Copper- and copper–N-heterocyclic carbene-catalyzed C─H activating carboxylation of terminal alkynes with CO₂ at ambient conditions

Abstract

The use of carbon dioxide as a renewable and environmentally friendly source of carbon in organic synthesis is a highly attractive approach, but its real world applications remain a great challenge. The major obstacles for commercialization of most current protocols are their low catalytic performances, harsh reaction conditions, and limited substrate scope. It is important to develop new reactions and new protocols for CO₂ transformations at mild conditions and in cost-efficient ways. Herein, a copper-catalyzed and copper–N-heterocyclic carbene-cocatalyzed transformation of CO₂ to carboxylic acids via C─H bond activation of terminal alkynes with or without base additives is reported. Various propiolic acids were synthesized in good to excellent yields under ambient conditions without consumption of any organometallic or organic reagent additives. This system has a wide scope of substrates and functional group tolerances and provides a powerful tool for the synthesis of highly functionalized propiolic acids. This catalytic system is a simple and economically viable protocol with great potential in practical applications.

The chemical fixation and transformation of carbon dioxide (CO₂) has attracted much attention in view of environmental, legal, and social issues in the past few decades (1–7). Carbon dioxide is an attractive C1 building block in organic synthesis because it is an abundant, renewable carbon source and an environmentally friendly chemical reagent (8–13). The utilization, as opposed to the storage of CO₂, is indeed more attractive especially if the conversion process to useful bulk products is an economical one. Significant efforts have been devoted toward exploring technologies for CO₂ transformation, whereby harsh and severe reaction conditions are one of the major limitations for their practical applications (1–15). Therefore, the development of efficient catalyst systems for CO₂ utilization under mild conditions is highly desired, especially for real world applications.

Carboxylic acids are one of the most important types of compounds in medicinal chemistry and also in fine-chemicals synthesis (16, 17). Although there are many well-established protocols for the preparation of carboxylic acids, the direct carboxylation of carbon nucleophiles using CO₂ as the electrophile is the most attractive and straightforward method (16, 17). The formation of a stable C─C bond is desired for CO₂ fixation and remains the most challenging aspect thus far. Typically, this type of reaction is facilitated by the insertion of CO₂ into a metal-carbon bond (16–26). Widespread use of these methods is limited by the synthesis organometallic reagents as precursors and the restricted substrate scope.

In the past decades, several interesting systems have been reported for metal-mediated reductive carboxylation of alkynes (27, 28), allenes (29, 30), and alkyenes (31) with CO₂ to form carboxylic acids or esters. However, most of those systems need either a stoichiometric amount of transition metals as reactants or an excess amount of organometallic reagents for transmetallation processes. An alternative possibility to achieve the catalytic synthesis of carboxylic acid from CO₂ is by direct C─H bond activation and carboxylation. Very recently, Nolan’s group reported a gold-catalyzed CO₂ carboxylation of C─H bonds of highly activated arenes and heterocycles (32). Herein, we reported a copper- and copper–N-heterocyclic carbene (NHC)-catalyzed transformation of CO₂ to carboxylic acid through C─H bond activation and carboxylation of terminal alkynes. Various propiolic acids were synthesized in good to excellent yields under ambient conditions. This catalytic system is a simple and economically viable protocol with great potential in practical applications.

Results and Discussion

The ubiquity of alkynyl carboxylic acids in a vast array of medicinally important compounds as well as its tremendous utility as a synthon in organic synthesis makes them particularly attractive targets for pharmaceutical and fine-chemical as well as conductive polymer synthesis (16, 17, 33, 34). A plethora of well-established methods for the preparation of alkynyl carboxylic acids includes CO₂ insertion into the metal-carbon bond of organometallic reagents, the well-known hydrolysis of bromide and related derivatives, and the oxidation of preoxidized substrates, such as alcohols or aldehydes (Fig. 1 A–C) (35). Despite the efficiency of these conventional procedures, their major drawback includes severe reaction conditions and restrictions of organometallic reagents that dramatically limit the synthesis of a wide scope of functionalized propiolic acids. Therefore, a functional group tolerant and straightforward method for accessing alkynyl carboxylic acids (Fig. 1D) is highly desired and will provide opportunities in organic, pharmaceutical, and materials synthesis.

