Strained few-layer MoS₂ with atomic copper and selectively exposed in-plane sulfur vacancies for CO₂ hydrogenation to methanol

The study addresses the challenge of selectively hydrogenating CO₂ to methanol, a valuable C₁ chemical and energy carrier central to the "methanol economy." Conventional catalysts (e.g., Cu/ZnO/Al₂O₃, In₂O₃-based oxides, solid solutions, metal alloys, Mo-containing solids) have been widely studied, but achieving high methanol selectivity at moderate temperatures remains difficult. Recently, few-layer 2H-MoS₂ with selectively exposed in-plane sulfur vacancies (S_v) was identified as an effective catalyst for methanol synthesis, whereas edge S_v favor methane formation; other MoS₂ phases (1T, 3R) are ineffective. Thus, the key research question is how to selectively expose and activate in-plane S_v on the inert basal plane of 2H-MoS₂ while minimizing edge sites, and whether strain engineering and atomic Cu promoters can further enhance CO₂-to-methanol performance. The study proposes physically constrained sulfidation of MoO₂ within mesoporous silica to produce few-layer, fullerene-like MoS₂ with curvature-induced strain that enriches in-plane S_v, and investigates the role of atomically dispersed Cu in H₂ activation and S_v formation.

Prior work established that MoS₂ structure (polytype, layer number, site type) critically governs selectivity: few-layer 2H-MoS₂ with in-plane S_v drives methanol formation, while edge S_v promote methane. Conventional syntheses often yield multilayer, thick MoS₂ with inert basal planes and randomly exposed vacancies, limiting methanol selectivity. Strain engineering in 2D transition metal dichalcogenides has been shown to activate basal planes, modulate electronic structure, and promote vacancy formation, typically via transfer from wrinkled/patterned substrates—methods that are complex, costly, and ill-suited to high-pressure heterogeneous catalysis. Metal promotion (e.g., Cu, In, Pd, NiGa) can facilitate H₂ activation/spillover and improve CO₂ hydrogenation to methanol across oxide and alloy systems. However, a precisely controllable synthesis of strained, few-layer, fullerene-like 2H-MoS₂ selectively exposing in-plane S_v and its synergy with atomic Cu had not been demonstrated for CO₂ hydrogenation.

Synthesis: - MoO₂ nanocores: One-pot hydrothermal synthesis using ammonium heptamolybdate (AMT), ethanol, and polyvinylpyrrolidone (PVP) at 180 °C for 16 h; particle size tunable (31–147 nm) by AMT/PVP amounts. - MoO₂@SiO₂ core–shell: Deposition of mesoporous silica shell around MoO₂ in water/methanol using CTAC and 2-methylimidazole, followed by TEOS hydrolysis/condensation; shell thickness tunable by deposition time. - MoS₂@SiO₂: Hydrothermal sulfidation of MoO₂@SiO₂ with thioacetamide (TAA) at 200 °C for 24 h in water, then calcination in Ar at 700 °C for 2 h; yields hollow, few-layer (2–4) fullerene-like MoS₂ confined within the silica shell due to physically constrained topological transformation; parameter studies on temperature, time, solvent, and S source. - Cu/MoS₂@SiO₂: Incipient wetness impregnation of MoS₂@SiO₂ with copper acetate in ethanol, drying (80 °C, 1 h), calcination in Ar (500 °C, 3 h); nominal Cu loading 1.5 wt% (varied 0–5 wt%). - References: MoS₂ nanoparticles (direct MoO₂ → MoS₂ without silica), hydrothermal MoS₂ (MoS₂-HT), commercial MoS₂. Characterization: - Morphology/structure: FESEM, TEM/HRTEM, HAADF-STEM with EDS mapping/line scans; observation of hollow fullerene-like MoS₂ (2–4 layers), curvature, and confinement within mesoporous silica; control samples show thicker multilayer stacks. - XRD: Identification of 2H-MoS₂ (100)/(110) peaks; absence of (002) in MoS₂@SiO₂ indicates few-layer structure; MoS₂-NPs, MoS₂-HT, commercial show strong (002) stacking. - N₂ physisorption: Type IV isotherm with H4 hysteresis; BET ~98.1 m²/g; pore volume 0.15 cm³/g. - XPS: Mo 3d and S 2p confirming MoS₂; lower signal for encapsulated samples consistent with core confinement. - EPR: Signal at g≈2.0 to quantify S_v; higher intensity for MoS₂@SiO₂, further increased with Cu. - Raman: E₂g¹ and A₁g¹ positions indicate Cu is not substitutional in MoS₂ lattice. - XAS (XANES/EXAFS): Cu K-edge indicates atomic Cu coordinated to S (Cu–S ~2.22 Å, CN≈1.5), absence of Cu–Cu; WT-EXAFS corroborates Cu–S; small amounts of Cu clusters present by linear combination fitting. - In situ/operando: High-pressure in situ DRIFTS (CO₂+H₂ at 250 °C, 30 bar) to track intermediates (CO*, CH₃O*, CH₃OH bands; CH₄ band); operando Cu K-edge XANES/EXAFS under H₂ reduction, CO₂, and reaction gas (10 bar, 260 °C) to follow Cu coordination evolution (Cu–O → Cu–S, maintained under reaction). - H₂-TPD: Assess H adsorption/desorption; higher H₂ uptake for Cu/MoS₂@SiO₂. Catalytic testing: - Fixed-bed high-pressure flow reactor; 150 mg catalyst; pretreatment H₂ 20 mL/min at 300 °C for 3 h; reaction at 5 MPa, typical H₂:CO₂=4:1, GHSV 8000–24000 mL g_cat⁻¹ h⁻¹, 180–260 °C; effluent analyzed by GC (TCD/FID). Stability tests up to 150 h; parameter sweeps of GHSV and H₂:CO₂ ratio. Computational (DFT): - VASP, PBE-GGA with D3 dispersion, PAW, 400 eV cutoff; Γ-point (films) and 1×1×2 k-sampling (nanotubes); convergence 1e-5 eV, 0.03 eV/Å. Models: 5×5 and 6×6 MoS₂ films with S_v (including two adjacent S_v for CO₂RR), biaxial strain from −5% to +15%; MoS₂ nanotubes to emulate curvature (interior compressive, exterior tensile). Free energy corrections include ZPVE, vibrational enthalpy/entropy; gas-phase corrections; operating conditions incorporated; additional 0.44 eV correction for CO₂.

