Methane Formation Induced via Face-to-Face Orientation of Cyclic Fe Porphyrin Dimer in Photocatalytic CO₂ Reduction

Abstract

Iron porphyrins are known to provide CH₄ as an eight-electron reduction product of CO₂ in a photochemical reaction. However, there are still some aspects of the reaction mechanism that remain unclear. In this study, we synthesized iron porphyrin dimers and carried out the photochemical CO₂ reduction reactions in N,N-dimethylacetamide (DMA) containing a photosensitizer in the presence of 1,3-dimethyl-2-phenyl-2,3-dihydro-1H-benzo[d]imidazole (BIH) as an electron donor. We found that, despite a low catalytic turnover number, CH₄ was produced only when these porphyrins were facing each other. The close proximity of the cyclic dimers, distinguishing them from a linear Fe porphyrin dimer and monomers, induced multi-electron CO₂ reduction, emphasizing the unique role of their structural arrangement in CH₄ formation.

1. Introduction

The rise in atmospheric CO₂ concentration has led to serious impacts on the environment, emphasizing the urgent requirement for CO₂ reduction. Many attempts have been made to utilize light energy to reduce CO₂ and convert it into energy-rich substances similar to photosynthesis. One significant challenge is the selective reduction of CO₂ while suppressing thermodynamically favorable proton reduction, which generates hydrogen, and this has been addressed using metal complexes. For instance, in artificial Z-scheme systems using semiconductor photocatalysts, the incorporation of metal complex catalysts at reduction sites enables selective CO₂ reduction under visible light irradiation. This process effectively extracts electrons from water, and simultaneously minimizes hydrogen generation, producing CO and formic acid [1,2]. Abundant earth elements have been used as the central metal; among them, Fe porphyrin complexes have been extensively studied owing to their high selectivity and activity [3,4]. The introduction of peripheral proton-donating groups (Fe-o-OH in Figure 1) considerably improved the catalytic activity of Fe porphyrins, resulting in selective CO production [5]. In addition, linking Fe porphyrins through an o-phenylene or a urea group enhanced CO production [6,7]. The Fe porphyrin substituted with p-trimethylammoniophenyl groups (Fe-p-TMA in Figure 1) also exhibited high CO production activity in the electrochemical CO₂ reduction [8]. Interestingly, under photochemical CO₂ reduction conditions using tris(2-phenylpyridine)iridium (Ir(ppy)₃) as a photosensitizer and triethylamine (TEA) as a sacrificial electron donor in acetonitrile, Fe-o-OH, and Fe-p-TMA yielded CH₄ along with CO [9]. The formation of CH₄ was observed when using an organic photosensitizer, 3,7-(4-biphenyl)-1-naphthalene-10-phenoxazine (Phen2), instead of Ir(ppy)₃ in N,N-dimethylformamide (DMF) solution containing Fe-p-TMA and sacrificial electron donors [10]. The reaction mechanism indicated the involvement of the Fe(II)–CO intermediate. However, although computational studies have proposed this reaction mechanism [11], the actual formation mechanism of CH₄, an eight-electron reduction product of CO₂, remains unclear.

We previously synthesized cyclic Zn porphyrin dimers connected via 2,2′-bipyridine (bpy) and isophthalamide (Zn₂-CP2_m) or terephthalamide linkers (Zn₂-CP2_p) [12]. Herein, we substituted the central Zn ions of porphyrins with Fe ions to yield two types of cyclic Fe porphyrin dimers with different distances between the porphyrin planes (Fe₂-CP2_m and Fe₂-CP2_p in Figure 1). The photocatalytic CO₂ reduction, in the presence of Ir(ppy)₃ [9] or Phen2 [10] as the photosensitizer and 1,3-dimethyl-2-phenyl-2,3-dihydro-1H-benzo[d]imidazole (BIH) [13] as the electron donor, surprisingly produced CH₄ only when using the cyclic Fe porphyrin dimers (Fe₂-CP2_m and Fe₂-CP2_p). In contrast, a linear Fe porphyrin dimer (Fe₂-P2) and monomeric Fe porphyrins did not produce CH₄, indicating that the close proximity of two Fe porphyrins facing each other induced the multielectron reduction of CO₂.

