We began our study by using glycosyl stannane
1a bearing nontransferable butyl ligand and 4-fluorophenyl sulfonium salt
2a as model substrates. The choice of the TIPS-protected glycosyl stannane can be rationalized from the compromise of its stability and reactivity. Moreover, novel approaches were designed to avoid contamination by organotin residues, making the reaction more competitive [
54]. In order to accomplish the
C-glycosylation process, the combination of a palladium complex and a copper salt was chosen as the catalysts. Initial screening with a series of fluorides proposed that the tin moiety was substituted and formed polymeric tin precipitation, showing that KF represented the best reactivity in DMF at elevated temperature (
Table 1, entries 1–3). Switching the solvent from DMF to less polar solvent such as toluene or dioxane resulted in a considerable decrease in the yields (
Table 1, entries 4 and 5). The identification of copper salts revealed that CuI was optimal (
Table 1, entry 6). Surprisingly, the yield was improved dramatically when the reaction was conducted at room temperature even shortening the reaction time (
Table 1, entry 7). It might be attributed to the suppression of side reactions under mild conditions [
55]. The examination of other palladium source gave an inferior result (
Table 1, entry 8). We then turned out to use Pd(PPh
3)
4 as the catalyst. Reducing the loading of palladium catalyst from 10 mol% to 2 mol% did not affect the yield significantly (
Table 1, entry 9). However, the reaction afforded a slightly lower yield with 1 mol% palladium catalyst (
Table 1, entry 10). Finally, we reached the optimal conditions for efficient
C-glycosylation in the presence of 2 mol% Pd(PPh
3)
4 and 10 mol% CuI with KF as the base at ambient temperature (
Table 1, entry 9). Control experiments revealed that both Pd and Cu played a crucial role in the coupling (
Table 1, entries 11–13). Reactions conducted without palladium catalyst completely did not give the desired product, the starting material
2a remaining intact, whereas the reaction performed without copper catalyst led to the decomposition of sulfonium salt
2a. Additionally, it was noteworthy that the desilylation of glycosyl donor
1a was not observed during the optimization.