α, β-Unsaturated amides are important structural motifs ever present in wide range of natural products and materials science frameworks [1-8]. For instance, α, β-unsaturated amides derivatives display biological activities applied in the treatment of psychological diseases and showed remarkable potential for the treatment of cancer (Fig. 1) [9-11]. Besides, α, β-unsaturated amides are versatile building blocks in organic synthesis [12-17]. Synthetic methods for α, β-unsaturated amides preparation has attracted broad interest due to important applications in medicine and organic synthesis. The development of greener, low-cost, energy-efficient and selective synthetic methods for α, β-unsaturated amides preparation has attracted broad interests.

Traditional synthetic methods for unsaturated amides are generally, based on nucleophilic substitution of amines with carboxylic acids derivatives and acyl chlorides or cyclization with amic acids in the presence of activating reagents (Scheme 1a) [18-23]. Transition metal catalyzed hydroaminocarbonylation provides a high atom- and step-economy pathway for producing high value-added unsaturated amides [24-37]. Reports on catalytic hydroaminocarbonylation of alkynes have been documented in literature by reacting primary and secondary amines for the preparation of the corresponding α, β-unsaturated amides with high chemo- and region-selectivity (Scheme 1b) [38-46]. However, noble metals, costly ligands or additives were necessary to achieve reasonable yields and good regioselectivity [27, 38, 39, 41-43]. The development of more efficient carbonylation methods for the synthesis of unsaturated primary amides remains desirable.

Solid ammonium salts are cheaper and easy-to-handle ammonia source widely applied in aminocarbonylation reactions to produce various amide compounds [47-56]. Recently, we have developed the efficient iron-catalyzed aminocarbonylation of alkynes to produce succinimides with NH₄HCO₃ [54]. Huang and co-workers developed a palladium-catalyzed aminocarbonylation method for transforming alkenes with NH₄Cl into the corresponding amides [55]. Besides, Huang et al. reported a selective palladium-catalyzed hydroaminocarbonylation reaction between alkynes and NH₄Cl to branched α, β-unsaturated primary amides [56]. Liu's group reported a hydroaminocarbonylation reaction with alkynes and NH₄HCO₃ to generate branched α, β-unsaturated primary amides (Scheme 1c) [37]. Nevertheless, developing non-noble metal-catalyzed hydroamin-ocarbonylation with NH₄HCO₃ for the preparation linear α, β-unsaturated primary amides has not been achieved up to date. Herein, we report the first example of ligand-free iron-catalyzed hydroaminocarbonylation of alkynes to generate linear α, β-unsaturated primary amides using NH₄HCO₃ as ammonia source (Scheme 1d).

Initially, phenylacetylene 1a was selected as the model substrate for this hydroaminocarbonylation reaction. A series of commercially available catalysts were examined with NH₄HCO₃ used as ammonia source in the presence of CO. When iron salts such as FeCl₃ and FeCl₂ were used, no desired α, β-unsaturated amide product 2a could be observed. To our delight, when Fe₃(CO)₁₂ was used as catalyst the hydroaminocarbonylation reaction proceeded successfully and gave the desired cinnamamide 2a in 46% yield. The use of Fe₂(CO)₉ instead of Fe₃(CO)₁₂ as catalyst afforded the desired product in only 13% yield for 2a (Table 1, entries 1–4). In addition, other metal carbonyls were found inefficient for this transformation, with no desired product could be detected (Table S1 in Supporting information). Moreover, carrying out the reaction without Fe₃(CO)₁₂ resulted in undetected product, while adding PPh₃ as ligand was ineffective for hydroaminocarbonylation reaction (Table 1, entries 5 and 6). We observed that reaction temperature and CO pressure played important role in this reaction. Cinnamamide 2a was prepared in 81% yield when the reaction was performed at 140 ℃ under 30 bar CO (Table 1, entries 7–11). The yield was slightly decreased to 62% at 160 ℃ (Table 1, entry 12). Increasing the amount of NH₄HCO₃ to 5.0 mmol improved the yield significantly, and the desired product 2a was obtained in 91% yield (Table 1, entries 13 and 14). Further screening of other reaction parameters such as solvent and the catalyst loading did not improve the reaction yield (Table 1, entry 15 and Table S2 in Supporting information). We also examined the effect of the ammonium salts and found (NH₄)₂CO₃ and HCOONH₄ proved to be effective as ammonia source. However, NH₄Cl or NH₄OAC were found inefficient for the hydroaminoca-rbonylation reaction (Table S3 in Supporting information).