2. Results and Discussion

Fe₂-CP2_m and Fe₂-CP2_p were, respectively, prepared by demetallation of Zn₂-CP2_m and Zn₂-CP2_p [12] to obtain the corresponding free-base porphyrins, followed by the introduction of Fe ions. Fe₂-P2 was prepared by the introduction of Fe ions into the corresponding free-base precursor [12]. In the UV–vis absorption spectra, the Soret band of Fe₂-P2 red shifted compared with those of Fe₂-CP2_m, Fe₂-CP2_p, and Fe(III) tetraphenylporphyrin chloride (FeTPP(Cl)) (Figure 2, left). The red-shifted band is attributed to the head-to-tail excitonic coupling between the two transition dipoles of X in the anticonformation of Fe₂-P2, indicating that a linear structure is stable in Fe₂-P2, as observed for the corresponding Zn porphyrins (Figure 1) [12]. The cyclic voltammogram (CV) and differential pulse voltammogram (DPV) of Fe₂-P2 in Ar-saturated DMF showed three reversible redox waves at −0.74, −1.55, and −2.23 V vs. Fc/Fc⁺ (−0.27, −1.08, and −1.76 V vs. SCE) [14], which corresponded, respectively, to Fe(III/II), Fe(II/I), and Fe(I/0) couples (Figure 2, right and Figure S12). The observation of only three redox waves indicates that each porphyrin is reduced independently and that the electronic interaction between the two porphyrins through the bpy linker is negligible. Meanwhile, although the CVs of Fe₂-CP2_m and Fe₂-CP2_p showed redox waves at the similar positions to those of Fe₂-P2, the second redox wave of Fe₂-CP2_m split into two, indicating that the porphyrins in close proximity, arranged in a face-to-face configuration, exhibit electrical interaction with each other. Under a CO₂ atmosphere, Fe₂-P2, Fe₂-CP2_m, and Fe₂-CP2_p showed similar catalytic currents at the third waves of Fe(I/0) of the Fe porphyrin dimers (Figure 2, right). Lewis acids, known to enhance the activity of Fe porphyrin catalysts for CO₂ reduction [15,16], were anticipated to exhibit differences in their interactions with the bpy moiety between cyclic and linear structures. However, no significant difference among three porphyrin dimers in the catalytic currents was observed for the electrochemical CO₂ reduction even when using water and/or metal ion additives (Figures S13–S15 and Table S1).

We first attempted to perform photocatalytic CO₂ reduction using Fe₂-P2 in the presence of Ir(ppy)₃ as a photosensitizer and TEA as an electron donor, but no reduction product was detected. The present Fe porphyrin dimers exhibited more negative reduction potentials (−2.17 to −2.23 V vs. Fc/Fc⁺ and −1.70 to −1.76 V vs. SCE for Fe(I/0)) compared with the reported Fe porphyrins, such as Fe-o-OH (−1.57 V vs. SCE for Fe(I/0)) and Fe-p-TMA (−1.47 V vs. SCE for Fe(I/0)) [17]. According to the energy diagram (Figure S16), the oxidation quenching process [18] of the excited Ir(ppy)₃ by either Fe₂-P2 or Fe₂-CP2_m is thermodynamically less favorable when using TEA. Consequently, we used BIH with a stronger reducing power as an electron donor instead of TEA. The energy diagram and results of a phosphorescence quenching experiment support that electron transfer from BIH to the excited Ir(ppy)₃ (i.e., reductive quenching) [18], followed by a reduction in the Fe(I) porphyrins, can occur (Figures S17–S19). Here, BIH was used in much larger quantities (10 mM) than the catalyst (10 μM), and it was expected that the reaction could proceed, although it was slightly unfavorable thermodynamically. Under the reaction conditions for the photocatalytic CO₂ reduction using BIH and Fe-o-OH, we observed the production of CO during the catalytic reaction (Table S2). However, the amount of BIH consumed was significantly greater than the production of CO (Figure S21). The NMR spectra after irradiation in the presence of CO₂ revealed the formation of unidentified BIH decomposition products (Figure S22) rather than the formation of BI⁺, which is typically observed in reactions with the [Ru(bpy)₃]²⁺ photosensitizer as a two-electron oxidation product of BIH [13]. A highly reactive BI radical, formed via the oxidation and deprotonation of BIH, is likely to react with CO₂ to generate the unidentified products (Figure S23). This is hypothesized because Ir(ppy)₃ would not accept the electron from the BI radical. The investigation of the effects of solvents and additives showed that N,N-dimethylacetamide (DMA) [19] promoted CO production and suppressed BIH consumption more than acetonitrile (entries 4−6 in Table S3 and Figure S24). Therefore, in the subsequent experiments, the photoreactions were performed using DMA.

