Mechanism of CO 2 Electroreduction to Multicarbon Products over Iron Phthalocyanine Single-Atom Catalysts

: Carbon dioxide reduction reaction (CO 2 RR) is a promising method for converting CO 2 into value-added products. CO 2 RR over single-atom catalysts (SACs) is widely known to result in chemical compounds such as carbon monoxide and formic acid that contain only one carbon atom (C1). Indeed, at least two active sites are commonly believed to be required for C − C coupling to synthesize compounds, such as ethanol and propylene (C 2+ ), from CO 2 . However, experimental evidence suggests that iron phthalocyanine (PcFe), which possesses only a single metal center, can produce a trace amount of C 2+ products. To the best of our knowledge, the mechanism by which C 2+ products are formed over a SAC such as PcFe is still unknown. Using density functional theory (DFT), we analyzed the mechanism of the CO 2 RR to C1 and C 2+ products over PcFe. Due to the high concentration of bicarbonate at pH 7, CO 2 RR competes with HCO 3 − reduction. Our computations indicate that bicarbonate reduction is significantly more favorable. However, the rate of this reaction is influenced by the H 3 O + concentration. For the formation of C 2+ products, our computations reveal that C − C coupling proceeds through the reaction between in situ-formed CO and PcFe(“0”) − CH 2 or PcFe(“-I”) − CH 2 intermediates. This reaction step is highly exergonic and requires only low activation energies of 0.44 and 0.24 eV for PcFe(“0”) − CH 2 and PcFe(“-I”) − CH 2 . The DFT results, in line with experimental evidence, suggest that C 2+ compounds are produced over PcFe at low potentials whereas CH 4


■ INTRODUCTION
Electrocatalytic CO 2 reduction relies on multiple proton and electron transfers to form a variety of reduction products.Considering the number of carbon atoms, the products in CO 2 electroreduction can be classified into C1 (carbon monoxide, formic acid, methanol, etc.) and C 2+ (ethylene, ethanol, propylene, propanol, etc.) paths. 1 Compared to C1, the C 2+ products exhibit a higher energy density.Furthermore, some multicarbon products are highly demanded chemical feedstocks.−5 For instance, Cu-nanoplate catalysts reported by Ajmal et al. display selectivity toward C 2 H 4 and C 2 H 6 formation. 6In comparison with the Cu-planar, the Cu-nanoplate catalyst exhibited a 24 times increase in faradaic efficiency.Koper et al. investigated the reaction mechanism of electrochemical CO 2 reduction to C1 and C 2+ on copper electrodes. 7According to their suggestion, the key intermediate for the C 2+ pathway is the formation of a CO dimer through electrochemical C−C coupling.This CO dimer is then reduced to form the C 2+ products.In addition to uniform Cu catalysts, copper alloy materials, which have tunable adsorption strengths for key intermediates, can facilitate CO 2 reduction to C 2+ products.Using density functional theory (DFT), Zhi et al. analyzed the challenging kinetics of C−C coupling for electrochemical CO 2 reduction on copper alloys with Pt, Pd, or Au. 8 They concluded that OC−COH coupling is kinetically more favorable than OC−CHO coupling on the copper alloy surfaces due to the increase in the reactivity of the adsorbed CO species.However, effective CO 2 reduction to C 2+ species remains a challenge 9−12 due to the following factors: (1) it competes with the undesired hydrogen evolution reaction (HER), 13,14 (2) a high activation barrier needs to be overcome for C−C couplings to form long hydrocarbon chains (C 2+ products), 15 and (3) there is a high number of possible reaction pathways for C 2+ formation. 15t this point, it appears from most of the literature that at least two active sites on the catalyst are required to produce C 2+ .This is due to the fact that C 2+ formation requires two nearby active sites for carbonous intermediates, which then can couple.However, our previous study 16 surprisingly demonstrated that C 2+ products can also be formed over single-atom molecular catalysts like iron phthalocyanine (PcFe) albeit with a very low faradaic efficiency, while these systems are wellknown to only produce C1 products. 17,18Based on the abovementioned study, 16 PcFe is capable of producing light hydrocarbons such as C 2 H 4 and C 3 H 6 through the electrocatalytic CO 2 reduction in aqueous solutions under neutral pH conditions.Based on their experimental evaluations, we suggested that CO is an essential reactant in the pathways of CO 2 reduction to C 2+ .There is, however, no clear understanding of how C−C coupling occurs over PcFe, which contains only one active site (Fe).Developing a deeper understanding of the C−C coupling mechanism over PcFe could lead to the design of more effective single-atom catalysts (SACs) with a higher faradic efficiency for C 2+ .
