Our study addresses two aspects going beyond this state of the art: 1) considering the intermediate, semiquinone, of the reaction (see Figure 1(b)) and thus the thermodynamic overpotential; and 2) incorporating the effects of the electrode/electrolyte interface, where the actual electrochemical processes take place (see Figure 1(c)). Since the exact atomic structure of the carbon felt is unknown, we use different model carbon morphologies to calculate the thermodynamic overpotential of the given reaction.
Grand-canonical DFT in combination with the linearized Poisson–Boltzmann implicit solvent model [Abidi et al., 2020] was used to calculate the overpotential (versus SHE) of the two-step reduction reaction of non-substituted p-benzoquinone (BQ) on 6 different models for carbon surfaces - graphite basal plane (0001), and edges (1100), (1000), zigzag nanotube, armchair nanotube and Buckyball. The overpotential on these surfaces was found to be between 0.2-0.5 V. The solution-phase overpotential due to the intermediate was 0.3 V for non substituted quinone, and 0.5 V for both the substituted quinones that we studied (R=CN and R=OMe). At the graphite basal plane, while for the non-substituted quinone, the overpotential negligibly increased, for the substituted ones, the overpotential reduced by ~0.2 V each (see Figure 1(d)). This showed an interesting relationship between the substituents of the quinone and the electrode surface.
Extending our study further, we explored the effect of defects in the graphite basal plane: substitutional B or N doping, and the Stone-Wales (SW) defect. We found that the overpotential of the BQ reduction reaction did not change for the SW defect, and it increased by 0.1 V if a C-atom was substituted by a B-atom and decreased by 0.1 V, for N-atom substitution. The N-doped surface had the least overpotential (0.25 V), which suggests that doping the electrode with nitrogen would make better electrodes. From these results, we concluded that the atomic-level morphology of the surface does not affect the overpotential of the reaction as much as the dopants in the surface do.
In summary, our results demonstrate that grand-canonical DFT leads to detailed insights on the chemistry of ORFB at the electrode. From an electrochemical viewpoint, we show that estimating the overpotential of ORFB requires the consideration of the intermediate in the reduction reaction, and that substituting the quinones with CN or OMe decreases the overpotential on the electrode surface. Finally, the effect of the atomic-level morphology of the electrode does not significantly affect the overpotential, while substitutional N-doping lowers the overpotential.
References:
Abidi, N., Lim, K. R. G., Seh, Z. W., & Steinmann, S. N. (2020). WIREs Computational Molecular Science, 11(3), e1499.
Er, S., Suh, C., Marshak, M. P., & Aspuru-Guzik, A. (2015).Chemical Science, 6(2), 885 - 893.
Symons, P. (2021). Current Opinion in Electrochemistry, 29, 100759.
Zhang, Q., Khetan, A., & Er, S. (2020). Scientific Reports, 10(1), 22149.
Figure 1: (a) One-step coupled 2e-,2H+ reduction reaction of benzoquinone (Q) into hydroquinone (HQ); (b) Two-step e-,H+ coupled reduction reaction via the intermediate, semiquinone (SQ); (c) Two-step e-,H+ transfer reduction reaction at the interface of carbon electrode and electrolyte; For substituted quinones, we used model examples of electron withdrawing substituents Ri={1,2,3,4}=CN and electron donating sbstituents Ri={1,3}=OMe (d) Overpotential of reduction reaction of different quinones in solution phase (blue) and at the graphite basal plane (0001) (red) is compared. While for non-substituted quinone, the overpotential only slightly increases due to the surface, for the substituted cases, the overpotential due to the surface decreases by 0.2 V.