Considering these challenges, we began investigating the reaction using 2-(cyclopropylidene(phenyl)methyl)aniline (
1a) as the substrate under 1 atm of CO
2 (
Table 1). The desired product
2a was obtained in 25% yield with LiO
tBu as the base in DMF at 140 ℃ for 24 h (
Table 1, entry 1). The use of other solvents, such as DMAc or
o-xylene, resulted in lower yields of
2a, while the use of THF resulted in a yield comparable to that obtained using DMF (
Table 1, entries 2–4). This transformation exhibited good performance in a non-polar solvent [
26,
27], while the carbonylation with CO
2 was typically more effective in a dipole solution [
28]. Therefore, we combined the two advantages and used a mixture of a dipole solvent and non-polar solvent as the reaction medium to further screen the reaction. The yield of the desired product considerably improved under these conditions (
Table 1, entries 5 and 6), especially with a combination of DMF and
o-xylene, for which the yield reached 49%. From the perspective of the reaction mechanism, the reason for low yield might be the sluggish rearrangement after carbonylation and many side reactions. Based on this hypothesis, we tested various Lewis acids in the reaction to promote the opening of the ternary ring and improve the efficiency of the reaction. After a series of screenings, we found that the addition of a particular amount of FeCl
3 improved the yield substantially to 80% (
Table 1, entries 7–10. Please see more information in Supporting information), thus confirming the validity of our hypothesis. Next, we evaluated various mixture proportions of solvents, reaction times and amounts of LiO
tBu (
Table 1, entries 11–15). However, no yield better than 80% was achieved. When the amount of LiO
tBu was reduced, there was a noticeable decrease in the yield of the product. Combined with our previous research [
28], this may because LiO
tBu does not only act as a base in the reaction but also fix CO
2 for the formation of intermediates. Therefore, the amount of base has a significant effect on the reaction (
Table 1, entries 14 and 15). Once the optimal solvent proportion, reaction time and amount of LiO
tBu had been confirmed, we further screened various reaction temperatures, aiming for milder conditions. We found a slight enhancement to the product yield to 82% when reacting at 110 ℃ (
Table 1, entries 16). Finally, we confirmed that FeCl
3 (
Table 1, entry 18), CO
2 (
Table 1, entry 19), and LiO
tBu (
Table 1, entry 20) all played a crucial role in promoting the reaction.