Independent of the target products, CO 2 is commonly believed to be the starting point of the carbon dioxide reduction reaction (CO 2 RR). 19,20−23 As a result of carbonate equilibrium in the aqueous solution, CO 2 reduction competes with carbonate reduction. 21Indeed, HCO 3 − and CO 3 2− dominate the equilibrium at neutral to alkaline pH, while CO 2 and H 2 CO 3 dominate at acidic pH.According to our previous studies, 21,24 bicarbonate is one of the key reactants in CO 2 reduction reactions over Fe porphyrins.Accordingly, it is necessary to consider bicarbonate or carbonic acid reduction when studying the CO 2 reduction reaction in an aqueous solution.
In this study, we will present the mechanisms of the CO 2 RR toward C 2+ products over PcFe (Figure 1a).The understanding of the detailed mechanisms for the formation of C 2+ products on PcFe facilitates the development of more effective catalysts.The mechanisms presented in this study are derived from DFT calculations.

■ COMPUTATIONAL DETAILS
All DFT calculations were performed using Gaussian 16 25 at the M06-L level of theory. 26−33 For all the calculations, solvent effects were considered using the SMD solvation model 34 with water as the solvent.The SDD basis set 35,36 with effective core potential was chosen to describe iron.The 6-31G(d) basis set 37 was used for all other atoms.Frequency calculations were carried out at the same level of theory as those for structural optimization.Transition states were located by using the Berny algorithm.Intrinsic reaction coordinate calculations were used to confirm the connectivity between transition structures and minima. 38,39To further refine the energies obtained from the SMD/M06-L/SDD,6-31G(d) calculations, we carried out single-point energy calculations using the M06-L functional method with the SMD solvation model in water along with the def2-TZVP basis set 32 on all atoms for the optimized structures.The tight convergence criterion and ultrafine integral grid were utilized to enhance the accuracy of the single-point calculations.In the case of bicarbonate reduction, where we were unable to find the TS, potential energy scans were utilized (see Supporting Information).The most likely spin states were also tested, and the reaction profiles have been developed based on the most energetically favorable spin state.Accordingly, we assume that the spin changes only by ±1/2 in electrochemical steps.In chemical steps, the spin sate remains the same.
A computational normal hydrogen electrode was used to model proton-coupled electron transfer (PCET), which assumes that H 2 in the gas phase is a reference. 40In order to compare to the experimental results, all potentials were then converted to RHE.With formic acid as a reference compound, pK a values were determined by the isodesmic method. 41To correct the deficiencies of the implicit solvation model, eq 1 was used to scale the obtained pK a values. 42+ K K p (scaled) 0.49p (DFT) 3.2 a a (1) The effective absolute potential method was used to compute pure electron transfer steps which were not followed by an acid−base reaction. 43Transition states for proton transfer reactions were modeled using an explicit water cluster containing two H 2 O and one H 3 O +44,45 in conjunction with the SMD.The TS values were then calculated based on the Gibbs free energy difference between the TS and the optimized structures of the reactant solvated by the same water cluster.A water solution with a pH of 7 was assumed as the electrolyte.
■ RESULTS AND DISCUSSION Speciation.To determine the active form of PcFe under the reaction conditions (pH = 7), pure electron transfer without protonation was calculated. 21,24According to our computations, the reduction of formal PcFe("II") to PcFe("I") by electron transfer requires a potential of −0.48 V (Figure 1b).Upon receiving the second electron, PcFe("I") forms PcFe("0") at a potential of −0.73 V.In a subsequent step, PcFe("0") is reduced to PcFe("-I") at −0.91 V (Figure 1b).The oxidation states mentioned in this study are formal, and the additional electrons in each step can be delocalized to the ligand.As determined by DFT calculations, H 2 O weakly binds to the catalyst in all oxidation states and the PcFe−O distance is almost 3 Å in all cases.Therefore, binding of water to PcFe can be neglected.
