Selective hydrogenation via precise hydrogen bond interactions on catalytic scaffolds

The study addresses how precisely engineered microenvironments around heterogeneous catalytic sites can mimic enzyme-like weak interactions, particularly hydrogen bonding, to control hydrogenation activity and selectivity. Traditional heterogeneous catalysts lack structural precision to exploit such interactions. Porous organic polymers (POPs), with tunable functional groups and high surface areas, offer a platform to design hydrophilic (–OH) or hydrophobic (–CH3) environments near metal nanoparticles. The hypothesis is that substrates capable of accepting hydrogen bonds (e.g., carbonyl-containing molecules) will interact more strongly with –OH-functional scaffolds, enhancing adsorption, activating the C=O bond, and improving rates/selectivity, whereas non- or weakly H-bonding substrates (e.g., toluene) will be favored on hydrophobic –CH3 environments.

Recent advances emphasize tuning catalyst microenvironments at the nanoscale, including MOFs with tailorable coordination and surface modifications (e.g., PDMS-modified UiO-66) to alter hydrophobicity and enrich organics, improving activity. Single-atom catalysts can be tuned via local coordination/electronic states. POPs have emerged as versatile scaffolds; encapsulating metal nanocrystals in amine-based POPs changed Pd-catalyzed CO oxidation behavior, and POPs bearing solvent-mimicking functional groups created 'solid solvent' environments that enhanced fructose dehydration without high-boiling solvents. Enzyme catalysis relies on weak interactions, notably hydrogen bonds (e.g., human transketolase involving a key H-bond at residue 366), motivating catalyst designs that precisely control substrate–environment interactions to regulate reaction pathways.

• Catalyst synthesis: Hyper-crosslinked porous polymers (HCPs) were prepared via Friedel–Crafts alkylation using phenol (for HCP-OH) or toluene (for HCP-CH3) monomers, with dimethoxymethane as crosslinker and FeCl3 catalyst in 1,2-dichloroethane; heated at 45 °C for 5 h and 80 °C for 19 h. Triphenylamine (~20% theoretical) was co-monomer to introduce N sites for metal anchoring. Polymers were washed, Soxhlet-extracted in methanol (48 h), and dried. Ir nanoparticles were introduced by wet impregnation from H2IrCl6 in water/ethanol (1:1), followed by drying and H2 reduction at 553 K for 5 h. Variations in phenol:toluene ratio tuned –OH density. Pd and Pt versions were similarly prepared for control studies.
• Characterization: ss13C CP/MAS NMR and FT-IR confirmed scaffold structures and functional groups; wettability tests (contact angle) probed hydrophilicity/hydrophobicity; SEM and FIB-SEM 3D reconstructions assessed morphology/porosity; TEM/HAADF and EDS mapping evaluated metal dispersion/particle size; XRD checked crystallinity; N2 adsorption (BET) characterized porosity; XPS and CO-DRIFTS probed metal electronic states; in situ DRIFTS and ATR-IR monitored substrate adsorption modes and strength; adsorption isotherms for furfural and toluene measured in reaction solvent.
• Catalytic testing: Hydrogenations performed in a 120 mL Parr reactor. Typical furfural run: 1 mmol furfural, 20 mL hexane, 20 mg catalyst, H2 300 psi, target temperature under agitation; GC-MS and GC-FID used for product ID/quantification with internal standards. Conditions used in comparative rate plots included: substrate 1 mmol, catalyst 40 mg, hexane 10 mL, H2 300 psi, 120 °C. Substrates included toluene, furfural, 2-methylfuran, and cinnamaldehyde.
• NMR H-bond probing: Concentration-dependent 1H-NMR of phenol with furfural, 2-methylfuran (2-MF), and toluene to detect H-bond-induced chemical shift changes and estimate ΔG.
• DFT calculations: VASP (PBE+D3, PAW, 500 eV cutoff, smearing 0.05 eV) on graphene ribbon models (armchair and zigzag) with terminal H replaced by –OH or –CH3 to represent functional groups. k-point meshes: armchair (1×3×1), zigzag (3×1×1), vacuum ≥15 Å in-plane and 25 Å out-of-plane. Adsorption energies of furfural computed as Eads = Esurf+furfural − (Esurf + Efurfural). Geometries optimized to force <0.02 eV/Å. Complementary Gaussian 09 B3LYP/6-311+G(d,p) with BSSE correction computed interaction energies between furfural and monomers (phenol, toluene).
• Recyclability: Multi-run tests at ~15% conversion; hot filtration and ICP to assess leaching; TEM post-reaction for sintering assessment.

