CO₂ conversion to formamide using a fluoride catalyst and metallic silicon as a reducing agent

Carbon dioxide emissions have been increasing since the Industrial Revolution, and UNEP projects that a 7.6% annual reduction in CO2 emissions for at least a decade is needed to limit warming to 1.5 °C. Utilizing and converting CO2 as a sustainable C1 building block has gained attention because CO2 is abundant, inexpensive, nontoxic, and renewable. However, due to CO2’s thermodynamic stability and kinetic inertness, activation typically requires highly reactive nucleophiles and/or transition-metal catalysts. Organocatalysis for CO2 conversion has also emerged, with N-heterocyclic carbenes, TBD, thiazolium carbenes, 1,3,2-diazaphosphatrane, carbodi-carbenes, tetrabutylammonium formate, and others showing high catalytic activity comparable to metal-based catalysts. Hydrosilanes are common reductants for CO2, yielding silyl formate exothermically, but they are relatively expensive and used stoichiometrically. Metallic silicon (Si) is an inexpensive alternative reducing agent. Large volumes of crystalline silicon from PV panel waste are expected by 2050, yet their economic value is low despite high purity. Utilizing waste Si from solar panel production as a reductant for CO2 conversion to organic chemicals could enable a circular economy. Previous work demonstrated reduction of CO2 to formic acid and methanol using powdered silicon, but details of fluoride catalysis for reductive functionalization of CO2 and the behavior of fluoride on both surface and interior of Si powder remained unclear. This study reports efficient synthesis of formamides from various amines using CO2 and metallic silicon (from solar panel production) as a reducing agent, enabled by catalytic fluoride. Spectroscopic, kinetic, and isotopic experiments, including in situ measurements, elucidate a mechanism involving Si–H intermediates and oxidation of both surface and interior Si, demonstrating the first highly efficient formamide synthesis from CO2 using metallic silicon as a reductant.

The study situates itself within efforts to utilize CO2 as a C1 synthon via both metal-catalyzed and organocatalytic approaches. Prior organocatalysts effective for CO2 functionalization include N-heterocyclic carbenes, TBD, thiazolium carbenes, 1,3,2-diazaphosphatrane, carbodi-carbenes, and tetrabutylammonium formate, among others, often paired with hydrosilane reductants. Hydrosilane-based reductions of CO2 are exothermic but limited by cost and stoichiometric consumption. Earlier work using silicon-based reducing agents showed conversion of CO2 to formic acid and methanol and explored fluoride catalysis with disilanes and powdered Si. Solvent coordination (e.g., DMSO, NMP) enhances hydrosilylation of CO2 by activating Si centers. The present work extends fluoride-catalyzed systems to achieve direct amide formation from CO2, H2O, and amines using metallic silicon, while probing the unique role of fluoride in oxidizing both surface and bulk Si during catalysis and generating mesoporosity, aspects previously underexplored.

