The electrochemical reduction of CO₂ is a promising technology that could potentially provide carbonneutral fuels with relatively high coulombic efficiency.^1-3) The CO₂ reduction using homogeneous electrocatalysts has been extensively studied due to its intrinsic advantage of the relative low overpotential compared to the direct CO₂ reduction on metal electrodes [1-3]. Among the homogeneous electrocatalysts, Ni(II)-cyclam (cyclam: 1,4,8,11-tetrazazcyclotetradecane) and its derivatives have been known to exhibit remarkable efficiency and selectivity in converting CO₂ to CO [4-9]. It has been revealed that Ni(I)-cyclam, generated by one-electron-reduction of Ni(II)-cyclam, is the active catalytic species and the cyclam ligand stabilizes the Ni(I) state [5,7,10]. Ni(II)-cyclam, however, still exhibits rather high overpotential (ca. −0.6 V), leading to the low energy efficiency for CO₂ conversion. In addition, Ni(II)-cyclam necessitates a mercury electrode to avoid hydrogen evolution and shows much less electrocatalytic activities on other environmentally-benign electrodes [5-7].

In contrast to the extensive works on Ni(II)-cyclam, the studies on the cyclam complexes of other transition metals (e.g., Co, Cu) are scarce [11-15].It has been known that Cu(I)-cyclam is unstable so that Cu(I) ion is liberated from the complex upon a cathodic polarization, which is a serious drawback as an electrocatalyst [12,13]. This demetalation issue, however, has not been fully addressed yet. It was reported that the both the smaller and larger macrocyclic complexes than 14-membered cyclam leads to stabilization of Cu(I) state, and thus suppressed demetalation behavior [12,13]. It was also claimed that the N-alkylation of cyclam stabilizes the Cu(I)- and Ni(I)-cyclams both by increasing the cavity size and by decreasing the ligand field strength imparted by the nitrogen donor atom [15], which leads to a positive shift in the M(II)/M(I) potential [14,15]. In the prior works, however, the catalytic activities of the Cu(II)-complexes for CO₂ reduction were not reported.

In this study, Cu(II) complexes of cyclam, propylene cross-bridged cyclam (PCB-cyclam), and propylene cross-bridged cyclam diacetate (PCB-TE2A) (Fig. 1) were prepared, and investigated as homogeneous electrocatalysts for CO₂ reduction. With an aim to address the demetalation issue of Cu(II)-cyclam, N-alkylation with a cross-bridging propylene group in Cu(II)-PCB-cyclam was adopted with the following two rationales: (1) altering the macrocycle size to stabilize the central Cu ion, (2) blocking one of the two open sites of Cu(II)-cyclam. In Cu(II)-PCB-TE2A, N-diacetate groups are supposed to add further steric and electronic effects on Cu(II)-PCB-cyclam. The redox properties and demetalation behaviors of the Cu(II)-complexes were examined using cyclic voltammetry in aqueous solution on glassy carbon electrode in comparison with Ni(II)-cyclam.

Before investigating the electrocatalytic activity of Cu(II)-cyclam derivatives, the redox reactions of Ni(II)-cyclam, the most representative electrocatalyst for CO₂ reduction, was examined for comparison. The cyclic voltammograms (CVs) of Ni(II)-cyclam in tri-HCl buffer solution (pH 8.25) are presented in Fig. 2. In a N₂-saturated condition, an irreversible reduction wave of Ni(II)/Ni(I) couple apprears at −1.22 V, which matches with prior studies reporting irreversible behaviors of Ni(II)/Ni(I) redox couple on a glassy carbon electrode in tri-HCl buffer and KCl solution [7,8]. Some other studies, however, reported quasi-reversible behaviors for the Ni(II)/Ni(I) couple [5,6]. It seems that the reversibility of Ni(II)/Ni(I)-cyclam is affected by the experimental conditions (e.g., supporting electrolyte, pH, working electrode materials), which deserves further study. In a CO₂-saturated solution, a large reduction current corresponding to electrocatalytic CO₂ reduction is observed. The onset potential of the CO₂ reduction is determined to be −1.18 V, which is consistent with prior studies [5-8].

The CVs of Cu(II)-cyclam in tris-HCl buffer with and without CO₂ are shown in Fig. 3. The reduction of Cu(II)-cyclam takes place through two steps. A Cu(II)/Cu(I) peak appears at −0.85 V (peak A), a Cu(I)/Cu(0) peak at −1.05 V (peak B) [12], and the solvent decomposition current follows at below −1.2 V. In a CO₂-saturated solution, the reduction currents are notably increased and the peak potentials are shifted to positive direction (i.e., −0.81 V (peak A’) and −0.97 V (peak B’)). The onset potential of CO₂ reduction by Cu(II)- cyclam is determined about −0.7 V, which is much less negative than that by Ni(II)-cyclam (−1.18 V). The theoretical potential for the CO₂/CO conversion is known to be −0.78 V vs. Ag/AgCl in pH 8 tris-HCl buffer solution saturated with 1 atm CO₂ [8]. Despite the slightly different experimental conditions, it can be stated that the electrocatalytic CO₂ reduction by Cu(II)-cyclam takes place near the thermodynamic potential. During the anodic scan, however, a large oxidation peak (peak C or peak C’) appears at around 0.0 V, which is attributed to the oxidation of Cu(I) ion/Cu atom released from Cu(I)-cyclam [12]. This manifests a severe demetalation behavior of Cu(II)-cyclam, which is a critical drawback as an electrocatalyst for CO₂ reduction.

