Introduction

CO2 is the main component of greenhouse gases and has an important impact on climate change1. However, CO2 is also an abundant, sustainable, and non-toxic one-carbon (C1) building block that could react with olefins2,3,4,5,6,7,8,9,10,11,12,13,14, alkynes15,16,17,18, imines19,20, cyclic ethers21,22,23, halides24,25,26,27,28,29,30,31, arenes32,33,34, etc. to synthesize many types of compounds35,36. Electron-deficient alkenes do not react easily with CO2 due to the descend electron cloud density of carbon-carbon double bond37. To date, despite some hydrocarboxylation methods of electron-deficient alkenes with CO2 have been reported, the reaction substrates were mainly focus on acrylates. Mikami’s group used expensive [RhCl(cod)]2 as catalyst and highly flammable ZnEt2 as reducing agent to achieve the hydrocarboxylation of acrylate with CO2 (Fig. 1a i)38. In recent years, with the rise of green organic synthesis, organic photochemistry has developed rapidly. Yu’s lab achieved β-carboxylation of acrylates with CO2 through an electron donor-acceptor (EDA) photochemical strategy (Fig. 1a ii)39.

Fig. 1: Hydrocarboxylation of electron-deficient alkenes with CO2.
figure 1

a Hydrocarboxylation reactions of acrylates with CO2. b Previous hydrocarboxylation of acrylamides with CO2. c This work: Electrochemical selective hydrocarboxylation of acrylamides with CO2.

Electrochemistry, as another synthesis strategy of green chemistry, has attracted the attention of organic chemists because it avoids the need for chemical oxidation-reduction agents40,41,42,43,44,45,46,47,48. Buckly’s group successfully accomplished the hydrocarboxylation of acrylate with CO2 through electrochemistry, yielding β-carboxylation products (Fig. 1a iii)49. Unlike acrylate, the conjugation of the nitrogen atom and the carbonyl group in acrylamide molecule results in a more dispersed electron cloud around the carbon-carbon double bond, which weakens the nucleophilicity of acrylamide50. Consequently, the hydrocarboxylation of acrylamide with CO2 presents a significant challenge. Despite relevant research reports, these reactions could only be achieved through photocatalytic methods, and were limited to the β-carboxylation of acrylamide (Fig. 1b)51,52.

Based on our research foundations in electrochemical organic synthesis53,54,55,56,57,58,59, and our research interest in CO2 conversion. Herein, we report an electrochemical approach that enables the selective hydrocarboxylation of acrylamide to obtain α- and β-carboxylation products by adjusting solvents and electrolytes under mild reaction conditions (Fig. 1c).

Results and Discussion

We commenced our study using N-methyl-N-phenylmethacrylamide (1a) as the model substrate in an undivided cell under 1 atm of CO2 atmosphere (Table 1). Pleasantly, the β-carboxylation product 2a was obtained in a 70% isolated yield when a magnesium sheet was used as the anode, and TBAC was used as a supporting electrolyte in DMF at room temperature (entry 1). By substituting Mg with Al or Zn as the anode (entries 2, 3), the reaction efficiency was significantly decreased. Surprisingly, the reaction still occurred when graphite carbon was used as the anode (entry 4). This might be due to the oxidation of chloride ions at the anode. When TBAI (entry 5), TBABF4 (entry 6), TBAPF6 (entry 7), and TBAClO4 (entry 8) were employed as the supporting electrolyte, the yield of 2a was low. Several other solvents, including DMAc, NMP, DMSO, and CH3CN, were evaluated; the results were not satisfactory (entries 9-12). However, interestingly, α-carboxylation product 3a was obtained when using DMSO and CH3CN as solvents. This indicates that the solvent was crucial in determining the reaction site. Subsequently, using CH3CN as the solvent, we examined the electrochemical conditions of the α-carboxylation reaction. After replacing TBAC with other electrolytes, the yield of 3a did not increase significantly, whereas the formation of 2a was markedly reduced (entries 13-16). Notably, the use of TBAClO4 as the electrolyte led to the exclusive formation of the α-carboxylation product 3a (entry 17). The synergistic effect of TBAClO4 and CH3CN enables the selective control of the carboxylation reaction at the α position. Employing TEAClO4 as the electrolyte further improved the yield of 3a (entry 18). This might be because, during the initial stage of the reaction, the tetraethylammonium cation with lower steric hindrance is more conducive to stabilizing the carboxylate formed at the α position60. After reducing the concentration of 1a, the reaction efficiency has improved (entry 19). We speculate that the structure of compound 3a may be thermodynamically unstable compared to compound 2a. Therefore, the reaction for generating 3a might be a kinetic-controlled process, and low temperatures may be beneficial for the reaction. Based on this, we attempted to lower the reaction temperature (entries 20-22). The experimental results were in line with expectations, and the target product 3a was obtained with a yield of 70% at -10 °C (entry 21). In contrast, reducing the temperature will not increase the yield of 2a (entry 23).