Photocatalytic CO₂ reductions using Fe₂-P2 and Fe₂-CP2_m (10 μM) in DMA containing BIH (10 mM) and Ir(ppy)₃ (0.2 mM) under 450 nm light were performed. The turnover numbers (TONs) of the reduction products against the Fe porphyrin dimers are shown in Figure 3. No detectable amounts of CO, H₂, or CH₄ were found in the absence of any one of the Fe porphyrin dimers, Ir(ppy)₃, and light, whereas only H₂ was detected under Ar instead of CO₂. While Fe₂-P2 and Fe₂-CP2_m produced CO, the amount of CO was smaller in Fe₂-CP2_m. Interestingly, a small amount of CH₄ was formed in Fe₂-CP2_m (Figure 3b). Under the same conditions, CH₄ was not detected in Fe₂-P2 (Figure 3a), Fe-o-OH, or FeP-phen, which is a model monomer with a diimine ligand. Figure 3b shows CH₄ production with an induction period and lower CO production than that in Figure 3a, indicating that the CH₄ was formed via the reduction in CO, as reported previously [9,10].

Next, we investigated the photocatalytic CO₂ reduction reactions in which the organic dye Phen2, rather than Ir(ppy)₃, was used as the photosensitizer (Figure 4) [10]. During the irradiation of the solution in the absence of the catalyst (Figure S28), BIH decomposition was still observed without reduction products of CO₂. However, reducing the light intensity to 5 mW suppressed BIH decomposition. When Fe₂-P2 was used as the catalyst, a linear CO formation was observed for up to 4 h under the 5 mW light intensity (Figure S29). Figure 4 shows the TONs of the reduction products against the Fe porphyrin dimers in DMA containing BIH (10 mM) and Phen2 (1.0 mM) under 420 nm light (5 mW). As observed with Ir(ppy)₃, CH₄ was formed with induction periods in the cyclic structure (Fe₂-CP2_m and Fe₂-CP2_p), whereas no CH₄ was detected in the linear structure (Fe₂-P2). In addition, the CO productions in Figure 4b,c were smaller than those in Figure 4a. In the previous systems involving Fe-p-TMA as the catalyst and TEA as the electron donor, the addition of a proton source, such as trifluoroethanol (TFE), enhanced the formation of CO and CH₄ [9,10]. However, using the cyclic porphyrin dimer, the CH₄ production decreased with the addition of TFE and was completely suppressed by PhOH, while TFE and PhOH enhanced the CO production. The addition of Mg ions decreased the production of CO and CH₄ (Figure S30).

In conventional catalytic reactions using Phen2 as the photosensitizer, the reaction typically involves an oxidative quenching process, where electrons are transferred from the excited Phen2 to the catalyst [18,20]. However, this system would proceed via a reductive quenching process involving electron transfer from the electron donor to the excited Phen2. The fluorescence quenching experiments of Phen2 by BIH demonstrate that the electron transfer from BIH to the excited singlet state of Phen was efficient (Figure S32). However, the quenching efficiency of the excited Phen2 (η_q), which was estimated from the Stern–Volmer plot [18], strongly depended on the concentration of BIH because of the shorter fluorescence lifetime ([BIH] = 10 mM, η_q = 7%; [BIH] = 100 mM, and η_q = 41%). Meanwhile, we observed that the TONs were less dependent on the BIH concentration (10−100 mM in Figure 5), suggesting that the electron transfer from BIH mainly occurred not via the excited singlet state but via the long-lived excited triplet state of Phen2 [21], which has a lifetime of 480 μs [22].