- Structure and strain engineering: - Physically constrained sulfidation inside mesoporous silica yields hollow, few-layer (2–4 layer) fullerene-like 2H-MoS₂ with curvature-induced strain and abundant in-plane S_v while minimizing edge exposure. - Strain tunable by MoO₂ core size and silica shell thickness; calculated compressive strains (simple geometric estimate) of ca. −4.1%, −2.9%, and −2.1% for varying radii. - EPR shows higher S_v density for MoS₂@SiO₂ than MoS₂-NPs/HT/commercial; S_v signal further increases upon Cu addition. - Atomic Cu speciation and role: - XANES/EXAFS reveal atomically dispersed Cu coordinated to S (Cu–S ~2.22 Å, CN≈1.5), no Cu–Cu peak; operando EXAFS shows stable Cu–S under reaction; minor Cu clusters present. - H₂-TPD indicates higher H adsorption/desorption for Cu/MoS₂@SiO₂, evidencing enhanced H₂ activation. - DFT: Single-atom Cu at Mo-atop/hollow sites lowers nearby S_v formation energies by ~1.5 eV; Cu migrates between adjacent S_v with low barrier (0.33 eV), facilitating dynamic vacancy environments. - Catalytic performance (CO₂ hydrogenation to methanol): - Products: methanol as main product; CO and CH₄ as by-products; trace DME. - MoS₂@SiO₂ outperforms MoS₂-HT, MoS₂-NPs, commercial MoS₂ across 180–260 °C; at 260 °C, 5 MPa, GHSV 8000 mL g_cat⁻¹ h⁻¹ achieves CO₂ conversion 11.28%, MeOH selectivity 52.16%, specific MeOH yield 1.89 mol_MeOH mol_MoS2⁻¹ h⁻¹; lower CH₄ selectivity than references. - Cu promotion (1.5 wt%): at 260 °C, 5 MPa, GHSV 8000 mL g_cat⁻¹ h⁻¹ reaches CO₂ conversion 12.73%, MeOH selectivity 59.2%, specific MeOH yield 2.42 mol_MeOH mol_MoS2⁻¹ h⁻¹; Cu@SiO₂ alone is weak (2.5% conversion, 50.4% selectivity), highlighting Cu–MoS₂ interface necessity. - Optimization vs Cu loading (GHSV 24000): volcano in MeOH selectivity and specific yield; best at 1.5 wt% Cu with MeOH selectivity 66.6% and specific MeOH yield 5.30 mol_MeOH mol_cat⁻¹ h⁻¹. - Long-term stability: Over 150 h at 260 °C, 5 MPa, GHSV 24000, both MoS₂@SiO₂ and Cu/MoS₂@SiO₂ show increasing performance initially (due to growing S_v under reducing conditions) and then stable operation, with preserved structure (HRTEM, XRD, XPS) and stronger EPR signals post-reaction. - Benchmark: After 150 h on stream, Cu/MoS₂@SiO₂ delivers specific MeOH yield 6.11 mol_MeOH mol_Mo⁻¹ h⁻¹ with 72.5% MeOH selectivity at 260 °C, 5 MPa, GHSV 24000 mL g_cat⁻¹ h⁻¹, surpassing state-of-the-art Mo- and Cu-based catalysts including commercial Cu/ZnO/Al₂O₃ (under similar conditions). - Mechanistic insights: - In situ DRIFTS detects intermediates: linearly adsorbed CO* (~2076 cm⁻¹), CH₃O* (2810–3000 cm⁻¹), CH₃OH (1005–1054 cm⁻¹), Mo=O (900–965 cm⁻¹); CH₄ band (~3014 cm⁻¹) weaker with Cu, indicating suppression of overhydrogenation. - DFT on strain: Compressive strain lowers S_v formation energy (facilitates vacancy creation and CO₂ activation), while tensile strain facilitates H₂ dissociation and accelerates O hydrogenation (site regeneration). Energetic span analysis shows 2% compressive strain lowers methanol-formation span (1.55 → 1.46 eV) and 2% tensile lowers O-hydrogenation span (2.34 → 2.16 eV), explaining synergistic roles of interior (compressive) and exterior (tensile) surfaces of curved MoS₂.