With the optimized reaction conditions in hand, we explored the scope of alkyne substrates for the synthesis of linear α, β-unsaturated primary amides (Scheme 2). Gratifyingly, a range of alkynes are suitable substrates to react with NH₄HCO₃ and CO under the optimized reaction conditions. The desired linear α, β-unsaturated primary amides 2a-2q were obtained in 46%−91% yields. The electronic properties of the substituents on the aromatic ring of the aromatic alkynes have weaker influence on the reactivity and selectivity. The results of meta- and para-substituted aromatic alkynes showed insignificant electronic effects. And surprisingly, the reaction of sterically hindered 1-ethynyl-2-methoxybenzene (1b) provided the desired linear α, β-unsaturated primary amides in good yield (2b, 88%). Moreover, aromatic alkynes bearing electron-withdrawing substituents such as fluoro- and chloro- groups have less influence on the reactivity (2c, 84%; 2d, 83%; 2g, 81%; 2k, 83%). Similarly, aromatic alkynes with strong electron withdrawing groups like trifluoromethyl and esters substitution afforded linear α, β-unsaturated primary amides in moderate yields (2l, 79%; 2m, 81%). To our delight, aliphatic alkynes were also transformed in moderate to good yields, e.g., 2n, 2o and 2p. Trace amounts of succinimide was detected under the optimized reaction conditions. The method could be applied for the internal alkyne 1, 2-diphenylethyne (1q) to give moderate yield (46%).

To demonstrate the synthetic utility of this method, vanilloid receptor-1 antagonist TRPV-1 was prepared. 3-(4-(tert-Butyl)phenyl)acrylamide 2j preparation proceeded smoothly at gram-scale and 1.68 g of 2j was obtained under the slightly modified reaction conditions in 83% yield (Scheme 3a). The subsequent coupling reaction of 3-(4-(tert-butyl)phenyl) acrylamide afforded vanilloid receptor-1 antagonist TRPV-1 4jr on a gram scale (Scheme 3b) [57].

In order to better understand the mechanism of iron-catalyzed hydroaminocarbonylation of alkynes using NH₄HCO₃ as the ammonia source, a series of experiments were performed (Scheme 3 and Fig. S5 in Supporting information). Firstly, hydroaminocarbonylation reaction in the absence of Fe₃(CO)₁₂ of phenylacetylene 1a with NH₄HCO₃ using CO was unable to provide the desired cinnamamide 2a (Scheme 4a). Furthermore, carrying out the reaction without CO resulted in undetected cinnamamide 2a (Scheme 4b). Cinnamamide 2a was detected when the reaction was performed with Fe₃(CO)₁₂ serving as catalyst and CO source (Schemes 4c and d). Interestingly, cinnamamide 2a and succinimides 3a were detected when the reaction was performed with gaseous NH₃ (6 bar) serving as ammonia source instead of NH₄HCO₃ under the optimized reaction conditions (Scheme 4e). Various solid ammonium salts used as ammonia source provided different results and showed clear influence on the reactivity (Table S3 in Supporting information). It was indicated that NH₄HCO₃ promoted the formation [Fe]-H species responsible for the efficient hydroaminocarbonylation of alkynes, enabling the highly efficient and regioselective synthesis of linear α, β-unsaturated primary amides [26-28]. We established that Fe(CO)₅ is formed from Fe₃(CO)₁₂ reaction with NH₄HCO₃ based on in situ ¹³C NMR experiments using NH₄HCO₃ and Fe₃(CO)₁₂ [54].

Based on the experimental results and recent experimental data on the hydroaminocarbonylation of alkynes and NH₄HCO₄ [24-28, 54-56, 58-62], we propose a possible mechanism pathway as shown in Fig. 2. Initially, the active mononuclear iron carbonyl Fe(CO)₅ was formed in situ through interactions of Fe₃(CO)₁₂ with NH₄HCO₃. Meanwhile, NH₃ and H₂CO₃ is released through in situ decomposition of NH₄HCO₃. Then, Fe(CO)₅, NH₄HCO₃ and CO generate the intermediate A. Intermediate A reacts with alkyne to produce intermediate B or B'. The steric hindrance of the terminal alkyne plays the major role for the formation of the kinetically favored terminal alkenyl-iron intermediate B. Subsequent CO insertion forms intermediate C and C', which then affords the final carbonylation product D and D' with the presence of NH₃ released from NH₄HCO₃. Besides, the coordination of alkyne substrates and NH₃ with Fe center makes CO insertion and product D formation much more accessible, which is similarly observed in Beller's work using organic amines as the substrates [26].

In summary, we have demonstrated the first example of ligand-free iron-catalyzed hydroaminocarbonylation of alkynes synthesis of linear α, β-unsaturated primary amides using NH₄HCO₃ as the ammonia source. In the presence of NH₄HCO₃ and non-noble Fe₃(CO)₁₂ serving as catalyst, a variety of alkynes, including aromatic alkynes, aliphatic alkynes, terminal alkynes, internal alkynes, were transformed into the desired linear α, β-unsaturated primary amides in good to excellent yields. The applicability of this methodology has been demonstrated by synthesis of α, β-unsaturated amides bio-active compound. Preliminary mechanistic studies reveal the activation model involving interactions of NH₄HCO₃ with Fe₃(CO)₁₂. Further investigations are currently underway to apply the method to other reactions.

Declaration of competing interest

The authors declare no conflict of interest.

Acknowledgments

We are grateful for the financial supports from the National Natural Science Foundation of China (Nos. 21772035, 22022204, 22072167, 21202206) and Natural Science Foundation of Hunan Province (Nos. 2021JJ40147). The authors would like to dedicate this work to Professor Matthias Beller (LIKAT) on the occasion of his 60^th birthday.

Supplementary materials

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2022.01.080.