Table 1 Optimization of the reaction conditions[a]

With the optimal reaction conditions for the selective hydrocarboxylation of acrylamide in hand, the substrate scope was further investigated under these conditions. We first tested the range of substrates for the β-carboxylation reaction (Fig. 2). N-methyl-N-phenylmethacrylamide bearing either electron-donating or electron-withdrawing groups on the benzene ring all reacted smoothly to afford the desired products 2a-2m in 32 to 90% yields. The substrate containing naphthalene ring or tetrahydroquinoline ring was gained products 2n and 2o with the yield of 75% and 53%, respectively. N-phenylmethacrylamide with different N-substituents showed very high reactivity to afford the corresponding products 2p-2s in good to excellent yields. The acrylamide substrates containing different substituents in the olefin could also undergo the reaction smoothly and yield the corresponding target products (2t-2w). When the benzoyl group was substituted for the phenyl group, the β-carboxylation products 2x-2z were still obtained with moderate yields. The structures of products 2r and 2x were determined by X-ray crystallographic analysis.

Fig. 2: Scope of the β-carboxylation reaction.
figure 2

Reaction conditions: undivided cell, Mg plate anode (10 mm × 20 mm × 0.5 mm), C plate cathode (10 mm × 20 mm × 1.0 mm), 1 (0.2 mmol), TBAC (0.2 mmol), CO2 (1 atm), DMF (3.0 mL), 8 mA, 3.5-5 h. a 4 mA, 6 h.

Later, a diverse range of aryl-substituted substrates have been examined in the α-carboxylation reaction of acrylamide with CO2, and the results were summarized in Fig. 3. The substrates containing various electron-donating groups on the phenyl group of compounds 1 were more suitable for this method (3a-3g). The yields decreased modestly when the 4-position of the phenyl group was substituted with electron-withdrawing groups (3h-3l). Substrates containing 3,4-dimethyl-substituted benzene ring, naphthalene ring, or tetrahydroquinoline ring also exhibited moderate reactivity (3m-3o). Altering the nitrogen protecting group to isopropyl, cyclohexyl, or benzyl would also lead to the desired products in medium to high yields (3p-3r). To our surprise, the unprotected NH substrates were also able to produce α-carboxylation products (3s-3t), which may be used for further functionalization to synthesize other kinds of important complex molecules. When there are other substituents at the α- or β-position of the double bond, the target products could be obtained with a moderate yield (3 u, 3 v). Unfortunately, the α-phenyl and β-phenyl substituted acrylamides have not yielded the target compounds due to the influence of electronic effects. Unexpectedly, replacing benzene ring with indole ring, the β-carboxylation products (5a-5f) were detected under the conditions of method B (Fig. 4).

Fig. 3: Scope of the α-carboxylation reaction.
figure 3

Reaction conditions: undivided cell, Mg plate anode (10 mm × 20 mm × 0.5 mm), C plate cathode (10 mm × 20 mm × 1.0 mm), 1 (0.1 mmol), TEAClO4 (0.2 mmol), CO2 1 atm), CH3CN (3.0 mL), 6 mA, 3.5-5 h.

Fig. 4: The other scope of the β-carboxylation reaction.
figure 4

Reaction conditions: undivided cell, Mg plate anode (10 mm × 20 mm × 0.5 mm), C plate cathode (10 mm × 20 mm × 1.0 mm), 4 (0.1 mmol), TEAClO4 (0.2 mmol), CO2 (1 atm), CH3CN (3.0 mL), 4 mA, 6 h.

To demonstrate the practicability of this electrocarboxylation reaction, large-scale reactions between 1a and CO2 were conducted under the conditions of methods A and B, respectively, the results were shown in Fig. 5a. Next, various control experiments were tested in order to gain more insights into the reaction mechanism. The free radical scavengers were added into the reaction systems of the conditions A and B respectively, the hydrocarboxylation reactions were almost inhibited, which suggested that the reactions presumably adopted a radical pathway (Fig. 5b). Meanwhile, compounds 6 and 7 were observed by HRMS, which indicated the existence of α-carbon radical intermediate in method A and CO2 radical anion intermediate in method B. In addition, sodium formate was discovered in the absence of 1a through 1H NMR and 13C NMR under the reaction conditions of method A and method B respectively, the experiment results showed that CO2 could also be reduced in both method A and B (see Supporting Information S5 for more details)39.

Fig. 5: Control experiments.
figure 5

a Large-scale reactions. b Radical trap experiments. c Deuterium labeling experiments. d Proton source exploration experiments.