Silicon source and preparation: A Czochralski monocrystalline silicon wafer (from AIST, Japan) intended for solar panel production was crushed in an alumina mortar and sieved to 20 µm particles. General reaction setup: Silicon powder, a fluoride catalyst (typically tetrabutylammonium fluoride, TBAF), water, and CO2 were charged into an autoclave with solvent, and the mixture was stirred in an oil bath at the target temperature and time. After reaction, mesitylene was added as an internal standard; products were analyzed by NMR, GC-MS, and GC-FID for qualitative and quantitative determination. Catalyst and condition screening: Fluoride sources (TBAF, TBAF(t-BuOH)4, TEAF, CsF, KF, NaF) and halide controls (TBACl, TBABr, TBAI) were tested with morpholine in NMP at 90 °C, 24 h, CO2 9 atm, H2O 10 mmol, Si powder 5.0 mmol. Solvent effects were examined with morpholine (1.0 mmol), Si (5.0 mmol), H2O (10 mmol), TBAF (0.05 mmol), CO2 (9 atm), 90 °C, 24 h across DMA, DMF, NMP, DMSO, THP, dioxane, MeCN, n-hexane, CH3Cl, toluene, and H2O. Pressure and temperature optimizations were done in DMSO at 120 °C (and 90 °C), 4–6 atm CO2, 72 h. Isotopic labeling: 13CO2 (ca. 1 atm) was used to trace incorporation of carbon into the formyl group; D2O (10 mmol) replaced H2O to assess the proton source. Product isotopologues were assessed by GC-MS (molecular ion shifts) and NMR; 13C incorporation and deuterium content were quantified. In situ spectroscopy (DRIFTS/ATR-FTIR): Si powder was mounted on TBAF(tBuOH)4. Spectra were collected at room temperature and at 100 °C, then after EtOH vapor addition (proton source), followed by CO2 exposure at 100 °C. Bands diagnostic for F–H/Si–H and formyl groups were monitored. XPS: Measurements were performed on an ULVAC-PHI Quantera SXM (dual Mg/Al X-ray source, hemispherical analyzer) under high vacuum (<1×10−7 Pa). Regions: O 1s, C 1s, F 1s, Si 2p. C 1s at 285 eV was used as internal reference. Fresh Si powder and recovered solids (with/without TBAF and with CO2 or Ar) were analyzed. Some recovered samples were milled before XPS to probe interior composition. XRD: Powder XRD patterns of fresh and spent Si were measured to assess crystallinity and formation of amorphous SiO2. N2 physisorption: Isotherms at 77 K were measured (BELSORP mini). Samples were outgassed at 473 K for 2 h to 1 Pa. BET surface areas (P/P0=0.30–0.70) and BJH pore size distributions were obtained to evaluate surface area increases and mesoporosity after reaction. Kinetic/time-course: Formation of formic acid and formamide was monitored over time to elucidate sequential pathways (formic acid formation preceding amide formation).

- Fluoride catalyst necessity and activity: TBAF and TEAF were active; TBAF delivered 1.91 mmol N-formylmorpholine from 0.05 mmol catalyst (ca. 40× more product than catalyst), indicating catalysis; yield based on CO2 in this setup was ~18%. Inorganic fluorides (CsF, KF, NaF) showed low activity; other halides (Cl, Br, I) were ineffective, underscoring the critical role of fluoride and counter-cation/solubility effects. - Optimized conditions: In DMSO (4 mL) at 120 °C, 6 atm CO2, 72 h, morpholine afforded 94% yield; high yields were maintained at reduced temperature (87% at 90 °C). No product formed without H2O or without CO2 (under Ar), and negligible formation occurred without TBAF. - Solvent effects: Aprotic polar solvents with C=O or S=O bonds enabled high yields: DMA, DMF, NMP each >99%; DMSO 95%. Less polar solvents were inferior: THP 52%, dioxane 22%, MeCN 13%, n-hexane 2%, CH3Cl 2%, toluene 1%, water <1%. This supports a role of solvent coordination to Si enhancing Si–H reactivity toward CO2. - Isotopic labeling: 13CO2 incorporation into the formyl group was 88%, confirming CO2 as carbon source. Using D2O led to a molecular ion shift (M from 115 to 116) and 77% deuterium incorporation at the formyl position, confirming water as the hydrogen source. - Mechanistic spectroscopy: In situ FTIR/DRIFTS showed formation of Si–H species (2000–2100 cm−1 shoulder) and F–H (~1900 cm−1) upon heating Si with TBAF(tBuOH)4, which diminished upon CO2 introduction with concurrent growth of a 1600–1700 cm−1 band assigned to formyl/formate species, indicating an Si–H-mediated hydrosilylation-like step forming formate (Si–O–C(O)H) intermediates. - Sequential pathway: Time-course data indicated initial formation of formic acid, followed by conversion with amines to formamides, implying a sequential mechanism via formate/formic acid. - XPS/XRD evidence of bulk and surface transformation: Fresh Si showed Si(0) at 99 eV. After catalysis with TBAF under CO2, only Si(+4) was observed, even after milling, indicating oxidation of both surface and interior Si. Without TBAF (under CO2), milling revealed Si(0), indicating fluoride accelerates internal oxidation. Under Ar with TBAF, oxidation was not observed, showing CO2 is required. F 1s spectra showed Si–F at 685.4 eV, intensified after milling, indicating Si–F formation within the particle interior. - Structural evolution and porosity: XRD of spent Si showed diminished crystalline Si peaks (≤1/10 intensity) and a broad hump at ~20° (amorphous SiO2), with complete disappearance of metallic peaks at higher TBAF loading. N2 sorption showed surface area increased from 8.2 m2 g−1 (fresh) to 299.6 m2 g−1 (spent with fluoride); without fluoride, surface area was only 1.2 m2 g−1. BJH analysis showed mesopores (~14 nm) in spent samples, indicating fluoride-driven degradation of bulk Si into mesoporous silica-like structures during catalysis. - Substrate scope: Various amines were converted to corresponding formamides in high yields under optimized conditions, demonstrating broad applicability (details referenced in the article’s scope statements). - Proposed mechanism: Fluoride coordinates/attacks Si–Si bonds in the presence of H2O to generate Si–H and Si–F species; Si–H reduces CO2 to formate in coordinating solvents; fluoride is regenerated via reaction of Si–F with hydroxide. Subsequent amine formylation yields formamides.