The photocatalytic CO₂ reduction using Fe₂-CP2_p and 100 mM BIH produced CH₄ with TON = 3.5 against the Fe porphyrin dimer after 35 h, exceeding the amount of the Fe porphyrin units (Figure S34). The capillary electrophoresis showed that the TON of formic acid reached 38 after irradiation for 18 h in the presence of Fe₂-CP2_p and BIH (100 mM, Figure S34). We conducted isotopic experiments under ¹²CO₂ and ¹³CO₂ atmospheres. In gas chromatography/mass spectrometry, ¹³CO (m/z = 29) and ¹³CH₄ (m/z = 17) were detected under a ¹³CO₂ atmosphere (Figures S35 and S36), confirming that the carbon source of CO and CH₄ was CO₂. The ¹H and ¹³C NMR spectra were measured in a DMA-d₉ solution during irradiation under ¹²CO₂ and ¹³CO₂ atmospheres (Figures S37–S39). The spectral changes during light irradiation showed that BIH was almost completely consumed after 21 h, indicating that the catalytic reaction stopped due to the disappearance of BIH. No reduction products, including methanol or formaldehyde, were observed, except for formic acid, which showed a doublet peak at 8.68 ppm with a coupling constant of J_13C-H = 175 Hz and a singlet peak at 8.72 ppm in ¹H NMR spectra under ¹²CO₂ and ¹³CO₂ atmospheres, respectively (Figure S38). A peak at 167 ppm was assigned to HC(O)-, which was correlated with the doublet proton peak at 8.68 ppm in the heteronuclear multiple bond connectivity (HMBC), observed in the ¹³C NMR spectrum under only a ¹³CO₂ atmosphere (Figures S39 and S40) [19]. Formic acid was not observed in the absence of the Fe porphyrins, indicating that formic acid is produced via the CO₂ reduction catalyzed by the Fe porphyrins, and it is not directly formed by the chemical reaction between BIH and CO₂ [23,24]. Furthermore, a peak appeared at 222 ppm in the ¹³C NMR spectrum only under a ¹³CO₂ atmosphere (indicated by an asterisk in Figure S39). The peak can be attributed to the Fe–¹³CO signal [25], indicating the formation of a carbonyl intermediate during irradiation, as observed in previous reports. In addition, only under a ¹³CO₂ atmosphere an intense peak was observed at 172 ppm (Figure S39), which was correlated with the proton peaks at 7.4, 3.4, 2.7, 2.4, and 1.8 ppm in HMBC (Figure S40). Although no clear attribution could be established, it likely corresponded to the reaction products of BIH and CO₂, an adduct of CO₂ with the BI radical caused by the oxidation and deprotonation of BIH.

