au/TiO₂, indicating enhanced hydrophobicity after thiol modification, consistent with the water contact angle results. The improved hydrophobicity favored the rightward shift of the POCM process, thereby preventing active‐site blocking by accumulated *H₂O and ultimately promoting C₂₊ products formation.

To further elucidate the influence of thiols modification on the POCM mechanism of Au/TiO₂‐based photocatalysts, DFT calculations were carried out. The dissociation of the first C─H bond in CH₄ was calculated on C₆S‐Au₅/TiO₂. As shown in Figure 6f, CH₄ underwent weak chemical adsorption on the TiO₂ and then dissociated into the *CH₃(Ti)‐*H(O_L) species with a ΔE of 1.45 eV. Subsequently, *CH₃ might either undergo further dehydrogenation to *CH₂ on Ti sites or migrate to adjacent Au sites. The much lower ΔE for *CH₃ migration from Ti to Au sites (ΔE = ‐1.02 eV) than for *CH₃ dehydrogenation to *CH₂ (ΔE = 2.50 eV) indicated that CH₃ migration was thermodynamically favored. Figure S17 revealed that Au₅/TiO₂ exhibited a similar dissociation pathway of C─H bond to that of C₆S‐Au₅/TiO₂, implying that thiol modification exerted negligible impact on the electronic structure of TiO₂. Moreover, Figure 6h suggested that the related d band center of Au was upshifted from ‐3.48 to ‐3.11 eV after thiol grafting, signifying the enhanced adsorption toward *CH₃ intermediates according to the well‐founded band center theory [66]. After the adsorption of *CH₃ on Au sites, energy barrier of subsequent C─C coupling process was calculated (Figures 6i and S18). The coupling energy barrier of *CH₃ on C₆S‐Au₅ sites was 0.65 eV, which was much lower than that on Au₅ cluster (0.85 eV), suggesting that the site‐specific thiol modification on Au sites could accelerate C─C coupling kinetically. Subsequently, the coupling of *CH₃ intermediates to form *C₂H₆ was thermodynamically favorable (ΔE = ‐0.36 eV for C₆S‐Au₅/TiO₂). As mentioned above, the site‐specific thiol modification could efficiently modulate electronic redistribution of Au sites in C₉S‐Au^δ+/Au/TiO₂, which contributed to optimizing the adsorption of *CH₃ intermediates and reducing the kinetic barrier for C─C bond coupling, thereby promoting the selective production of C₂₊ products and circumventing undesirable overoxidation.

Based on the aforementioned experimental and DFT calculation results, the POCM reaction mechanism on C₉S‐Au^δ+/Au/TiO₂ photocatalyst was proposed. As shown in Figure 6j, CH₄ was initially adsorbed on the TiO₂ substrate (step 1). The confined microenvironment created by alkyl chains of thiol molecules effectively increased the surface coverage of CH₄, thereby forming CH₄ reservoir on TiO₂ surface and facilitating the C─H bond activation. The Ti³⁺‐O_L ⁻ pairs could be formed after electrons and holes were photoexcited in TiO₂ (step 2). Subsequently, the C─H bond dissociation occurred on these Ti³⁺‐O_L ⁻ sites by generating *CH₃(Ti)‐*H(O_L) intermediates (step 3). After that, the *CH₃ species migrated to adjacent Au sites (step 4) followed by the adsorption and activation of another CH₄ on TiO₂ (step 5). The resulting *CH₃ species on the Au sites underwent C─C coupling to generate target C₂H₆ product (step 6). Concurrently, the *H species were observed to bind with the activated lattice oxygen of TiO₂ to generate *H₂O, thereby leaving O_v on the photocatalyst surface. The hydrophobic alkyl chains of thiols could effectively promote the desorption of *H₂O, thus shifting the equilibrium of POCM to the right and favoring the formation of C₂₊ products. Finally, the remaining O_v was replenished by the photoreduction of O₂, thereby completing the lattice oxygen cycle (step 7). In particular, compared with unmodified Au/TiO₂, the site‐specific thiol modification on Au cocatalysts triggered the local electronic redistribution, which concurrently steered the charge transfer dynamics and lowered the reaction barrier for C─C coupling, thus significantly promoting generation of C₂₊ products while suppressing undesired overoxidation pathway (Figure 6k).

3. Conclusion

In summary, the archetypal C₉S‐Au^δ+/Au/TiO₂ photocatalyst was constructed through site‑specific grafting of alkyl thiols onto Au sites supported on planar TiO₂, achieving a high C₂₊ production rate of 22.92 mmol g_cat ⁻¹ h⁻¹ with an excellent selectivity of 93.9%. Moreover, C₉S‐Au^δ+/Au/TiO₂ exhibited almost a 2.43‐fold increase in the Au‐mass normalized yield with improved selectivity simultaneously compared with pristine Au/TiO₂, highlighting the enhanced noble metal utilization via the site‐specific molecular engineering. Comprehensive characterization and mechanistic analysis demonstrated that the covalent Au‐thiolate interaction could precisely manipulate the electronic distribution of Au NPs, upshifting the d band center of Au sites, thereby promoting *CH₃ adsorption and substantially lowering the energy barrier of C─C coupling. Moreover, the reconstructed C₉S‐Au^δ+ species could create a unique electron transfer channel that accelerated the extraction of electrons from the adjacent Au NPs, significantly boosting the charge carrier separation efficiency. Besides, the alkyl thiol molecules grafted onto Au sites acted as “molecular fences” that effectively concentrated CH₄ reactants for C─H bond activation and stabilized the *CH₃ intermediates. This work highlights the pivotal role of site‐specific molecular engineering in simultaneously regulating radical reaction pathways and surface reaction kinetics for photocatalytic oxidative CH₄ coupling and offers a distinct design paradigm for surface molecular engineering toward upgrading light hydrocarbons to multi‐carbon products.

Author Contributions

Yaru Shao: methodology, investigation, writing – original draft, formal analysis and visualization. Wanying Guo: software, methodology, visualization. Wei Li: investigation, validation, formal analysis. Gaoyu Cui: investigation and formal analysis. Huijie Cheng: supervision, writing – review and editing. Lingfeng Zhu: supervision and formal analysis. Liqun Ye: review and formal analysis. Zihan Qiu: methodology, validation and formal analysis. Yu Shan: supervision, methodology and visualization. Rui Li: supervision and formal analysis. Tianyi Ma: supervision, conceptualization and writing – review and editing. Jungang Hou: supervision, validation, conceptualization, funding acquisition, writing – review and editing.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: anie74526‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.