The study demonstrates that metallic silicon, especially from solar panel production, serves as an efficient, low-cost reductant for CO2 conversion to formamides when paired with catalytic fluoride. The findings elucidate a mechanism in which TBAF generates reactive Si–H species from bulk Si that hydrosilylate CO2 to formate intermediates; subsequent reaction with amines yields formamides. In situ FTIR directly tracks Si–H formation and consumption and formyl formation. XPS and XRD establish that fluoride catalysis promotes oxidation of both surface and interior Si atoms to SiO2, transforming the crystalline Si into mesoporous structures, while F 1s signals confirm Si–F formation throughout the particle. The creation of mesoporosity and the large increase in surface area (to ~300 m2 g−1) rationalize sustained catalytic turnover by exposing interior Si for further reaction. Solvent effects support a role for Lewis-basic, polar aprotic solvents in stabilizing/reactivating Si–H toward CO2, consistent with hydrosilylation paradigms. Isotopic labeling firmly assigns CO2 as the carbon source and water as the hydrogen source in the formyl group. Together, these results address the challenge of cost-effective CO2 functionalization by enabling the use of abundant, waste-derived Si as a reductant and reveal a unique fluoride-mediated pathway that activates both the surface and bulk of Si particles, with implications for extending this approach to other difficult reductive transformations (e.g., biomass valorization, other CO2 reductions).

Silicon powder recovered from solar panel production is an efficient reducing agent for the synthesis of formamides from CO2 and water in the presence of catalytic TBAF. Various amines were formylated in high yields, representing the first catalytic conversion of CO2 to amides using metallic silicon as a reductant. Multi-technique analyses (XPS, FTIR, SEM-EDS, XRD, N2 adsorption/desorption) show that both external and internal Si surfaces are oxidized during reduction of CO2 to formic acid via Si–H intermediates, and the fluoride catalyst uniquely enables access to interior Si, generating mesoporosity and high surface areas. These insights suggest broader applicability of fluoride–Si systems to other challenging reductive reactions and support strategies for circular utilization of abundant silicon resources. Future work could explore expanded substrate scopes, optimization for lower pressures/temperatures, catalyst recovery/reuse, and scale-up with real PV waste streams.

- The reaction requires a fluoride catalyst (e.g., TBAF); inorganic fluorides with smaller, less-soluble cations showed poor activity, and alternative halides were ineffective. - High temperatures (90–120 °C) and elevated CO2 pressures (4–9 atm) were employed for optimal yields. - Solvent choice is critical; high yields were obtained in polar aprotic solvents with coordinating C=O or S=O groups (e.g., DMA, DMF, NMP, DMSO), whereas common solvents like MeCN and nonpolar solvents performed poorly. - The process uses excess silicon powder as a sacrificial reductant and generates oxidized silicon (amorphous SiO2/mesoporous solids), implying material consumption and transformation during operation. - Reported yields and scope are demonstrated primarily with model amines (detailed substrate scope not fully enumerated in the provided text), and long-term catalyst stability/turnover beyond the reported runs and scalability were not discussed in detail.