Publication date: 6th November 2020
Gold and iron-doped nitrogen graphene (Fe-N-C) are among the most active and selective catalysts for electrochemical CO2 reduction to CO. However, model DFT studies with standard GGA functionals on these materials predict CO poisoning on Fe-N-C and negligible CO coverage on Au, and the latter is contrary to observations in surface-enhanced spectroscopic experiments. Furthermore, the interfacial electric field and solvation can affect the binding strength, and these effects remain an open challenge for ab initio studies. To determine the CO adsorption energy and to assess the accuracy of DFT-GGA results, we return to a surface science experiment: temperature programmed desorption (TPD). By fitting TPD spectra to a model first order kinetic expression including the configuration entropy of the adsorbate, we determine the adsorption energy and equilibrium coverage of *CO on Au step facets [1] and Fe-N-C [2]. We find that the adsorption energies on these materials are remarkably similar, in contrast to DFT-GGA predictions. The results indicate that hybrid functionals are needed on Fe-N-C in order to accurately capture the binding strength. To investigate the impact of the electrochemical environment, we compare water adsorption energies from TPD with ab-initio molecular dynamics (AIMD) calculations. We find that the competition between water and CO adsorption can significantly affect the CO adsorption strength and result in different binding sites for *CO on Au in gas phase and electrochemical environments.
Funding from the Villum Fonden through the VSUSTAIN project (9455) is gratefully acknowledged. The authors thank Jens K. Nørskov for the discussion regarding TPD spectra. YK would like to acknowledge financial support from the Japan Society for the Promotion of Science (JSPS) KAKENHI Grant-in-Aid for Early-Career Scientists (19 K15360) and JSPS Open Partnership Joint Research Projects/Seminars (JPJSBP 120209925).