Using these calculations, it is possible to identify the most likely oxidation state of the active species under the reaction conditions.Considering a potential window of −0.6 to −1.0 V, PcFe("I"), PcFe("0"), and PcFe("-I") would appear to be the most likely active species.In order to simplify the explanation, The Journal of Physical Chemistry C we will refer to each intermediate by its oxidation state at the beginning of the mechanism, e.g., the PcFe("-I")−COOH intermediate originated from a bare PcFe("-I") catalyst.The same logic is applied to the naming of other intermediates.All voltages in this study are expressed vs RHE.
Hydrogen Evolution Reaction.The HER is always in competition with the CO 2 RR. 46,47It is therefore crucial to evaluate the HER before discussing the CO 2 RR mechanism.As the first step, the Volmer reaction (reaction 2) initiates the HER by forming PcFe−H.
Owing to the likely thermodynamic and almost barrierless reaction, the second PCET is both thermodynamically and kinetically favorable over all PcFe oxidation states.
Overall, the HER is blocked over PcFe("I") at pH = 7 as a result of the very acidic pK a of PcFe("I")−H.The low activation energies for more reduced complexes, on the other hand, indicate that H 2 is produced over PcFe("0") and PcFe("-I").Therefore, H 2 is, in line with experimental evidence, 16 likely to be formed at applied potentials lower than −0.7 V.
Indeed, the lower the applied potential, the greater the amount of H 2 formed in experiments.
CO 2 Reduction Reaction.PcFe−CO Formation.Considering different configurations of CO 2 over PcFe (see Supporting Information for details), carbon dioxide most likely binds via the carbon atom to form a PcFe−C bond. 48,49his reaction is highly endergonic by 0.61 eV over PcFe("I").Fe("I")'s poor nucleophilicity limits its ability to donate electron density for a PcFe−C bond formation, which is confirmed with the 0 net Mulliken charge transfer from PcFe to CO 2 .The O−C−O angle and the 3.2 Å PcFe−C distance are further indications that CO 2 is not activated on PcFe("I").An electron transfer to PcFe("I") forms PcFe("0") and decreases the PcFe−C bond formation energy to 0.49 eV (Figure 3; left arrow).However, PcFe("0")−CO 2 formation is still fairly endergonic.Indeed, the considerable loss of entropy associated with CO 2 adsorption contributes to the endergonic reaction energy.CO 2 is activated on PcFe("0") by altering the linear O−C−O bond angle (180°) to 138°, but still a high activation energy of 0.96 eV must be overcome (Figure 3; left arrow).This is accompanied by a charge transfer of −0.64 electrons to CO 2 .Over PcFe("-I"), PcFe−C bond formation is, finally, exergonic by −0.09 eV and possesses a rather low activation energy of 0.41 eV (Figure 3; left arrow).The significant Mulliken charge transfer of −0.97 e from PcFe("-I") to CO 2 indicates the formation of a PcFe−COO carboxylate and a strong PcFe−C bond.Overall, the CO 2 RR is inherently slow in all oxidation states when assuming CO 2 as the only reactant.The only exception is PcFe("-I").This is in line with our previous works on Fe porphyrins. 21,24An alternative to direct CO 2 reduction is the reduction of bicarbonate or carbonic acid through an S N 2 mechanism 21,24 (Figure 3; right arrow).Bicarbonate is present in high concentrations in the electrolyte at pH = 7.As shown in Figure 3 (right arrow), the S N 2 reaction for bicarbonate to form PcFe("0")−COO and PcFe("-I")−COO is exergonic by −0.82 and −1.41 eV, respectively.Similar to CO 2 activation, bicarbonate is not activated over PcFe("I") for the PcFe("I")−COO intermediate formation.
The activation energies for bicarbonate reduction are however 0.56 and 0.36 eV over PcFe("0") and PcFe("-I") assuming H 3 O + as the proton donor.The energy barrier is much lower than that required for CO 2 activation over PcFe("0").This, combined with the higher HCO 3 − concentration, renders the bicarbonate reduction path more likely for PcFe("0") (Figure S3).However, the bicarbonate reduction rate is sensitive to the proton concentration (Figures S3 and  S4).This results in the reduction of CO 2 over PcFe("-I") being more likely than bicarbonate reduction (Figure S4).As reported in the recent study by our group, 50 in a control experiment with Ar-saturated 0.1 M KHCO 3 at pH ≈ 8.5 (without purging any CO 2 gas), we only observed an extremely faint signal of CO, C 2+ over PcFe.Since there is almost no CO 2 in this experimental condition, the trace amount of carbon products is likely the result of bicarbonate reduction.Compared to pH = 7 where the calculations are performed, raising the pH reduces the proton concentration, which is critical in the suggested mechanism for bicarbonate reduction. 21Therefore, the rate of bicarbonate reduction also declines at higher pH values (Figures S3 and S4).