The work demonstrates that selectively exposing and activating in-plane S_v on few-layer 2H-MoS₂ is key to steering CO₂ hydrogenation toward methanol and away from methane. Confining the MoO₂→MoS₂ transformation within a mesoporous SiO₂ shell yields curvature-induced strain, which increases in-plane S_v density and accessibility while minimizing edge sites, thereby enhancing methanol selectivity. Atomic Cu anchored as Cu–S species further promotes performance by facilitating S_v generation, enabling H₂ activation/spillover, and enhancing stepwise hydrogenation of CO* intermediates to methoxy and methanol. In situ DRIFTS confirms similar intermediates on MoS₂@SiO₂ and Cu/MoS₂@SiO₂ but with stronger methanol-pathway features and weaker methane bands upon Cu promotion; operando XAS verifies stable atomically dispersed Cu–S under reaction conditions. DFT rationalizes the dual role of strain: compressive strain lowers S_v formation energies and favors CO₂ activation/methanol formation, while tensile strain facilitates H₂ dissociation and O hydrogenation to regenerate active sites, consistent with the spherical MoS₂ interior/exterior experiencing compressive/tensile strain. Collectively, the findings answer the central question by showing that strain-engineered, fullerene-like few-layer MoS₂ with atomic Cu maximizes in-plane S_v activity and achieves superior, stable methanol synthesis.

The study introduces a scalable, physically constrained synthesis of hollow, few-layer, fullerene-like 2H-MoS₂ within mesoporous silica, achieving curvature-induced strain that selectively exposes in-plane sulfur vacancies ideal for CO₂-to-methanol catalysis. Decorating the strained MoS₂ with atomically dispersed Cu–S species further enhances H₂ activation, vacancy formation, and methanol pathway kinetics. The optimized Cu/MoS₂@SiO₂ achieves a specific methanol yield of 6.11 mol_MeOH mol_Mo⁻¹ h⁻¹ with 72.5% selectivity at 260 °C and 5 MPa after 150 h on stream, outperforming leading benchmarks, and exhibits excellent structural stability. In situ/operando spectroscopies and DFT elucidate the mechanism and the complementary roles of compressive/tensile strain in activity and site regeneration. Future work could explore broader strain regimes via core size/shell thickness control, extend the strategy to other transition metal dichalcogenides and reactions (e.g., electrocatalysis), optimize Cu loading/dispersion and alternative promoters, and investigate scale-up and continuous operation over extended periods.

- The synthesis requires precise control of MoO₂ core size and SiO₂ shell thickness; deviations lead to incomplete sulfidation or MoS₂ growth outside the shell, impacting curvature/strain and performance. - Although XAS indicates predominantly atomic Cu–S species, linear combination fitting suggests small amounts of Cu clusters remain, which may contribute to activity. - The catalyst exhibits an induction period with performance increasing over tens to hundreds of hours, complicating direct early-stage benchmarking versus oxide catalysts. - Catalytic evaluations were performed at elevated pressure (up to 5 MPa); generalizability to lower-pressure conditions, large-scale reactors, and varying feed compositions requires further validation. - Direct imaging of single Cu atoms within the confined structure remains challenging; localization is inferred from spectroscopy and elemental mapping rather than atom-resolved microscopy.