Visible-Light-Driven Conversion of CO₂ to CH₄ with an Organic Sensitizer and an Iron Porphyrin Catalyst

Abstract

Using a phenoxazine-based organic photosensitizer and an iron porphyrin molecular catalyst, we demonstrated photochemical reduction of CO₂ to CO and CH₄ with turnover numbers (TONs) of 149 and 29, respectively, under visible-light irradiation (λ > 435 nm) with a tertiary amine as sacrificial electron donor. This work is the first example of a molecular system using an earth-abundant metal catalyst and an organic dye to effect complete 8e⁻/8H⁺ reduction of CO₂ to CH₄, as opposed to typical 2e⁻/2H⁺ products of CO or formic acid. The catalytic system continuously produced methane even after prolonged irradiation up to 4 days. Using CO as the feedstock, the same reactive system was able to produce CH₄ with 85% selectivity, 80 TON and a quantum yield of 0.47%. The redox properties of the organic photosensitizer and acidity of the proton source were shown to play a key role in driving the 8e⁻/8H⁺ processes.

In the quest of solar fuels production from CO₂, the ability to effect multielectron and -proton transfer processes with good selectivity remains a daunting challenge.^1–6 Molecular catalysts offer good reactivity and such systems have been shown predominantly to produce the 2e⁻/2H⁺ reduction products of CO (carbon monoxide) or HCOOH (formic acid).^7–9 In these 2e⁻/2H⁺ reduction systems, both electrochemical and photochemical approaches have been successfully developed using earth-abundant metal catalysts and sometimes metal free sensitizers in the case of photostimulated reactions.¹⁰ Recently, some of us were able to achieve 8e⁻/8H⁺ reduction of CO₂ to CH₄ (methane) by employing a dual catalytic approach that combined an iron porphyrin catalyst (Fe-p-TMA, Scheme 1) with an iridium-based photosensitizer fac-Ir(ppy)₃ (fac-tris[2-phenylpyridinato-C²,N]iridium(III)).¹¹ To date, it was the only molecular based system that could achieve such a complete 8e⁻/8H⁺ reduction process. Comparatively, heterogeneous systems have shown more advanced progress for CO₂-to-CH₄ reduction. Semiconductive solid materials doped with cocatalysts have been recently developed for CH₄ production at ambient temperature and pressure, with hydrocarbon production rate up to a few hundred μmol g_cat⁻¹ h⁻¹, ^12–18 and sometimes with excellent selectivity.¹⁹ In the latter example, TiO₂ was used as a photocatalyst with Pd₇Cu₁ alloy as a nano cocatalyst to yield CH₄ with a high selectivity of 96% at a rate of 19.6 μmol g_cat⁻¹ h⁻¹. Like their molecular counterpart, these heterogeneous systems principally employ noble metals to boost reactivity.

Scheme 1.

Molecular Structures of the Iron Catalyst (Fe-p-TMA) and Organic Photosensitizers (Phen1, Phe2) Investigated in This Study

Noble-metal-based photosensitizers, exemplified by polypyridyl ruthenium(II) and iridium(III) compounds,^20,21 are commonly used in photoredox catalysis²² and other important energy conversion processes. Despite their proven performance in numerous light-driven reactions, these noble-metal compounds pose long-term supply and cost issues that their replacement by organic photosensitizers is of significant interest.²³ In this context, a few of us have recently developed organic photosensitizers based on the dihydrophenazine (Phen1,²⁴ 5,10-di(2-naphthyl)−5,10-dihydrophenazine) and phenoxazine^25,26 (Phen2, 3,7-di(4-biphenyl)-1-naphthalene-10-phenoxazine) motifs (Scheme 1). These organic chromophores exhibit photon absorption in the visible light spectrum, redox reversibility, good triplet quantum yields [e.g., 2% (Phen1) and 90% (Phen2)], and long triplet lifetimes [ca. 4.3 μs (Phen1) and 480 μs (Phen2)].²⁷ In particular, Phen1 and Phen2 were specifically engineered as strong excited-state electron donors with highly negative excited state reduction potentials for oxidative quenching applications. The triplet excited state reduction potential values of Phen1 and Phen2 are E⁰(²Phen1^•+/³Phen1*) = −2.09 V vs Fc⁺/Fc and E⁰(²Phen2^•+/³Phen2*) = −2.20 V vs Fc⁺/Fc, respectively (values in N,N-dimethylacetamide as solvent); notably, these values closely match the E⁰(Ir(IV)/³Ir(III)*) = −2.13 V vs Fc⁺/Fc for Ir(ppy)₃, which were successfully employed for the photochemical reduction of CO₂ to CH₄. Given these properties, we hypothesized that Phen1 or Phen2 could directly replace Ir(ppy)₃ in the light-driven tandem catalysis with Fe-p-TMA for CO₂ reduction. Herein, we report a noble metal free molecular system for visible light-driven 8e⁻/8H⁺ reduction of CO₂ to CH₄ with an organic photosensitizer (Phen2) and an earth-abundant iron porphyrin catalyst (Fe-p-TMA). Such a premiere is expected to advance the field of photochemical CO₂ reduction and contribute to the mechanistic understanding of multielectron and -proton transfer processes in CO₂ reduction.