• Scaffold functionality controls substrate-specific rates: Ir-HCP-CH3 outperforms Ir-HCP-OH for toluene hydrogenation, whereas Ir-HCP-OH strongly enhances furfural hydrogenation; similar trends observed with Pd and Pt, indicating the environment rather than metal identity dominates the effect.
• Adsorption thermodynamics: Furfural adsorption isotherms show HCP-OH achieves ~2× higher saturation capacity (1.95 mmol/g) and ~3× higher affinity constant K than HCP-CH3; toluene saturation is similar on both supports and lower than furfural’s, indicating group-specific interactions with carbonyls.
• Spectroscopic evidence of binding modes: In situ DRIFTS shows stronger furfural adsorption on HCP-OH with carbonyl-region peaks at lower wavenumbers than on HCP-CH3. Deconvolution identifies modes at ~1720 cm−1 (physisorbed), 1698 cm−1 (ring adsorption), and 1670 cm−1 (η1 C=O interaction). HCP-OH exhibits higher fraction of η1 binding and higher desorption temperatures (170–220 °C vs. 130–170 °C for HCP-CH3), pointing to stronger C=O interactions.
• Direct H-bond detection: Concentration-dependent 1H-NMR of phenol shows active H shifts from 4.56 to 5.42 ppm upon furfural addition, giving ΔG ≈ −0.552 kcal/mol; minimal shift with 2-MF (ΔG ≈ −0.213 kcal/mol); no shift with toluene, confirming substrate-dependent H-bonding.
• DFT insights: On CH3-functional graphene, furfural adsorbs flat (ring adsorption); on OH-functional graphene, it tilts toward OH with η1 carbonyl interaction. Furfural adsorption is stronger on OH than CH3 by 0.07 eV (zigzag) and 0.16 eV (armchair). H-bond O···H distances: ~1.76 Å (zigzag), 1.85 Å (armchair); CH3 interaction is weaker/longer (~2.48 Å). H-bond elongates C=O from 1.230 Å to 1.242 Å, indicating bond weakening and partial activation.
• OH density effect: Increasing –OH content (via phenol fraction) increases furfural hydrogenation rate; ssNMR peak ratios (150 ppm/135 ppm) track –OH density and correlate with activity.
• Structure–activity relation to metal sites: For Ir-HCP-OH, TOF scales with particle size as d−2.54, suggesting edge sites in proximity to –OH are most active and a synergistic effect between metal sites and nearby OH groups.
• Selectivity control: In cinnamaldehyde hydrogenation, hydrophilic Ir-HCP-OH favors C=O hydrogenation (higher selectivity to cinnamyl alcohol) relative to hydrophobic Ir-HCP-CH3, consistent with H-bond-guided adsorption conformation.
• Recyclability: Ir-HCP-OH maintains activity into run 2 at ~15% conversion; activity drops by run 3 due to Ir sintering/structural collapse. Hot filtration and ICP indicate heterogeneous catalysis with no detectable leaching.
• Generality beyond POPs: Zeolites with higher Si–OH (Si-Beta) show ~10× higher ketone adsorption than Si-ZSM-5 (Si/Al=500), reinforcing the broad role of surface OH groups in H-bonding to carbonyls.

The results confirm the central hypothesis: precisely engineered microenvironments that can hydrogen-bond to carbonyl groups enhance both adsorption and catalytic activation of C=O functionalities, thereby increasing hydrogenation rates and steering selectivity. Hydrophilic –OH scaffolds promote η1 binding of the carbonyl, lower desorption temperatures, and partially weaken the C=O bond, translating into higher rates for H-acceptor substrates (e.g., furfural) and selective C=O hydrogenation in multifunctional molecules (e.g., cinnamaldehyde). In contrast, hydrophobic –CH3 environments favor non-H-bonding substrates (e.g., toluene). The correlation of activity with –OH density, and the d−2.54 TOF scaling, highlights the nano-scale synergy between proximal –OH groups and specific metal sites (likely edges). Cross-validation across metals (Ir, Pd, Pt) suggests the microenvironment as the dominant factor, while adsorption and in situ spectroscopies, NMR, and DFT provide convergent mechanistic evidence. The observation that zeolites with more surface –OH similarly enhance carbonyl adsorption suggests the principle is general to other porous scaffolds. These findings elevate microenvironment engineering as a powerful lever, akin to enzyme binding pockets, for rate and selectivity control in heterogeneous hydrogenations.

This work introduces enzyme-inspired hyper-crosslinked porous polymer catalysts that precisely position –OH or –CH3 groups around metal nanoparticles to control substrate–site interactions via hydrogen bonding. Hydrophilic –OH scaffolds strengthen adsorption and partially activate C=O groups, boosting rates for carbonyl-containing substrates and enabling selective C=O hydrogenation in multifunctional molecules. Hydrophobic –CH3 supports favor less polar substrates. Activity scales with –OH density and depends on metal site topology proximate to –OH groups. The combined spectroscopic, kinetic, and DFT evidence establishes hydrogen bonding as a tunable design element for heterogeneous catalysis. Future research could (i) tailor other functional groups (e.g., amide, carboxyl, pyridinic N) and spatial distributions to fine-tune interactions with different functionalities (e.g., nitro, nitrile), (ii) integrate stability-enhancing encapsulation/redispersion strategies to mitigate sintering, (iii) extend microenvironment designs to continuous-flow reactors and other transformations (oxidations, condensations), and (iv) develop more realistic amorphous polymer models to capture pore-scale effects computationally.

• Active site indeterminacy: Although metal particle sizes and electronic states (XPS, CO-DRIFTS) were similar across supports, other attributes (crystal facets, spatial location within pores) may influence reactivity and cannot be fully excluded.
• Structural modeling: POPs are amorphous; DFT used graphene ribbon surrogates with terminal –OH/–CH3 to represent local functionalities, which may not capture full pore geometry and heterogeneity.
• Stability: Recyclability is limited by Ir sintering/structural collapse after multiple runs; while no leaching was detected, long-term stability requires further engineering.
• Quantitative selectivity data: While trends are clear (elevated C=O selectivity on –OH), detailed selectivity numbers under varying conditions are not exhaustively reported in the presented text.