Mechanistic Pathway in the Electrochemical Reduction of CO₂ on RuO₂

Abstract

RuO 2 has been reported to reduce CO 2 electrochemically to methanol at low overpotential. Herein, we use density functional theory (DFT) to gain insight into the mechanism for CO 2 reduction on RuO 2 (110). We investigate the thermodynamic stability of various surface terminations in the electrochemical environment and find CO covered surfaces to be particularly stable, although their formation might be kinetically limited at mildly reducing conditions. We identify the lowest free energy pathways for CO 2 reduction to formic acid (HCOOH), methanol (CH 3 OH) and methane (CH 4 ) on partially reduced RuO 2 (110) covered with 0.25 and 0.5 ML of CO*. We find CO 2 is reduced to formic acid, which is further reduced to methanol and methane. At 0.25 ML CO* the reduction of formate (OCHO*) to formic acid is the thermodynamically most difficult step and becomes exergonic at potentials below -0.43 V vs. the reversible hydrogen electrode (RHE). On the other hand, at 0.5 ML CO*, the reduction of formic acid to H 2 COOH* is the thermodynamically most difficult step and becomes exergonic at potentials below -0.25 V vs.RHE. We find that CO 2 reduction activity on RuO 2 changes with CO coverage, which suggests that CO coverage can be used as a tool to tune the CO 2 reduction activity. We show the mechanism for CO 2 reduction on RuO 2 to be different from that on Cu. On Cu, hydrocarbons are formed at high Faradaic efficiency through reduction of CO* at ~1 V overpotential, while on RuO 2 , methanol and formate are formed through reduction of formic acid at lower overpotentials. Using our understanding of the CO 2 reduction mechanism on RuO 2 , we suggest reduction of formic acid on RuO 2 , which should lead to methanol and methane production at relatively low overpotentials.

Descriptors and Thermodynamic Limitations of Electrocatalytic Carbon Dioxide Reduction on Rutile Oxide Surfaces

“…The latter resulti si na greement with previous findings by Karamad et al [29] The calculated methanol onset potentials presented here take into account the potential requirement for OH* removal neededt ok eep the active site availablef or CO 2 reduction.A ll the calculated onsetp otentials are lower than or equal to the OH* removal potentialfor strongly bindings urfaces. Here it is seen that the palladium/platinum oxide overlayer shows thermodynamically limitedH 2 COOH* formation, whereas OH* removal is limiting for the Mo/Nb/Ta/Ti-containing overlayers.…”

“…[25,26] However,i th as been suggested that metal catalysts can be selective for formic acid production over HER if OCHO*i st he reactioni ntermediate instead of COOH* because of the lack of scaling between H* and OCHO*. Following previous theoreticalw ork on the CO2RR mechanism for rutile oxide catalyst, [29] we limit our study to ar educed reaction network with only C 1 products and intermediates that bind through O. This providesadditional scope to suppresst he HER activity over CO2RR on appropriate catalystsurfaces.…”

“…As the RuO 2 (110)s urfaceh as already been studied theoretically in the CO 2 electroreduction mechanism, [29] we use this as the ChemSusChem 2016, 9,3230 -3243 www.chemsuschem.org basis of our study.R uO 2 is found in the rutile-type structure with lattice parameters a = 4.491 and c = 3.106 . As the RuO 2 (110)s urfaceh as already been studied theoretically in the CO 2 electroreduction mechanism, [29] we use this as the ChemSusChem 2016, 9,3230 -3243 www.chemsuschem.org basis of our study.R uO 2 is found in the rutile-type structure with lattice parameters a = 4.491 and c = 3.106 .…”

“…As the RuO 2 (110)s urfaceh as already been studied theoretically in the CO 2 electroreduction mechanism, [29] we use this as the ChemSusChem 2016, 9,3230 -3243 www.chemsuschem.org basis of our study.R uO 2 is found in the rutile-type structure with lattice parameters a = 4.491 and c = 3.106 . [41] Following ap revious study by Karamad et al, [29] our supercell that represents the catalyst surfacec onsists of af ourt rilayer thick 2 1R uO 2 (110)s urface. [41] Following ap revious study by Karamad et al, [29] our supercell that represents the catalyst surfacec onsists of af ourt rilayer thick 2 1R uO 2 (110)s urface.…”

“…This facilitates the modeling of the catalytic activity with fewer parameters, the understanding of why some catalysts are better than others, and allows us to narrow the search window forg ood catalysts. [24][25][26][27][28] Recent theoretical studies [29] indicate that methanol and methane are produced throughadifferent pathway that involves an HCOOH* intermediate on RuO 2 catalysts. [14][15][16] Specifically for fuel-cellelectrocatalysts, the identification of scaling relations [13] and the construction of theoretical activity volcanoes have been used to establish why overpotentials exist.…”