[a] Conversion of the substrate was determined by 1H NMR analysis of the crude mixture. [b] Determined by 1H NMR analysis of the crude mixture.
Complete Catalytic Deoxygenation of CO2 into Formamidine Derivatives
Article first published online: 26 NOV 2012
Copyright © 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Volume 5, Issue 1, pages 117–120, January 2013
How to Cite
Jacquet, O., Das Neves Gomes, C., Ephritikhine, M. and Cantat, T. (2013), Complete Catalytic Deoxygenation of CO2 into Formamidine Derivatives. ChemCatChem, 5: 117–120. doi: 10.1002/cctc.201200732
- Issue published online: 23 DEC 2012
- Article first published online: 26 NOV 2012
- Manuscript Received: 18 OCT 2012
- FP7 Eurotalents Program
- French patent application. Grant Number: FR1255238
Because fossil resources are a limited feedstock and their extensive use results in the problematic accumulation of CO2 in the atmosphere, the organic-chemical industry will face important challenges over the coming few decades to circumvent the use of raw fossil materials.1 In particular, the fuel, petrochemical, and fine-chemicals industries have to find alternative feedstocks and carbon-free energy sources to embrace sustainability.2 In this regard, CO2 has been proposed as an “energy vector” for renewable energies,3 as a solution for hydrogen storage,4 and as a C1 building block for the synthesis of fine chemicals.5, 6 Yet, as a waste compound, CO2 is thermodynamically and kinetically difficult to transform and research efforts are still needed to promote shifts in technology in the chemical industry. Among the challenges that are associated with CO2 transformation, we must acknowledge that, despite recent progress, the scope of chemical functions that are available from CO2 is still very limited and mostly consists of molecules in which at least one CO bond from CO2 is retained.5, 6 In fact, the only catalytic reaction that results in the complete deoxygenation of CO2 is its reduction into methane by hydrogenation, hydrosilylation, or electrochemical methods.7 Interestingly, Wehmschulte and co-workers recently observed that toluene and diphenylmethane could be obtained as side-products in the silylium-catalyzed hydrosilylation of CO2 into methane in the presence of benzene.7f
To utilize CO2 as a “true” C1 building block and to prepare a wide spectrum of chemicals, catalytic reactions that are able to promote the complete deoxygenation of CO2 with the complete reconstruction of the carbon valence sphere are required (Scheme 1). Herein, we report the first solution to tackling this problem by using the cascade reductive functionalization of CO2 into benzimidazoles, quinazolinones, formamidines, and their derivatives.
We recently reported an organocatalytic formylation reaction of NH bonds by using CO2 and hydrosilanes to yield formamides.6 To substitute the CO bond in the formamide derivative and achieve complete deoxygenation, we reasoned that the amide function could be reacted in a cascade reaction with a nucleophile (Scheme 1), such as an amine. The formylation step was efficiently catalyzed by N-heterocyclic carbenes (NHCs), which were found to be efficient at room temperature for the conversion of a large scope of amines, anilines, imines, and N heterocycles.6b This reaction was mild, robust, and selective; thus, it offered a good starting point for the development of new cascade reactions. Alternatively, nitrogen bases, such 1,5,7-triazabicyclo[4.4.0]dec-5-ene (TBD), were also active catalysts in this reaction, but at higher temperatures (100 °C).6a By using a primary amine as a nucleophile, the condensation step with a formamide is known to be thermally available without needing to resort to catalysts. However, hard drying agents, such as phosphorus oxychloride, trifluoroacetic anhydride, and thionyl chloride, are typically required to promote this reaction.8 To avoid the use of such additives, which increase waste formation, we investigated the reactivity of a diamine, o-phenylene diamine (1 a), in the presence of CO2 and hydrosilanes, so as to favor an intramolecular condensation reaction (Table 1). By using 5.0 mol % of IPr in the presence of CO2 (2 bar) and 1 equivalent of phenylsilane, compound 1 a was converted in high yield into its formyl and N,N′-bisformyl derivatives (compounds 2 a (31 %) and 3 a (38 %) respectively) after 24 h at 25 °C (Table 1, entry 5). To our delight, a significant amount (16 %) of the desired benzimidazole (4 a) was also observed in the reaction mixture. This reaction demonstrated that the complete deoxygenation product (4 a) was available under our reaction conditions. However, the observed selectivity indicates the high rate of the formylation reaction in the presence of PhSiH3. Because compound 3 a is unreactive towards condensation, a less-reactive silane, that is, poly(methylhydrosiloxane) (PMHS),6b, 9 was employed to avoid the formylation of both amine functions. By using 3 equivalents of PMHS under similar reaction conditions (Table 1, entry 6), mono-formyl derivative 2 a was formed as the major compound (42 % yield) and compounds 3 a and 4 a were formed as side-products (in 20 % and 5 % yield, respectively). Therefore, the condensation step appears to be rate determining in this cascade strategy. As a consequence, raising the operating temperature to 70 °C was highly beneficial in terms of selectivity and allowed for the formation of compound 4 a in 90 % yield (Table 1, entry 8). Thus, the synthesis of benzimidazole 4 a from CO2, a diamine (1 a), and PMHS unambiguously demonstrates the feasibility of our strategy (Scheme 1) for promoting the complete deoxygenation of CO2 by catalytic reductive functionalization.