Overall, only PcFe("0") and PcFe("-I") are able to activate CO 2 with the formation of PcFe−COO.Considering the speciation, it is most likely that PcFe−COO is formed only at

The Journal of Physical Chemistry C
In PcFe−COO, H 3 O + prefers to attack the highly nucleophilic oxygen rather than the carbon. 24,51−55 Thus, PcFe−COO is reduced through a PCET to the oxygen atom to form PcFe−COOH.This reaction proceeds on PcFe("I") and PcFe("0") already at relatively high potentials of −0.45 and −0.47 V (Figure 3; right arrow).However, due to a very low pK a of 0, PcFe("I")−COO quickly loses its proton at pH = 7.In the case of PcFe("-I"), a potential of −0.91 V is required for the formation of PcFe("-I")−COOH, which is still achievable within the range of assumed applied potentials (−0.6 to −1 V).
PcFe("0")−COOH and PcFe("-I")−COOH species remain protonated due to the high pK a value of 9. Regardless of PcFe's oxidation states, PcFe−COOH formation is barrierless using H 3 O + as the proton source (Figure 3; right arrow).This is not surprising since the lone pair electron of oxygen in PcFe− COO is easily shared with H 3 O + .The second proton-coupled electron attacks OH in PcFe−COOH to form PcFe−CO through water elimination.This reaction requires even more positive potentials of 0.14 and 0.13 V for PcFe("0") and PcFe("-I").Additionally, the formation of PcFe−CO is almost barrierless for both oxidation states (Figure 3; right arrow).
Overall, PcFe("I") is unable to activate CO 2 .The highly endergonic PcFe−COO formation over PcFe("0") renders the binding of CO 2 unlikely.In addition, CO 2 binding to PcFe("0") requires the overcoming of a high activation barrier of 0.96 eV.This is in contradiction to the experiment 16 in which CO is produced already at a moderately negative potential of −0.7 V. Bicarbonate reduction, on the other hand, requires a lower reaction energy of −0.82 eV and an activation energy of 0.56 eV over PcFe("0").Accordingly, the experimental results for the CO formation at −0.7 V can only be explained by considering bicarbonate as the reactant.It is noteworthy that bicarbonate reduction is, even for PcFe("-I"), still the kinetically favorable route.In addition, the concentration of bicarbonate is considerably higher than that of CO 2 at pH = 7.Therefore, bicarbonate is, regardless of the oxidation states, the most likely reactant for the CO 2 RR.
CO Release vs Post-CO.Once PcFe−CO has been formed, it can either release CO or be converted into post-CO intermediates.The route to the post-CO products is initialized by hydrogenation.PcFe−CO can form either PcFe−COH (reaction 4) or PcFe−CHO (reaction 5).Alternatively, it may also be released as CO (reaction 6).
CO is almost thermoneutral over PcFe("0") and PcFe("-I").Thus, CO formation is overall favorable (Figure 4; left arrow).However, the activation barrier required for CO release increases from 0.29 eV (PcFe("0")) to 0.44 eV for PcFe("-I") (Figure 4; left arrow).Accordingly, at lower potentials, the release of CO is less likely.This can be explained with the higher electron density on Fe, which enhances the π backdonation to the π* orbital of CO.This increases the PcFe−CO bond strength which, in turn, is responsible for the larger activation barrier.This effect is also visible in the PcFe−C bond length, which decreases from 1.88 to 1.82 Å, while the C−O bond length increases from 1.18 to 1.20 Å for PcFe("0") to PcFe("-I"), respectively (Figure 5).Therefore, more negative potentials shift the selectivity from CO toward post-CO products, which is observed experimentally. 16Fe−COH formation is unlikely due to the very low potential of almost −2.10 V independent of the formal oxidation state (see Supporting Information).The hydrogenation of PcFe−CO to form PcFe−CHO has a potential of approximately 0 V for all oxidation states (Figure 4; right arrow).Owing to the pK a values of 12 and 16, PcFe("0")− CHO and PcFe("-I")−CHO stay protonated at pH = 7. PcFe− CHO formation is a barrierless reaction for both formal oxidation states using H 3 O + as the proton source (Figure 4; right arrow).The high reactivity of PcFe−CO toward H 3 O + is attributed to the bent Fe−C−O angle of 149°(Figure 5), which indicates the existence of a nonbonding electron in the valence shell of the carbon.These results clearly indicate that the hydrogenation of PcFe−CO to form PcFe−CHO is more likely than CO release.Accordingly, a greater amount of post-CO products than the trace amounts observed experimentally 16 could be expected.This is, however, contrary to the experiment, which indicates that, irrespective of the applied potential, always the amount of CO is significantly higher than that of post-CO products. 16Thus, the activation energy cannot solely explain the observed selectivity.Indeed, the proton supply is highly limited by the very low concentration of H 3 O + at pH = 7 ([H + ] = 10 −7 ).The CO release reaction, on the other hand, is independent of H 3 O + and consequently is more likely than PcFe−CHO formation.Although the CO release is a more favorable path, PcFe−CHO formation is still feasible over PcFe("0") and PcFe("-I").