3. Materials and Methods

3.1. General Procedure

All chemicals and solvents were of commercial reagent quality and were used without further purification unless otherwise stated. Tris(2-phenylpyridine)iridium (Ir(ppy)₃) was purchased from Sigma-Aldrich (St. Louis, MO, USA). Fe(III) 5,10,15,20-tetrakis(2,6-dihydroxyphenyl)porphyrin chloride (Fe-o-OH) [5], 3,7-(4-biphenyl)-1-naphthalene-10-phenoxazine (Phen2) [10], the cyclic Zn porphyrin dimers (Zn₂-CP2_m and Zn₂-CP2_p), the linear free-base porphyrin dimer (Fb₂-P2) [12], 1,3-dimethyl-2-phenyl-2,3-dihydro-1H-benzo[d]imidazole (BIH) [13], and 5,10,15-tris(4-tert-butylphenyl)-20-(1,10-phenanthrolin-2-yl)-21H,23H-porphyrin (H₂P-phen) [26] were prepared according to the literature. N,N-dimethylformamide (DMF) was dried over molecular sieves of size 4 Å. The reactions were monitored on silica gel 60F₂₅₄ TLC plates (Merck, Darmstadt, Germany). The following silica gels utilized for the column chromatography were purchased from Kanto Chemical Co. Inc. (Tokyo, Japan): silica gels (Spherical, Neutral) 40–100 μm and (Flash) 40–50 μm. ¹H, ¹³C NMR, and ¹H-¹H correlation spectroscopy (COSY), ¹H−¹³C heteronuclear single quantum correlation (HSQC), and ¹H−¹³C heteronuclear multiple bond correlation (HMBC) spectra were recorded using a JEOL JNM-ECZ-400 and a JEOL JNM-ECA-500 (JEOL, Tokyo, Japan). Chemical shifts were recorded in parts per million (ppm) relative to tetramethylsilane. MALDI–TOF mass spectra were collected on a JEOL JMS S-3000 with dithranol as a matrix with sodium iodide (NaI). UV-vis absorption spectra were collected using a square cell (path length = 1.0 cm) on a JASCO V-650 spectrometer (JASCO, Tokyo, Japan). The steady-state emission spectra were collected on an Hitachi F-4500 spectrometer and corrected for the response of the detector system. The fluorescence intensities were normalized at the absorbance of the excitation wavelength. Cyclic voltammogram (CV) and differential pulse voltammogram (DPV) were measured using an ALS-H/CHI Model 612E electrochemical analyzer (BAS, Tokyo, Japan) in a micro-cell equipped with a glassy carbon working electrode (ϕ 1.6 mm) and a Pt counter electrode. The micro-cell was connected via a Luggin capillary with a reference electrode of Ag/AgNO₃ (10 mM in DMA). Tetrabutylammonium hexafluorophosphate (ⁿBu₄NPF₆) recrystallized from ethyl acetate was used as a supporting electrolyte. Ferrocene was used as an external standard, and all potentials were referenced to the ferrocene/ferrocenium couple. The currents (i) were normalized by the peak currents (i_p⁰) corresponding to the Fe(III)/Fe(II) wave in the absence of CO₂ and additives. High-performance liquid chromatographies (HPLCs) were carried out using a JASCO PU-2089 and an MD-44010 system (JASCO, Tokyo, Japan) equipped with a TSKgel ODS-100S column (4.6 mm I.D. × 25 cm; Tosoh, Tokyo, Japan) using acetonitrile/H₂O = 4/1 (v/v) as an eluent.

3.2. Synthesis of Fb₂-CP2_m

The TFA (0.50 mL) was slowly added to Zn₂-CP2_m (10 mg, 5.9 × 10⁻⁶ mol) in a 10 mL flask, and the mixture was stirred for 2 h. The solution was slowly poured into a saturated NaHCO₃ aqueous solution in an ice bath. The organic layer was transferred to a PFA-coated funnel, and the aqueous layer was extracted with CHCl₃ (10 mL × 3). The combined organic layer was washed with water and dried over anhydrous Na₂SO₄. The solvent was evaporated to dryness, resulting in a purple solid, Fb₂-CP2_m (8.3 mg, 90%). ¹H NMR (500 MHz, pyridine-d₅) δ/ppm = 10.85 (s, 2H, NH), 8.91 (s, 1H, Ph), 8.83 (m, 2H, Ph), 8.74 (m, 4H, β-pyrrole), 8.70 (m, 2H, Py), 8.58–8.47 (m, 12H, β-pyrrole, Ph, Py), 8.36 (m, 4H, β-pyrrole), 8.31 (m, 2H, Py), 8.11 (m, 2H, Ph), 7.74 (s, 2H, Ph), 7.67 (m, 2H, Ph), 7.55 (m, 1H, Ph), 7.27 (s, 4H, mesityl), 7.08 (s, 4H, mesityl), 2.48 (s, 12H, CH₃), 1.64 (s, 12H, CH₃), 1.54 (s, 12H, CH₃), −2.84 (s, 4H, inner NH); MALDI-TOF mass: m/z [M + H]⁺ 1557.6808, calcd. for [C₁₀₆H₈₅N₁₂O₂]⁺ 1557.6913.