|Entry||Silane (equiv)||Catalyst||T||Conversion[a]||Product distribution[b] [%]|
|[°C]||[%]||2 a||3 a||4 a|
The scope of this new catalytic reaction was explored by using a variety of 1,2-diamines with different substitution patterns and linkers (Scheme 2a). Functionalized benzimidazoles 4 b and 4 c were isolated in high yields (>70 %) by using N-alkyl and N-aryl diamines 1 b and 1 c, respectively. Substitution on the phenyl linker in compound 1 a with methyl, chloro, and fused-arene groups enabled the formation of benzimidazoles 4 d–4 h in good yields (>40 %). The relative reactivities of these substrates directly followed that of their corresponding anilines in the presence of CO2 and hydrosilanes:6 The introduction of electron-withdrawing groups deactivated the aniline towards formylation, thereby leading to lower yields in the formation of compounds 4 f–4 h. This behavior was even more pronounced when pyridines and pyrazines were used as linkers (1 i–1 k). However, high yields were still achievable (up to 72 %) with these substrates by switching to the more-reactive PhSiH3 as the reductant (Scheme 2). Following a similar pathway, the 8-aminonaphtylamine (1 l) could be converted into its corresponding N heterocycle (4 l) in a modest 35 % yield, thereby resulting in the formation of a six-membered ring (Scheme 2b).
To expand the scope of this strategy, it is desirable to promote the deoxygenation of the intermediate formamide with a variety of nucleophiles (NuH2; Scheme 1). This hypothesis was first tested by using less-nucleophilic amide groups in the form of anthranilamide derivatives. For example, the simple anthranilamide 5 a was successfully converted into its corresponding 4-quinazolinone heterocycle (6 a, 65 % yield) when reacted with CO2 and 3 equivalents of PMHS in the presence of 5.0 mol % of IPr (Scheme 3). By using an excess of PMHS (6 equiv), the yield was increased to 85 %. We have previously observed that amides are unreactive substrates in the formylation of NH bonds when using CO2 and hydrosilanes.6 Therefore, from a mechanistic point of view, the amine group likely undergoes a formylation reaction first and the amide group plays the role of a nucleophile in a subsequent condensation step. Notably, substituted amides 5 b–5 d remained active substrates, albeit with lower conversions (Scheme 3). This observation is presumably owing to the increased steric congestion in the substituted amides, which slows down the condensation step in the cascade reaction. Phenyl-substituted anthranilamide 5 d exhibits somewhat higher conversion (56 %) than the benzyl (5 c) or methyl derivatives (5 b; 34 and 27 % yield, respectively). This trend follows the acidity of the amide group and could indicate a base-activation step to promote the condensation reaction on the formamide group.