Once the barrier associated with the initial hydrogenation step is overcome, subsequent PCET can result in the formation of either PcFe−CHOH (reaction 7) or PcFe−CH 2 O (reaction 8).The formation of PcFe−CHOH is unlikely due to the very low potential of −1.39 and −1.57V required for PcFe("0")− CHOH and PcFe("-I")−CHOH, respectively.A much higher potential of −0.29 V and a moderate activation barrier of 0.63 eV are, on the other hand, required for PcFe−CH 2 O formation from PcFe−CHO on PcFe("0") (Figure 4; right arrow).The potential increases to 0.11 V and the activation energy decreases to 0.52 eV, which result in even more likely PcFe−CH 2 O formation on PcFe("-I").This is followed by a PCET to the O atom in PcFe−CH 2 O, leading to PcFe− CH 2 OH (reaction 9).This reaction requires a potential of −0.05 and −0.50 V over PcFe("0") and PcFe("-I"), Overall, a lower applied potential results in a stronger PcFe− CO bond, which then requires a higher activation energy to release CO.Thus, the post-CO pathway opens at lower potentials as CO is more strongly bound to Fe in PcFe−Pc.Accordingly, the stronger the PcFe−CO binding, the greater the probability of entering the post-CO pathway.Indeed, the reaction is not limited by an overly high activation barrier or a too negative onset potential but is merely a result of the limited proton supply.
Methanol vs Methane Formation.Once PcFe−CH 2 is formed, the mechanism can bifurcate toward either CH 3 OH or CH 4 formation.Both paths are shown in Figure 6.Methanol is formed through hydrogenation of the carbon in PcFe− CH 2 OH, which then leads to CH 3 OH release from the catalyst (reaction 10).
This reaction is energetically favorable, as indicated by the very positive potential of 1.09 and 1.16 V for PcFe("0") and PcFe("-I"), respectively (Figure 6; left arrow).The very high TS value of 1.03 eV associated with this protodemetalation reaction, however, inhibits the formation of methanol over PcFe("0") (Figure 6; left arrow).For PcFe("-I")−CH 2 OH, H + is immediately drawn to the OH group and water elimination occurs; thus, it follows the CH 4 pathway (reaction 11).
According to our kinetic analysis, PcFe−CH 3 is rapidly formed through a barrierless proton transfer process for both oxidation states using H 3 O + as the proton source.Methane is then released in a final PCET step, which again requires only potentials of 1.44 and 1.45 V for PcFe("0") and PcFe("-I"), respectively (reaction 13 and right arrow in Figure 6).
Overall, the methanol pathway requires activation energies much higher than those of the PcFe−CH 2 pathway.Thus, the CH 4 pathway is more favorable than methanol formation.Based on our DFT calculations and in line with experimental evidence, 16 it appears that no methanol is produced over PcFe, while CH 4 is the major post-CO product.
C−C Coupling.Thus far, we have evaluated only the possible mechanisms for H 2 , CO, and CH 4 formation over PcFe in different formal oxidation states.Nevertheless, experiments clearly show that also minor amounts of C 2+ products are formed. 16It remains unclear how and where C 2+ is formed in the CO 2 RR mechanism over FePc.To address this key question, we first evaluated the CO insertion reaction, a common reaction in organometallic chemistry. 57However, in the present case, this very common mechanism was found to be unfavorable (see Supporting Information).This is not surprising since Fe does not possess a free adsorption site for CO.