3.3. Synthesis of Fe₂-CP2_m

In a 10 mL flask were placed Fb₂-CP2_m (8.3 mg, 5.3 × 10⁻⁶ mol), CHCl₃ (2.0 mL), and 2,6-lutidine (15 μL, 1.3 × 10⁻⁴ mol). Anhydrous FeCl₂ (37 mg, 2.9 × 10⁻⁴ mol) dissolved in methanol (1.0 mL) was added to it, and the mixture was stirred under reflux for 18 h. The resulting solution was transferred to a perfluoroalkoxy alkane (PFA)-coated flask, and diluted with CHCl₃ (ca. 10 mL). The organic solution was washed with 1 M HCl aqueous solution (×3) and water (×3) and then dried over anhydrous Na₂SO₄. The residue obtained by evaporation of the solvent was purified with a flush silica gel column (eluents: CHCl₃, CHCl₃:CH₃OH = 50:1, and CHCl₃:CH₃OH = 10:1). The fraction eluted with CHCl₃:CH₃OH = 50:1 was collected and the solvent was evaporated. The solid was dissolved in CHCl₃ (ca. 10 mL) was treated with 1 M HCl aqueous solution (×3) and water (×3), and passed through Phase Separator paper (Whatman, Maidstone, UK). The solvent was evaporated to dryness, giving the titled compound as a black solid (7.9 mg, 86%). MALDI-TOF mass: m/z [M − 2Cl + H]⁺ 1666.5287 (max), calcd. for [C₁₀₆H₈₁N₁₂O₂Fe₂]⁺ 1666.5333; UV-vis absorption (CHCl₃) λ _max/nm (ε/M⁻¹cm⁻¹) = 373 (6.2 × 10⁴, LMCT band), 415 (1.1 × 10⁵, Soret band), and 509 (1.4 × 10⁴, Q band).

3.4. Synthesis of Fb₂-CP2_p

The TFA (1.0 mL) was slowly added to Zn₂-CP2_p (8.7 mg, 5.2 × 10⁻⁶ mol) in a 10 mL flask, and the mixture was stirred for 2 h. The solution was slowly poured into a saturated NaHCO₃ aqueous solution in an ice bath. The organic layer was transferred to a PFA-coated funnel and the aqueous layer was extracted with CHCl₃ (10 mL × 3). The combined organic layer was washed with water and dried over anhydrous Na₂SO₄. The solvent was evaporated to dryness, giving a purple solid, Fb₂-CP2_p (7.1 mg, 89%). ¹H NMR (400 MHz, CDCl₃) δ/ppm = 8.67 (d, J = 8.2 Hz, 2H, Ph or Py), 8.63 (d, J = 7.8 Hz, 2H, Ph or Py), 8.52 (d, J = 4.7 Hz, 4H, β-pyrrole), 8.47 (t, J = 7.8 Hz, 2H, Ph or Py), 8.43–8.45 (2H, Ph or Py), 8.42 (d, J = 4.7 Hz, 4H, β-pyrrole), 8.39 (d, J = 4.7 Hz, 4H, β-pyrrole), 8.35 (d, J = 4.7 Hz, 4H, β-pyrrole), 8.11 (d, J = 7.3 Hz, 2H, Ph or Py), 7.70 (t, J = 7.9 Hz, 2H, Ph or Py), 7.59 (s, 4H, Ph), 7.19 (s, 4H, mesityl), 7.13 (s, 4H, mesityl), 7.00 (brs, 2H, Ph or Py), 2.59 (s, 12H, CH₃), 1.48 (s, 12H, CH₃), 1.36 (s, 12H, CH₃), −3.44 (s, 4H, inner NH); MALDI-TOF mass: m/z [M + H]⁺ 1557.6928, calcd. for [C₁₀₆H₈₅N₁₂O₂]⁺ 1557.6913.

