In transition-metal catalysed reductive carboxylation reactions using CO₂, stoichiometric amounts of metallic reductants (eg. Zn, Mn, etc.) or organic sacrificial electron-donors (eg. ⁱPr₂NEt, BI(OH)H, etc.) were usually required to regenerate active low-valent metals or metal hydrides from metal carboxylates. However, these CO₂ carboxylation methodologies accompanied generation of a significant amount of chemical wastes, which is beyond the original intention of green chemistry. Obviously, the best replacement of those reagents is to use the cleanest reductant, H₂. It should be noted that Leitner’s group described a formal hydrocarboxylation of olefins under high temperature and high pressure of CO₂/H₂ in 2013 (Angew. Chem. Int. Ed. 2013, 52, 12119.), but this reaction actually utilized CO generated by water-gas shift reaction.

After the long-term studies of developing novel reduction methods for carboxylations, such as R₃Al, formate salts, photoredox-amine, etc., our group had reached the stage to think about some really challenging transformations, that is, “hydrocarboxylation of olefins with CO₂ and H₂”, a 100% atom-economical reaction (Fig. 1).

Fig. 1. Hydrocarboxylation of olefinic compounds using CO₂ and H₂

Of course, the biggest challenge was how to suppress the undesired hydrogenation of olefins, which was usually considered as a rapid and easy transformation in the presence of metal catalysts. In addition, CO₂ is not very reactive. But it is absolutely true that without trials, no one can succeed. In fact, when we initiated our study based on our previous research on Rh-catalysed hydrocarboxylation reactions (Chem. Commun. 2017, 53, 3098.; Front. Chem. 2019, 7, 371.), we indeed obtained a large amount of hydrogenation product. We struggled for a long time with this unwanted side reaction, which almost brought us to a dead end.

The situation started to change when we found that one of the Buchwald-type ligands, DavePhos, behaved quite unusual as compared to other ligands. The combination of Rh-catalyst and DavePhos showed high reactivity for hydrocarboxylation, while the hydrogenation was somehow suppressed. After extensive optimization of the conditions using Davephos, we have succeeded in developing an acceptable system to afford the desired hydrocarboxylation products in good yields. The reaction was established, but two new questions remained unclarified: (a) Why is DavePhos good? (b) Why is an excess equivalent of DavePhos necessary?

These two mysteries haunted us for almost two years, until we finally defined the structure of a phosphorus-containing species in the reaction mixture to be a phosphonium salt derived from DavePhos. This structure suggested the function of DavePhos in this reaction system to be not only as a ligand, but also as an electron-donor for the initial generation of Rh–H species from the cationic Rh precursor (Fig. 2). This is the most rational explanation of the two mysteries so far. Fig. 2. Generation of Rh–H and observation of DavePhos phosphonium salt

In this study, we have developed a photocatalytic hydrocarboxylation of styrene derivatives with CO₂ and H₂, which is a 100% atomic economical reaction (Fig. 3). We hope that our findings will extend the toolbox of more efficient and sustainable reactions for the utilization of CO₂. Please find more details of the original paper of our work in Nature Communications: “Catalytic direct hydrocarboxylation of styrenes with CO₂ and H₂” (doi: 10.1038/s41467-022-35293-3).

Fig. 3. Catalytic direct hydrocarboxylation of styrenes with CO₂ and H₂