Deuterium labeling experiments confirmed the presence of β-carboanion intermediates in the reaction pathway of method A (Fig. 5c, i and ii)8,19. And in method B, the β-site was easily deuterated in both CO2 and N2 atmospheres (Fig. 5c, iii and iv), indicating that the proton may have undergone an addition reaction with 1a prior to CO2, and water provided protons in the reaction. After that, we conducted the following experiments to investigate the source of water. We conducted experiments using a solvent dried with anhydrous sodium sulfate and found a significant decrease in the yield of 3a (Fig. 5d, i). And 3a-D was detected when using CD3CN as a solvent (Fig. 5d, ii). The above experimental results indicate that protons mainly originate from the trace water present in acetonitrile and acetonitrile itself61. Perchlorate may have facilitated the proton dissociation process of acetonitrile (Fig. S9). At the same time, the results of DFT calculations (Fig. 6) showed that the β-carboxylation product 2a would be gained whether it is 1a reduction or CO2 reduction32, and the carboxylation of α-site was not initiated by single electron reduction62, which is consistent with the results of deuterium labeling experiments.

Fig. 6: Calculated reduction potentials and reaction diagrams for the carboxylation of 1a.
figure 6

1a is reduced (top) and CO2 is reduced (bottom).

Cyclic voltammetry experiments were conducted to observe the changes of redox profile (Fig. 7). Compared with CO262, 1a is more difficult to reduce (Fig. 7a, red), and the reduction potential of 1a did not change significantly in different solvents (see Supporting Information Fig. S6 for more details). When the solution was saturated with CO2, the profile of the reduction wave changed significantly (Fig. 7a, green), because of the reaction between the radical anion and CO26. The cyclic voltammetry test results indicate that 4a could not only be reduced (-1.73 V) but also easily oxidized (0.07 V) (Fig. 7b, blue). Therefore, indole-type compounds were detected in the reaction solution47,63. However, under the reaction conditions of Method A, no carboxylation products were observed, which might be related to the presence of multiple reduction peaks in this condition (see Supporting Information S15 and Fig. S8 for more details).

Fig. 7: Cyclic voltammetry studies.
figure 7

a Cyclic voltammograms of method A. b Cyclic voltammograms of 4a in CH3CN.

On the basis of the above mechanistic studies, the plausible reaction pathways have been proposed (Fig. 8). In method A, 1a was reduced to a radical anion (), which underwent nucleophilic attack to CO2 to give the carboxylic radical anion (), the radical intermediate () was reduced through another one-electron reduction process to give the carboxylation product 2a after work-up64. The α-carboxylation process proceeded as follows: 1a and proton undergo Markovnikov’s addition, then two single electron reductions occurred, resulting in the formation of a carbon anion intermediate (). Subsequently, the intermediate was attacked by CO2 to form VI, which was then protonated to yield 3a3,65.

Fig. 8: Plausible mechanism.
figure 8

The reaction pathways of α-carboxylation (top) and β-carboxylation (bottom).

Conclusion

Although there have been reports on the hydrocarboxylation reaction of acrylamide with CO2, effective control over the reaction site has not yet been achieved. This study presents an effective electrochemical selective hydrocarboxylation method of acrylamide with CO2 under mild conditions. By selecting electrochemical conditions such as electrolytes and solvents, the selective carboxylation of the α- and β-positions can be easily achieved. Control experiments and DFT calculations results show that the carboxylation of the α- and β- sites follows different reaction processes. Specifically, β-carboxylation is initiated by the reduction of 1a or CO2, while α-carboxylation is initiated by the hydrogenation reduction of 1a.

Methods

Synthesis of compounds 2a-2z (general procedure)

Acrylamide (1) (0.2 mmol) and TBAC (0.2 mmol, 1.0 equiv.) were added to a 10 mL reaction tube containing a stir bar. The reaction tube is then installed with a graphite carbon cathode and Mg plate anode and evacuated and backfilled with CO2 for at least 3 times. Then dry DMF (3.0 mL) was added to the reaction mixture by syringe. The reaction mixture was electrolyzed under a constant current of 8 mA for 3.5 h at room temperature. After the electrolytic reaction, the mixture was carefully quenched with HCl (2 N, 2 mL) and extracted three times with ethyl acetate (3 × 20 mL). The combined extracts were washed with brine, and dried over anhydrous Na2SO4. The solvents were removed under reduced pressure, and the crude residue was purified by flash column chromatography to give the desired products 2a-2z.

Synthesis of compounds 3a-3v and 5a-5f (general procedure)

Acrylamide (1) (0.1 mmol) and TEAClO4 (0.2 mmol, 2.0 equiv.) were added to a 10 mL reaction tube containing a stir bar. The reaction tube is then installed with a graphite carbon cathode and Mg plate anode and evacuated and backfilled with CO2 for at least 3 times. Then dry CH3CN (3.0 mL) was added to the reaction mixture by syringe. The reaction mixture was electrolyzed under a constant current of 6 mA for 3.5 h at −10 °C. After the electrolytic reaction, the mixture was carefully quenched with HCl (2 N, 2 mL) and extracted three times with ethyl acetate (3 × 20 mL). The combined extracts were washed with brine and dried over anhydrous Na2SO4. The solvents were removed under reduced pressure, and the crude residue was purified by flash column chromatography to give the desired products 3a-3v and 5a-5f.