3.5. Synthesis of Fe₂-CP2_p

In a 10 mL flask were placed Fb₂-CP2_p (7.1 mg, 4.6 × 10⁻⁶ mol), CHCl₃ (2.0 mL), and 2,6-lutidine (14 μL, 1.2 × 10⁻⁴ mol). Anhydrous FeCl₂ (37 mg, 2.9 × 10⁻⁴ mol) dissolved in methanol (1.0 mL) was added to it, and the mixture was stirred under reflux for 18 h. The resulting solution was transferred to a perfluoroalkoxy alkanes (PFA)-coated flask, and diluted with CHCl₃ (ca. 10 mL). The organic solution was washed with 1 M HCl aqueous solution (×3) and water (×3), and then dried over anhydrous Na₂SO₄. The residue obtained by evaporation of the solvent was purified with a flush silica gel column (eluents: CHCl₃, CHCl₃:CH₃OH = 50:1, and CHCl₃:CH₃OH = 10:1). The fraction eluted with CHCl₃:CH₃OH = 50:1 was collected and the solvent was evaporated. The solid was dissolved in CHCl₃ (ca. 10 mL) was treated with 1 M HCl aqueous solution (×3) and water (×3) and passed through Phase Separator paper (Whatman). The solvent was evaporated to dryness, resulting in the titled compound as a black solid (6.3 mg, 79%). MALDI-TOF mass: m/z [M−2Cl + H]⁺ 1666.5417 (max), calcd. for [C₁₀₆H₈₁N₁₂O₂Fe₂]⁺ 1666.5333; UV-vis absorption (CHCl₃) λ_max/nm (ε/M⁻¹cm⁻¹) = 375 (7.3 × 10⁴, LMCT band), 416 (1.7 × 10⁵, Soret band), 510 (1.4 × 10⁴, Q band).

3.6. Synthesis of Fe₂-P2

In a 10 mL flask were placed 5Fb₂ (26 mg, 1.6×10⁻⁵ mol), CHCl₃ (5.0 mL), and 2,6-lutidine (30 μL, 2.5 × 10⁻⁴ mol). Anhydrous FeCl₂ (100 mg, 7.9 × 10⁻⁴ mol) dissolved in methanol (3.0 mL) was added to it, and the mixture was stirred under reflux for 24 h. The reaction was quenched by adding EDTA aqueous solution, and the resulting solution was transferred to a perfluoroalkoxy alkanes (PFA)-coated flask, and diluted with CHCl₃ (ca. 50 mL). The organic solution was washed with water (×3) and brine (×1), and then dried over anhydrous Na₂SO₄. The use of hydrochloric acid was avoided because the Boc group could be removed under acidic conditions. The residue obtained by evaporation of the solvent was purified with a flush silica gel column (eluents: CHCl₃:CH₃OH = 50:1 and CHCl₃:CH₃OH = 10:1). The fraction eluted with CHCl₃:CH₃OH = 10:1 was collected, and was stirred with brine for overnight. The organic layer was passed through Phase Separator paper (Whatman) and the solvent was evaporated to dryness, giving the titled compound as a black solid (24 mg, 85%). MALDI-TOF mass: m/z [M − 2Boc − 2Cl + H]⁺ 1534.5110 (max), calcd. for [C₉₈H₇₆Fe₂N₁₂]⁺ 1534.5121; UV-vis absorption (CHCl₃) λ_max/nm (ε/M⁻¹cm⁻¹) = 378 (8.4 × 10⁴, LMCT band), 421 (1.8 × 10⁵, Soret band), 509 (2.4 × 10⁴, Q band).

3.7. Synthesis of FeP-phen

In a 25 mL flask were placed H₂P-phen (30 mg, 3.4 × 10⁻⁵ mol), CHCl₃ (8.0 mL), and 2,6-lutidine (35 μL, 3.0 × 10⁻⁴ mol). Anhydrous FeCl₂ (107 mg, 8.5 × 10⁻⁴ mol) dissolved in methanol (3.0 mL) was added to it, and the mixture was stirred under reflux for 22 h. The reaction was quenched by adding EDTA (ethylenediaminetetraacetic acid) aqueous solution, and the resulting solution was transferred to a perfluoroalkoxy alkanes (PFA)-coated flask, and diluted with CHCl₃ (ca. 50 mL). The organic solution was washed with was washed with 1 M HCl (×3) and brine and then dried over anhydrous Na₂SO₄. The residue obtained by evaporation of the solvent was purified with a flush silica gel column (eluents: CHCl₃, CHCl₃:CH₃OH = 9:1). The fraction eluted with CHCl₃:CH₃OH = 9:1 was collected and the solvent was evaporated. The solid was dissolved in CHCl₃ (ca. 10 mL) and treated with 1 M HCl aqueous solution (×3) and brine, and passed through Phase Separator paper (Whatman). The solvent was evaporated to dryness, giving the titled compound as a black solid (22 mg, 69%). MALDI-TOF mass: m/z [M − Cl + H]⁺ 939.3940 (max), calcd. for [C₆₂H₅₅FeN₆]⁺ 939.3833.