Another route proceeds through the direct attack of CO on the carbon atom in PcFe−CH 2 (for other considered routes, see Supporting Information).This CO attack is in competition with the hydrogenation of PcFe−CH 2 (CH 4 formation path).Identical to the formation of CH 4 , the mechanism is divided into a CO attack (reaction 14 and the right arrow in Figure 7) and hydrogenation steps (reaction 12 and the left arrow in Figure 7).
CO on the carbon in PcFe−CH 2 is strongly exergonic by −1.09 eV for PcFe("0") and −1.12 eV for PcFe("-I") (Figure 7; right arrow).This reaction led to the formation of a PcFe− C 2 H 2 O intermediate.Only small activation energies of 0.44 and 0.24 eV need to be overcome for PcFe("0") and PcFe("-I") (Figure 7; right arrow).Solely based on activation energies, PcFe−CH 3 (left arrow) and, thus, CH 4 are the most likely products.Accordingly, C 2+ products should not be formed (Figure 7).However, the TS energy is again not the only parameter that affects the rate of the reactions.In fact, low concentrations of the reactants may significantly reduce the rate of the reaction.According to our DFT results and in line with experiment, 16 CO is extensively produced and released into the double layer; thus, a high concentration of CO is expected close to the catalyst surface.The recent study by us demonstrated, using finite element calculations, that the concentration of CO close to the electrode (almost within 5 mm from the working electrode) is in the range of 1.5−4.5 mM. 16Meanwhile, the CH 4 path entails the use of H 3 O + as a proton source.In pH = 7, the concentration of H 3 O + is 10 −7 M, which is 30,000 times less than that of CO.Furthermore, the catalyst continuously consumes H 3 O + in the double layer, The Journal of Physical Chemistry C thereby resulting in a concentration of H 3 O + even lower than 10 −7 M. The C−C coupling (Figure 7; right arrow) is therefore feasible, but the formation of CH 4 (Figure 7); left arrow) is still a more likely path due to the exponential effect of the activation energy in the Arrhenius rate equation.Consequently, CH 4 is the main post-CO product, while a trace amount of C 2+ products are also expected only at very negative applied potentials where the activation energy decreases to 0.24 eV for the CO attack on PcFe("-I").The C 2+ formation path is very unlikely over PcFe("0") owing to the high energy barrier of 0.44 eV for PcFe("0")−C 2 H 2 O formation compared to the barrierless reaction for the formation of PcFe("0")− CH 3 .These are in good agreement with the experimental evidence of CO 2 RR over PcFe where the formation of CH 4 is observed starting from −0.8 V, while that of C 2+ products only starts from −0.9 V. 16 In PcFe("-I")−C 2 H 2 O, the PcFe−C bond readily switches from one carbon to another through a rearrangement reaction (reaction 15 and Figure 7).This reaction is barrierless and exergonic by −0.96 eV (Figure 7; right arrow).In the subsequent PCET, the O side in PcFe("-I")−C 2 H 2 O is hydrogenated to form PcFe("-I")−C 2 H 2 OH at the potential of −0.57V. Using H 3 O + as the proton source, this reaction is again barrierless.
A PCET to PcFe("-I")−C 2 H 2 forms PcFe("-I")−C 2 H 3 .This reaction is barrierless and takes place at a very high potential of 1.18 V (reaction 18 and Figure 8; right arrow).However, this PCET step rate is limited by the concentration of the proton at pH = 7.Both the CO attack for C3 formation and a PCET to PcFe("-I")−C 2 H 2 are feasible.The latter reaction is, however, slightly more favorable unless there is a limitation on proton supply.
From the PcFe("-I")−C 2 H 3 intermediate, the mechanism is again divided into C 2 H 4 release (reaction 19 and Figure 9; left arrow) and C 2 H 6 formation (reaction 20 and Figure 9; right arrow) paths.A PCET on the carbon bonded to Fe in PcFe("-I")−C 2 H 3 releases C 2 H 4 at a very positive potential of 0.77 V and a low activation energy of 0.26 eV associated with the protodemetalation reaction (Figure 9; left arrow).The PCET to the second carbon (the carbon that is not bonded to Fe) in PcFe("-I")−C 2 H 3 is even more kinetically favorable with an energy barrier of 0.07 eV (Figure 9; right arrow).In the end, this pathway leads to the formation of C 2 H 6 .Thus, PcFe("-I") is expected to produce a greater amount of C 2 H 6 than C 2 H 4 .(20) In the next step of C 2 H 6 formation, PcFe("-I")−C 2 H 4 forms PcFe("-I")−C 2 H 5 through PCET with a high potential of 1.21 V.As this reaction is barrierless, it is also kinetically quite likely (Figure 9; right arrow).In the following reaction, C 2 H 6 is released by another PCET.This reaction requires an activation energy of 0.37 eV and a positive potential of 1.41 V (Figure 9; right arrow).