The CVs of Cu(II)-PCB-cyclam and Cu(II)-PCB-TE2A, which were rationally designed to mitigate the demetalation behavior of Cu(II)-cyclam are shown in Fig. 4. It is noted that N-alkylation affects the redox properties of Cu(II)-cyclam in the following three aspects: (1) Cu(II)-PCB-cyclam exhibits only a quasi-reversible redox peak of Cu(II)/Cu(I) couple without showing Cu(I)/Cu redox reaction (Fig. 4a). (2) the redox potential of Cu(II)/Cu(I) couple is much more positive (−0.50 V) than that of un-substituted Cu(II)-cyclam (−0.85 V). (3) Cu(II)-PCB-cyclam shows much smaller (< 50%) Cu oxidation current, indicating mitigated demetalation compared to Cu(II)-cyclam.

These results are consistent with the previous claims that the N-alkylation of cyclam stabilizes the Cu(I)- and Ni(I)-cyclams, leading to a positive shift in the M(II)/M(I) potential [14,15]. It can be viewed that the stabilized Cu(I)-PCB-cyclam also resulted in the absence of M(I)/M redox reaction and in the diminished demetalation. However, the cathodic behavior of Cu(II)-PCB-cyclam is hardly increased under CO₂, implying that the presence of a cross-bridging propylene group seriously deactivates the electrocatalytic ability of Cu(II)-cyclam. The negligible electrocatalytic activity of Cu(II)-PCB-cyclam seems to be due to the too positive reduction potential (−0.50 V) for the theoretical CO₂ reduction potential (−0.78 V). It is supposed that the optimum reduction potential of an electrocatalyst would be only slightly more negative than the CO₂ reduction potential.

Cu(II)-PCB-TE2A also exhibits a quasi-reversible Cu(II)/Cu(I) redox reaction centered at −0.82 V. The reduction potential of Cu(II)/Cu(I) couple is −0.88 V, which is similar to that of un-substituted Cu(II)-cyclam (−0.85 V) but much more negative than Cu(II)-PCB-cyclam (−0.50 V). It seems that diacetic groups in a PCB-cyclam backbone increase the electron density around a Cu(II) ion, and thus make the Cu(II)/Cu(I) reduction of Cu(II)-PCB-TE2A unfavored compared to more Cu(II)-PCB-cyclam. It seems also possible that the presence of diacetic groups returns the cavity size of Cu(II)-PCB-cyclam to that of Cu(II)-cyclam, probably by generating a counter-strain at the other side of cyclam plane of a cross-bridged propylene group.

Notably, the Cu oxidation current is greatly suppressed in Cu(II)-PCB-TE2A. The excellent resistance of Cu(II)-PCB-TE2A to reduction-induced demetalation is ascribed to thorough protection of Cu(II)-cyclam backbone by the two Cu-coordinating acetate groups as well as by a cross-bridging propylene moiety [17]. Cu(II)-PCB-TE2A, however, shows only marginal electrocatalytic activity for CO₂ reduction. This poor electrocatalytic activity of Cu(II)-PCB-TE2A is rather unexpected when considering the redox potential (−0.82 V) suitable for the CO₂ reduction (−0.78 V) and the enhanced stability of Cu(I) state of the complex. Although further investigation is needed, it is supposed that the steric hindrance around the central Cu ion in Cu(II)-PCB-TE2A is too large to allow the intimate interaction with CO₂ molecules.

Based on the results obtained so far, it can be inferred that, to design efficient electrocatalysts for CO₂ reduction, one should consider the stereochemical factor to allow facile interaction with CO₂ molecules in addition to the proper redox potential and chemical stability of the catalysts. The lessons learned from this study would be helpful to understand the detailed CO₂ reduction mechanism of homogeneous electrocatalyst and to develop novel catalysts surpassing Ni(II)-cyclam, the current most efficient electrocatalyst for CO₂ reduction.

Cu(II)-cyclam, Cu(II)-PCB-cyclam, and Cu(II)-PCB-TE2A were compared as homogeneous electrocatalysts for CO₂ reduction in tris-HCl buffer (pH 8.45) on a glassy carbon electrode. It was revealed that Cu(II)-cyclam catalyzes CO₂ reduction at much lower overpotential than Ni(II)-cyclam, the current most representative electrocatalyst. However, Cu(II)-cyclam showed severe demetalation behavior owing to the insufficient stability of Cu(I) state. The demetalation issue was greatly mitigated in Cu(II)-PCB-cyclam, in which a cross-bridging propylene is supposed to stabilize Cu(I)-PCB-cyclam. Cu(II)-PCB-cyclam, however, exhibited poor electrocatalytic activity, which is possible ascribed to its redox potential too high for CO₂ reduction. Cu(II)-PCB-TE2A exhibited further suppressed demetalation and appropriate redox potential for CO₂ reduction, but was turned out to show only marginal electrocatalytic activity, which is attributed to large steric hindrance preventing facile contact with CO₂ molecules. This study also suggests that stereochemical factor is as crucial as the redox potential and chemical stability of the electrocatalysts for CO₂ reduction.