3.8. Photocatalytic CO₂ Reduction

In glass tubes (8.0 mL, i.d. = 10 mm), 1.0 mL of CO₂-saturated DMA solutions containing BIH was added by 1.0 mL of Ar-saturated DMA solutions containing the Fe porphyrin dimer and the photosensitizer, and the reaction solutions were bubbled with CO₂ gas (purity ≥ 99.995%) for 15 min. Photo-irradiations were carried out using a merry-go-round irradiation apparatus (Iris-MG, Cell Systems, Yokohama, Japan) equipped with LED lamps at λ = 420 nm (FWHM = 18.4 nm). The gaseous reaction products (CO, H₂, and CH₄) were quantified with a gas chromatography system (GC-2014, Shimadzu Science, Kyoto, Japan) equipped with a Shincarbon column (i.d. 3.0 mm × 3.0 m) and a thermal conductivity detector (TCD). The product (formate) in the solutions was analyzed with a capillary electrophoresis system (Otuka Electronics Co. CAPI-3300I, Osaka, Japan).

3.9. 13CO₂-Labeling Experiment

In glass tubes (8.0 mL, i.d. = 10 mm), 2.0 mL of Ar-saturated DMA solutions containing BIH (0.10 M) was added by 2.0 mL of Ar-saturated DMA solutions containing Zn₂-CP2_p (10 μM) and Phen2 (1.0 mM), and the reaction solutions were bubbled with ¹³CO₂ gas, which was generated by the addition of 3.0 M sulfuric acid (5 mL) to Ba¹³CO₃ powder (2.5 g). After irradiation at 420 nm (5 mW) for 16 h with the merry-go-round irradiation apparatus, the gaseous reaction products were analyzed with a GC-MS (GCMS-QP2010 Plus, Shimadzu Science; RESTEK (Bellefonte, PA, USA); RT-Msieve 5A). In an NMR tube, a 0.5 mL DMA-d₉ solution containing Zn₂-CP2_p (0.10 mM), Phen2 (1.0 mM) and BIH (0.10 M) was bubbled with ¹³CO₂ gas, which was generated by addition of 3.0 M sulfuric acid (3 mL) to Ba¹³CO₃ powder (1.0 g). After irradiation at 420 nm (5 mW) for 21 h with the merry-go-round irradiation apparatus, the ¹H and ¹³C NMR spectra were measured.

3.10. Computational Methods

The DFT calculations were carried out using the Gaussian 09 package of programs [27]. Each structure was fully optimized using the B3LYP functional using the 6–31G(d) basis set for all atoms, except Fe, and the standard double-ζ type LANL2DZ basis set with the effective core potential of Hay−Wadt for Fe. The stationary points were verified using the vibrational analysis.

4. Conclusions

In this study, we show that the photochemical CO₂ reduction yielded CH₄, an eight-electron reduction product of CO₂, when Fe porphyrins were placed in a face-to-face arrangement. Although the catalytic turnover number of CH₄ was low in this study, this can be attributed to the significant degradation of BIH. Thus, it is expected that by addressing this issue, we can enhance the catalytic performance. While bimetallic porphyrin complexes have been reported to enhance CO production, to the best of our knowledge, this is the first report on the induction of CH₄ production using bimetallic porphyrins. We anticipate that this finding can contribute to the understanding CH₄ formation mechanisms and provide potential molecular design guidelines for selective CH₄ generation.