Fig. 1.

Protocols for synthesis of substituted propiolic acids.

Organocopper reagents are very unique because the metal-carbon bond is of moderate polarity and is ready for CO₂ insertion under ambient conditions, and they are also tolerant to most functional groups (18). Copper catalysts can also catalyze various C─H and C-halogen activation reactions, and many of them involve intermediates with a Cu─C bond (36–38). These facts make copper catalysts a very promising choice for CO₂ transformation, especially with the formation of new C─C bonds. Hay (39) developed oxidative homocoupling of terminal alkynes via C─H activation and copper acetylide intermediate. After that, carbon dioxide was successfully introduced into the copper–terminal alkyne system, and the insertion of CO₂ into the copper acetylide intermediate was observed by Inoue’s group (40). However, the copper propynoate intermediate is unstable under the applied reaction conditions (100 °C, 1 atm CO₂). Propiolic acid products could not be isolated (40). We assumed that the copper propynoate intermediate should be stable under milder conditions, such as room temperature, and the carboxylate end product could be dissociated from the copper catalyst under normal basic conditions. On the basis of this hypothesis, the initial proof of concept experiment was conducted by using 2 mol% of CuCl, 1.5 mol% of TMEDA (N,N,N^′,N^′-tetramethylethylenediamine) ligand (L1), and K₂CO₃ as the base for the carboxylation of 1-ethynylbenzene 1a at ambient temperature and atmospheric pressure (Table 1, entry 1). Remarkably, phenylpropiolic acid 1b was produced in excellent yield after acid workup. The isolated pure product was characterized by NMR and elemental analysis. Phenylpropiolic acid was further converted into the methyl ester and characterized by NMR and GC/MS. Interestingly, the catalytic reaction could be performed without any base additives to give 55% of 1b (Tables S1 and S2). Further studies indicated that a catalytic amount of base is important to activate the CuCl catalyst. Meanwhile, a weak basic environment is necessary to stabilize the acid product and to ensure that the reaction is completed. In a catalyst system with basic ligands (Table S2, entries 1–3) or a catalytic amount of base (Table S2, entry 13), a moderate yield of 1b was obtained when the reaction was conducted in dimethylformamide (DMF) (a weak Lewis base) but not in THF, CH₃CN, or DMSO (Table S2). Stoichiometric amounts of base additives further promote the reaction, and a good yield of 1b could be obtained in DMF as well as in THF, CH₃CN, and DMSO (Tables S1 and S3). This result demonstrated a valuable example of a base free C─C coupling reaction system.

Table 1.

Copper-catalyzed carboxylation of terminal alkynes with CO₂

				Isolated yields, %
Entry	Alkynes	Time, h	Base	L1	L13
1		16	K2CO3	90	95
2		18	K2CO3	81	85
3		18	K2CO3	86	90
4		18	K2CO3	89	92
5		16	K2CO3	80	82
6		16	K2CO3	85	86
7		16	K2CO3	88	86
8		16	K2CO3	84	90
9		16	K2CO3	86	88
10		16	K2CO3	86	90
11		24	K2CO3	89	93
12		24	Cs2CO3	80	83
13		24	Cs2CO3	82	88
14		24	Cs2CO3	82	90
15		24	Cs2CO3	85	85
16		24	Cs2CO3	83	85

Reaction conditions: for L1, CuCl (2.0 mol%), TMEDA, 1.5 mol%; for L13, P(NHC)_0.5(NHC─Cu)_0.5, 5 mol%; alkynes (2.0 mmol), base (2.4 mmol), CO₂ (1 atm), DMF (4 mL), room temperature (25 °C).