Overall, the CH 4 pathway wins the competition over the formation of C 2+ products.However, given the very low concentration of H 3 O + at pH = 7 and the high concentration of CO produced by CO 2 close to the electrode, C 2+ pathways are also feasible at low potentials since they mostly depend on the C−C coupling step.This is also consistent with the experimental observations where only trace amounts of C 2+ products are observed at low potentials. 16The overall proposed mechanism is illustrated in Figure 10.

■ CONCLUSIONS
In conclusion, our study has shed light on the unexpected formation of C 2+ compounds during CO 2 reduction over a SAC, namely, PcFe.While the CO 2 RR over SACs is typically associated with the generation of single-carbon (C1) products, PcFe has shown the ability to yield small amounts of larger C 2+ products, thereby challenging the conventional understanding that C−C coupling necessitates multiple active sites.Our The Journal of Physical Chemistry C investigations, employing DFT, have elucidated key aspects of C 2+ products formation mechanism over SACs, particularly the C−C coupling step.According to our results, we suggest a pathway involving a reaction between in situ-generated CO and a PcFe−CH 2 carbene intermediate with nucleophilic character.These reactions were found to be highly exergonic, requiring low activation energies of 0.44 and 0.24 eV for PcFe("0")−CH 2 or PcFe("-I")−CH 2 .This study contributes to the understanding of how C−C coupling is achievable over SACs and molecular catalysts and provides opportunities for

Figure 3 .
Figure 3. CO 2 reduction vs bicarbonate reduction to the PcFe−CO intermediate over PcFe("I"), PcFe("0"), and PcFe("-I") at pH = 7. Potentials are reported vs RHE.Pure chemical steps are specified by red arrows.In all steps, the TS values are only calculated for chemical reactions.

Figure 5 .
Figure 5. Summary of important bond lengths for the Fe−CO intermediate.

Figure 6 .
Figure 6.Methanol release (left arrow) vs further reduction of the PcFe−CH 2 OH intermediate to CH 4 release (right arrow) over PcFe("0") and PcFe("-I") at pH = 7. Potentials are reported vs RHE.In all steps, the TS values are only calculated for chemical reactions.

Figure 7 .
Figure 7. CH 4 release (left arrow) vs C−C coupling on PcFe−CH 2 (right arrow) over PcFe("0") and PcFe("-I") at pH = 7. Potentials are reported vs RHE.Pure chemical steps are specified by red arrows.In all steps, the TS values are only calculated for chemical reactions.
PcFe("-I")−C 2 H 2 OH to form PcFe("-I")− C 2 H 2 requires a very low activation energy of 0.36 eV.Chemically, the PcFe("-I")−C 2 H 2 intermediate is quite similar to the PcFe("-I")−CH 2 intermediate with respect to the metal carbene type complexes.Hence, the same reaction mechanism observed may also be active for the formation of longer carbon chains.

Figure 8 .
Figure 8. C3 formation path (left arrow) vs C 2 H 4 and C 2 H 6 formation (right arrow) over PcFe("-I") at pH = 7. Potentials are reported vs RHE.Pure chemical steps are specified by red arrows.In all steps, the TS values are only calculated for chemical reactions.

Figure 9 .
Figure 9. C 2 H 6 release (left arrow) vs C 2 H 6 formation path (right arrow) over PcFe("-I") at pH = 7. Potentials are reported vs RHE.In all steps, the TS values are only calculated for chemical reactions.

Figure 10 .
Figure 10.Overall suggested mechanism for CO 2 reduction over Fe Phthalocyanine.Only paths to products which are thermodynamically and kinetically accessible are displayed..
This is in line with experimental evidence demonstrating that the amount of C 2 H 6 produced is almost two times that of C 2 H 4 .16