<p>CO is the simplest product from CO<sub>2</sub>
electroreduction (CO<sub>2</sub>R), but the identity and nature of its rate
limiting step remains controversial.
Here we
investigate the activity of both transition metals (TMs) and metal-nitrogen
doped carbon catalysts (MNCs), and a present unified mechanistic picture of CO<sub>2</sub>R to for both these classes of
catalysts. By consideration of the
electronic structure through a Newns-Andersen model, we find that on MNCs, like
TMs, electron transfer to CO<sub>2</sub><sub>
</sub>is facile, such that CO<sub>2</sub>
(g) adsorption is driven by adsorbate dipole-field interactions. Using density
functional theory with explicit consideration of the interfacial field, we find
CO<sub>2</sub> * adsorption to
generally be limiting on TMs, while MNCs can be limited by either CO<sub>2</sub>* adsorption or by the proton-electron transfer reaction
to form COOH*. We evaluate these computed mechanisms against pH-dependent experimental
activity measurements on CO<sub>2</sub>R
to CO activity for Au,
FeNC, and NiNC. We present a unified
activity volcano that, in contrast to previous analyses, includes the decisive CO<sub>2</sub>*<sub> </sub>and COOH* binding strengths as well as the
critical adsorbate dipole-field interactions.
We furthermore show that MNC catalysts are tunable towards higher
activity away from transition metal scaling, due to the stabilization of larger dipoles resulting from their
discrete and narrow <i>d</i>-states. The analysis suggests two design principles
for ideal catalysts: moderate CO<sub>2</sub>* and COOH* binding strengths
as well as large dipoles on the CO<sub>2</sub>*<sub> </sub>intermediate. We suggest that these principles can be
exploited in materials with similar electronic structure to MNCs, such as supported
single-atom catalysts, molecules, and nanoclusters, 2D materials, and ionic
compounds towards higher CO<sub>2</sub>R
activity. This work captures the
decisive impact of adsorbate dipole-field interactions in CO<sub>2</sub>R to CO and paves the way for computational-guided
design of new catalysts for this reaction.</p>