When the reaction was conducted with 2.0 mol% CuCl catalyst in the absence of ligands, the yield of 1b dropped to 50% (Table S4, entry 8). This result indicates that σ donor ligand can increase the catalyst activity (41). Other σ donor ligands, such as N,N^′-dimethylethanediamine, 1,3-dimesitylimidazol-2-ylidene, and 1,8-diazabicyclo[5.4.0]undec-7-ene, also work well for this catalytic system (Table S4). For the carboxylation of 1-ethynylbenzene 1a, when the catalyst loadings was reduced from 2 mol% to 0.5 mol%, good to excellent yields were still obtained after prolonged reaction time (24 h) (Table S4). The yield of 1b decreased sharply to 8% when the catalyst loading was further reduced to 0.1 mol%. No reaction was observed for the control experiments with K₂CO₃ or Cs₂CO₃ as the base and without copper catalyst (Table S4, entries 1 and 14) (42).

Kinetics studies of this reaction showed that 1b was formed rapidly in the first 4 h, reaching a yield of 70%, before the gradual increased to the final yield of 90% in 16 h. The yield remained constant around 90% even with further increasing the reaction time from 16 to 24 h (Fig. S1).

With the optimized reaction conditions of 2.0 mol% CuCl, 1.5 mol% TMEDA, and 120 mol% K₂CO₃ in DMF for 16 h, the substrate scope of the reaction was studied. For aromatic alkynes with or without electron donation functional groups, the corresponding alkynyl carboxylic acids were obtained in 80–91% yields (isolated) under standard conditions (Table 1, entries 1–11). The catalytic system is not sensitive to the position of substituents on the benzene ring. The related acid yields of p-, m-, o-substituted 1-ethylbenzene are approximately in the same range. The transformations proceeded smoothly without any side product formation. The initial trials for the carboxylation of the alkyl-substituted alkynes were unsatisfactory with a low yield of the corresponding acids (∼20%). The low reactivity of alkyl-substituted alkynes is probably because of the weak acidity of the alkyne proton. The conjugation system between the benzene ring and alkyne C≡C bond in aryl alkynes imposes more negative charge on C1 carbon and makes it a stronger nucleophile than alkyl alkynes. A density functional theory calculation [B3LYP/6-31G(d,p) level] indicated that the negative charge on C1 carbon of 1-ethynylbenzene is -0.534 and that of 1-hexyne is -0.461. After investigating further, a stronger base Cs₂CO₃ was used instead of K₂CO₃, and the yield of corresponding alkyl-substituted propiolic acids was raised to 80–91% (Table 1, entries 12–16).

In general, terminal aromatic alkynes with an electron withdrawing group are deactivated and often inert to many transformations (43, 44). With the electron withdrawing group on the phenyl ring, the nucleophilicity of the C1 carbon of alkynes dropped dramatically. The carboxylation of 4-nitro-1-ethynylbenzene 19a was unsatisfactory with a very low yield of the corresponding acids 19b (0 ∼ 8%) under standard conditions even with a very strong base, such as KO^tBu (Table S5). The low yields (∼2%) were also observed as the reaction temperature was adjusted to 0 °C and 50 °C. This result may be because of the low reaction rate at low temperature and instability of the reaction intermediate at high temperature (Table S5). The key step for this transformation is CO₂ insertion into the copper acetylide intermediate. Increasing the nucleophilicity of the carbanionic intermediate may increase the yield of the carboxylic acid product. Various ligands (L1–L12, Table 2) (45) were screened in the reaction with 4-nitro-1-ethynylbenzene 19a and the yields of the acid product ranged from 3% to 47%. The catalyst with the strongest electron donation ligand phenanthroline L10 gave the highest yield. It is well known that NHCs can activate CO₂ in various catalytic transformations (5, 7). With that, a unique NHC-Cu cocatalyst was designed by using poly-N-heterocyclic carbene (PNHC) as both ligand and catalyst. PNHC has a three-dimensional network structure, and the carbene units are located and fixed in the backbone of the network (Fig. 2) (46, 47). P(NHC)_0.5(NHC─Cu)_0.5 (P1) catalyst was prepared by the reaction of one equivalent of CuCl with two equivalents of PNHC. In the structure of this catalyst, only half of the carbene species coordinated with copper, and the other half remained as free carbenes (46). The initial experiment was conducted by using 5 mol% of P1 and Cs₂CO₃ as a base for the carboxylation of 4-nitro-1-ethynylbenzene 19a with CO₂ at ambient conditions. Remarkably, 4-nitro-phenylpropiolic acid 19b was produced in 70% yield after acid workup. Good yields were also achieved for terminal aromatic alkynes with different electron withdrawing groups in 36–48 h (Table 2). The longer reaction time may be because of the heterogeneous reaction behavior in this solid catalyst system. The mechanism of this unique process is intriguing, and related control experiments were conducted. We used 4-nitro-1-ethynylbenzene as a model substrate. No reaction was observed in a system with PNHC only. When an additional two portions of CuCl were added into the P1 catalyst reaction system, the yield of the desired product dropped dramatically to 18% (Table 2, entry 15). This result indicated that free carbene species in the catalyst play an important role for high activity. Furthermore, a reaction intermediate P(NHC─CO₂)_0.5(NHC─Cu)_0.5 was synthesized by the reaction of P(NHC)_0.5(NHC─Cu)_0.5 (P1) with CO₂. This intermediate was directly used to react with a stoichiometric amount of 1-ethynylbenzene (equivalent to NHC─CO₂) under standard conditions without an additional CO₂ source. A 52% yield of phenylpropiolic acid was obtained in 24 h. With these experiment results, it is believed that the unique structure of P1 catalyst is the key to the high activities. The free carbene species in the structure are randomly located around the copper center and act as an organocatalyst to activate CO₂. This essential step may reduce the activation energy barrier for CO₂ insertion.

