Trade-Off between Redox Potential and the Strength of Electrochemical CO2 Capture in Quinones

Electrochemical carbon dioxide capture recently emerged as a promising alternative approach to conventional energy-intensive carbon-capture methods. A common electrochemical capture approach is to employ redox-active molecules such as quinones. Upon electrochemical reduction, quinones become activated for the capture of CO2 through a chemical reaction. A key disadvantage of this method is the possibility of side-reactions with oxygen, which is present in almost all gas mixtures of interest for carbon capture. This issue can potentially be mitigated by fine-tuning redox potentials through the introduction of electron-withdrawing groups on the quinone ring. In this article, we investigate the thermodynamics of the electron transfer and chemical steps of CO2 capture in different quinone derivatives with a range of substituents. By combining density functional theory calculations and cyclic voltammetry experiments, we support a previously described trade-off between the redox potential and the strength of CO2 capture. We show that redox potentials can readily be tuned to more positive values to impart stability to oxygen, but significant decreases in CO2 binding free energies are observed as a consequence. Our calculations support this effect for a large series of anthraquinones and benzoquinones. Different trade-off relationships were observed for the two classes of molecules. These trade-offs must be taken into consideration in the design of improved redox-active molecules for electrochemical CO2 capture.


Orbital analysis of species in EEC and EECC
The frontier molecular orbitals (FMO) of the neutral Q are shown in Figure S1(a). The singly occupied molecular orbital (SOMO) of Qand the HOMO of Q 2have the same form as the LUMO of Q, which is consistent with electrons being added to the LUMO upon reduction. The LUMO energy has been shown to correlate with the reduction potential. S1 The coefficient of the LUMO of Q on the 2-position is larger than the 1-position so in Q 2-, more electron density lies on the 2-position than the 1-position and this has implication on effect of position of substitution. As discussed in the main text, there is greater delocalisation of electron density when an EWG is substituted at the 2-position, resulting in E • EE being more positive than a substitution at the 1-position.
Natural bonding orbital (NBO) analysis S2 is carried out to optimally transform the DFT wavefunction into localised forms, corresponding to the one-centre lone pair (LP) and twocentre bond (BD) elements. The NBOs involving the O centre which captures CO 2 for each species in the EECC scheme are shown in Figure S1(b-e).
S-3 2 Hydrogen bonding in OH case Figure S2: Effect of intramolecular hydrogen bonding on the OH bond lengths of the neutral and reduced form of 1-OH-AQ from DFT When EDGs are substituted, the reduction potentials are expected to become more negative as the reduced forms become disfavoured. In the case of -OH, however, intramolecular hydrogen-bonding stabilises the Q 2form so an overall increase in E • EE is observed. This is supported by the change in DFT geometry: the hydroxyl proton in Q transfers to quinone O following electrochemical reduction, as depicted by the change in OH bond lengths, as shown in .

S-4
3 Substitutions of Me-series Figure S3: Variation in the reduction potential for two-electron reduction for Me-substituted AQ derivatives Figure S3 shows that the computed E • EE decreases from −1.63 V to −2.13 V against Fc + /Fc upon substitution of one to eight Me, and is generally inversely proportional to the number of Me substitutions made. As more electron density is withdrawn from the Os in the Q 2anion, the reduced form is more delocalised and stabilised against re-oxidation. The increase in reduction potential varies slightly for different positions of substitutions for Me, as seen previously in the case of F substitutions. Overall, increasing the number of EDGs leads to approximately linear decrease in the reduction potentials of AQ, which is the opposite effect by F substitutions shown in the main text.
S-5 Figure S4: Computed Gibbs free energy change and redox potentials for the F-series in the gas and solvated phases.
For a more detailed look at how thermodynamic quantities are changed going from gas to solution phase, the Gibbs free energy change for each chemical step C1 and C2 (Eqs (3,4) in main text) is plotted against the absolute 2-electron reduction potential (Eq (5) in main text) computed for the F-subsituted AQ derivatives in both gas phase and in DMSO, as shown in Figure S4. For each F-series plotted, there is a linear relationship between the reduction potential and the Gibbs free energy change of both chemical steps. In both gas phase and in DMSO, there is a larger Gibbs free energy change and therefore a smaller driving force for the capture of the second CO 2 than the first CO 2 . This is supported by with the O CO 2 bond lengths increasing from 1.495Å to 1.553Å going from Q(CO 2 ) 2to Q(CO 2 ) 2-2 for AQ.
Solvation by the SMD model increases the reduction potentials of all derivatives, accounting for the gain in solvation energy going from the neutral Q to the anion Q 2-. This gain in solvation energy is larger for AQ than for AQ-F 8 as AQ-F 8 in its neutral form is already well solvated. Therefore, the redox window in the solution phase is narrower than that in the gas phase. S-6

DFT reduction potentials and free energy changes
The reduction potentials and free energy changes associated with the EEC and EECC schemes from DFT calculations (B3LYP/6-311++G** level, S3-S6 SMD solvation model with DMSO S7 ) are given in Tables S1, S2 and S3.     Figure S5 shows the CV of O 2 in DMSO, which has a half-wave potential of −1.24 V vs Fc + /Fc. This means that O 2 can potentially react with reduced AQ via a redox process. S-9