Benzimidazole and quinazolinone derivatives are useful chemicals as pharmaceuticals, where they are utilized as antifungal, antibacterial, anticancer, and anticonvulsant agents.10 Their successful synthesis by the catalytic reductive functionalization of CO2 encouraged us to investigate the reactivity of less-reactive substrates for the preparation of other N heterocycles. To determine whether the aromaticity of the aforementioned benzimidazole and quinazolinone heterocycles was the driving force for this cascade strategy and a prerequisite for the successful complete deoxygenation of CO2, we attempted the synthesis of non-aromatic 3,4-dihydroquinazoline heterocycles from their corresponding 2-aminobenzylamines. Satisfactorily, 2-aminobenzylamine 7 a could be successfully transformed into compound 8 a in 33 % yield (Scheme 4). Although methyl-substituted aminobenzylamine 7 b exhibited poor conversion (10 %), benzyl and phenyl derivatives 7 c and 7 d were transformed into compounds 8 c and 8 d, respectively, in very good yields (>56 %).
The formylation reaction of NH bonds by using CO2 as a carbon source and hydrosilanes as reductants was introduced as an illustration of a diagonal approach for CO2 transformation.6 This strategy relied on the tandem use of a functionalization reagent and a reductant that could be independently modified to access a wide range of chemicals from CO2. This new catalytic reaction for the conversion of CO2 into formamidine derivatives 4, 6, and 8 offers a new dimension to the use of CO2 as an oxygen-free C1 building block (Figure 1). Indeed, by adding a second functionalizing agent (amines, amides, etc.), a 3D chemical space (Figure 1) could be explored for the conversion of CO2 into deoxygenated fine chemicals. For a given oxidation state of the carbon center, tuning the two functionalizing agents describes the base of a cone. Although we showed that the nature of the nucleophiles could be easily adjusted and connected together by using both saturated and unsaturated linkers, the two horizontal axes in the 3D diagram are not truly independent and N heterocycles were obtained. As such, the completely intermolecular reaction was identified as the next challenge that faced this approach (Table 2). By using IPr as a catalyst, the four-component reaction shown in Table 2 was completely inefficient at producing the expected N,N′-diphenyl formamidine (9), even after prolonged reaction times (48 h) or in the presence of PhSiH3 at 100 °C (Table 2, entries 1–3). Instead, a mixture of the mono- and bis-formyl derivatives of aniline was recovered. Surprisingly, less-active catalyst TBD was quite selective in this reaction and the use of a substoichiometric amount of PhSiH3 (0.6 equiv) and 2 equivalents of aniline facilitated the transformation of CO2 into formamidine 9 in a good (44 %) yield after 48 h at 100 °C. Overall, the transformation of CO2 into compound 9 (or compounds 4, 6, and 8) can be described as a “cut-and-paste” reaction, in which the four CO bonds of CO2 are cleaved and the valence sphere of the C center is completely rebuilt with the formation of four new bonds (one CN double bond, one CN bond, and one CH bond). Importantly, this tandem strategy occurs in a single operation; all of the reactants and the catalyst are introduced at the start of the reaction. These results show that simple organocatalysts can promote and synchronize complex chemical transformations, including the functionalization and reduction of highly stable CO2, in “one pot” under mild conditions. The differences in selectivity between the two classes of organocatalysts (guanidines versus NHCs) are puzzling and suggest that different activation pathways are in play in the reductive functionalization of CO2.
|Entry||Silane (R3SiH)||n [equiv]||t [h]||Catalyst[a]||Yield [%]|
For financial support of this work, we acknowledge the CEA, the CNRS, the ANR (Fellowship to C.G. and a Starting Grant to T.C.), the FP7 Eurotalents Program (PD Fellowship to O.J.), and the CEA Physical Sciences Division (PD Fellowship to O.J., a Basic Research on Low Carbon Energies Grant to T.C.). Parts of this work have been filed as a French patent application (FR1255238; submitted: 05 June 2012 by the CEA).
- 1Sustainability in the Chemical Industry: Grand Challenges and Research Needs—A Workshop Report, National Academies Press, Washington, D.C., 2005.
- 5cCarbon dioxide as chemical feedstock, Wiley-VCH, Weinheim, 2010;,
- 8For representative examples, see:
- 9For a review on PMHS, see:
As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.
Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.