Table 2.

Cu-NHC–catalyzed carboxylation of deactivated terminal alkynes with CO₂

Entry	Alkyne	Ligand (mol%)	CuCl (mol%)	Time h	Yields %
1		L1 (10)	5	24	< 5
2		L2 (10)	5	60	2
3		L3 (10)	5	60	3
4		L4 (10)	5	60	8
5		L5 (10)	5	60	5
6		L6 (10)	5	60	9
7		L7 (10)	5	60	9
8		L8 (10)	5	48	21
9		L9 (10)	5	48	40
10		L10 (10)	5	48	47
11		L11 (10)	5	48	30
12		L12 (20)	5	48	32
13		L12 (10)	5	48	25
14		L13 (10)*	5	48	70
15		L13 (10)*	15	48	18
16		L13 (10)*	5	36	68
17		L13 (10)*	5	24	72
18		L13 (10)*	5	36	73
19		L13 (10)*	5	36	79

Reaction conditions: alkynes (2.0 mmol), CuCl, Cs₂CO₃ (2.4 mmol), ligand, CO₂ (1 atm), DMF (4 mL), RT.

^*10 mol% of NHC.

Fig. 2.

Structures of poly─NHC and P(NHC)_0.5(NHC─Cu)_0.5 catalyst.

Finally, with this unique P1 catalyst, different terminal alkynes with various functional groups were all successfully converted into the related carboxylic acids in good to excellent yield with carbon dioxide under very mild reaction conditions (2) (Table 1). The most remarkable advantage of this copper or copper–NHC catalyst system is its wide scope of substrate and functional groups tolerance. The catalytic system is not sensitive to a variety of functional groups, such as ─COOR, ─OH, ─CHO, ─CN, ─NO₂, etc. It provides a powerful tool for the synthesis of highly functionalized propiolic acids.

It is known that copper acetylide is the key intermediate for copper-catalyzed C─H activation of terminal alkynes and the Cu─C bond is active for CO₂ insertion (18, 39–41, 48). A catalytic cycle for copper-catalyzed carboxylation of terminal alkynes with CO₂ is proposed as shown in Fig. 3. The copper acetylide intermediate A was formed from the reaction of the terminal alkyne and L₂CuCl in the presence of a base. Subsequent CO₂ insertion into the polar Cu─C bond will form propynoate intermediate B, in which it will undergo metathesis with the terminal alkyne under basic conditions. This step would release propiolic acid and regenerate intermediate A (Fig. 3A). However, it must be noted that the copper propynoate intermediate B is not stable at elevated temperatures (40). In this reaction, intermediate B may decompose over heat to reform A through a decarboxylation process (Fig. 3B). As the temperature was raised from ambient temperature (25 °C) to 60 °C for the reaction of 1a, the yield of 1b dropped from 90% to 42%. Instead, some homocoupling by-product c was observed. Under room temperature condition, because of the quick insertion of CO₂ into intermediate A and the absence of an oxidant, production of c is prohibited. On heating, intermediate B decomposes to A, and CO₂ may also act as an oxidant in the production of c. The same reaction conducted at 0 °C showed lower activity but high selectivity. This observation is well in agreement with the proposed hypothesis.

Fig. 3.

Proposed catalytic cycle (A) and decomposition pathway (B).

In the P1 catalyst system, it is proposed that the copper center activates the terminal alkyne with a base to form the copper acetylide intermediate, whereas the free carbene activates CO₂ to form NHC carboxylate (Fig. 4). The NHC carboxylate will coordinate to a nearby copper center, which will induce the nucleophilic carbanion of the alkyne into attacking the carboxylic carbon. Following the formation of the new C─C bond, the CO₂ unit is transferred from the carbene center to the copper center, which will be regenerated by metathesis with the alkyne. This system demonstrated an interesting synergistic effect of an organocatalyst and an organometallic catalyst in one system.

Fig. 4.

Proposed copper-carbene cocatalyzed reaction mechanism.

In summary, we have successfully developed a process where copper and copper–NHC catalyzed the transformation of CO₂ to carboxylic acids through C─H bond activation of terminal alkynes. Various propiolic acids were synthesized in good to excellent yields under ambient conditions. The most remarkable advantage of this mild reaction system is its tolerance toward a wide substrate scope. This protocol opens up access to a pool of highly functionalized propiolic acids from CO₂. In addition, the poly─NHC─Cu system demonstrated a concept for the coeffect of organo and organometallic catalysts.

Materials and Methods

All solvents were anhydrous and bought from Sigma-Aldrich (99.8%). The alkynes were used without purification from commercial suppliers, unless otherwise indicated. The carbonates were all dried under vacuum with heating before use. Poly─NHC and poly─NHC─Cu were synthesized on the basis of the literature (46). All reactions were performed in oven-dried (140 °C) or flame-dried glassware under an inert atmosphere of dry N₂ or Ar.

Preparation of P(NHC)_0.5(NHC─Cu)_0.5.

NaO^tBu (60 mg, 0.6 mmol) was added to a DMF (10 ml) suspension of poly-imidazolium (46) (250 mg) in a reaction flask. The reaction mixture was stirred for 1 h, and then CuCl (25 mg, 0.25 mmol) was added. The resulting mixture was stirred at 80 °C for 6 h. The solid product was filtered and dried to obtain a pale yellow powder P(NHC)_0.5(NHC─Cu)_0.5. The catalyst is directly used for reaction. The coexistence of a metal center and free carbene was studied in ref. 46.

General Procedure for Carboxylation of the Terminal Alkynes (1b as an Example).

CuCl (4.0 mg, 0.04 mmol, 2.0 mol%), TMEDA (3.5 mg, 0.03 mmol, 1.5 mol%), and K₂CO₃ or Cs₂CO₃ (2.4 mmol) were added to DMF (4 mL) in the reaction tube (10 mL). CO₂ (balloon) and 2 mmol of terminal alkynes (1a, 204 mg) were introduced into the reaction mixture under stirring. The reaction mixture was stirred at room temperature (about 24 °C) for 16 h. After completion of the reaction, the reaction mixture was transferred to potassium carbonate solution (2 N, 5 mL) and the mixture was stirred for 30 min. The mixture was extracted with dichloromethane (3 × 5 mL), and the aqueous layer was acidified with concentrated HCl to pH = 1 and then extracted with diethyl ether (3 × 5 mL) again. The combined organic layers were dried with anhydrous Na₂SO₄ and filtered and the solution was concentrated in vacuum, affording pure product 1b.