In Figure 1 (open symbols), under visible-light irradiation (λ > 435 nm), we monitored the evolution of products as a function of time for a CO₂-saturated DMF (N,N’-dimethyl-formamide) solution containing 10 μM Fe-p-TMA, 1 mM Phen2, and 0.1 M TEA (triethylamine) acting as a sacrificial electron donor (SD). Large excess of Phen2 was used to ensure a strong light absorption as well as an efficient bimolecular subsequent reaction with the catalyst. Gratifyingly, we observed CH₄ production, albeit in moderate yield (TON of 8 after 47 h), alongside with the formation of H₂ (dihydrogen) and CO (TON of 8 and 50, respectively, Table 1 entry 1); note that the TON is defined as the mol number of product divided by the mol number of Fe-p-TMA. No other products such as formic acid, formaldehyde or methanol were detected. Importantly, the omission of any single reactive component (Fe catalyst, organic sensitizer, SD, CO₂, or light) produced no CH₄ product. Further, we found that with the addition of 0.1 M 2,2,2-trifluoroethanol (TFE) as an external acid, the production of CO and CH₄ was noticeably improved (TON of 71 and 14, respectively, see Figure 1 and Table 1 entry 2), whereas the production of H₂ remains almost unchanged (compare entries 1 and 2 in Table 1). GC/MS experiments performed under a ¹³CO₂ atmosphere confirmed that the produced CH₄ originated from CO₂ (Figure S1). Moreover, long-term irradiation (over 100 h) led to a TON in CH₄ of 29 and CO of 140 (Figure 2 and Table 1 entry 3). Catalytic selectivity for methane is 15%. The stability of the system was followed by UV–vis absorption spectroscopy over the entire irradiation course (Figure S3) and it showed no major Fe-p-TMA or Phen2 degradation. We have previously demonstrated that the dual catalysis employing Fe-p-TMA and Ir(ppy)₃ could also use CO as a starting substrate.¹¹ The rationale was that CO is an intermediary species toward the highly reduced CH₄ product. In Figure 3, using Phen2 as a photosensitizer in a CO-saturated solution, we observed the formation of CH₄ with a TON of 10 (Table 1 entry 4) and a selectivity of 30%, while H₂ was formed as the major product. The addition of 0.1 M TFE, however, significantly boosted CH₄ formation with a TON of 45 (Table 1 entry 5) and 87% selectivity upon 47 h of irradiation. The nature of the SD (Table 1, entries 5 to 8) had only minor effects on CH₄ production, e.g. similar results were obtained with TEA, DIPEA (N,N-diisopropylethylamine) and BIH (1,3-dimethyl-2-phenyl-2,3-dihydro-1H-benzo[d]imidazole), while TEOA (triethanolamine) gave lower CH₄ production. GC/MS experiments performed under a ¹³CO atmosphere again confirmed that the produced CH₄ was originated from CO (Figure S2). On the contrary, the types of acids used had a marked influence on CH₄ yield (Figure 3). Water, being a weaker acid than TFE, resulted in a much lower amount of CH₄ even with concentration up to 0.5 M (Table 1, entries 12 and 13). Conversely, the addition of 0.1 M PhOH (phenol, Table 1 entry 14), which is a stronger proton donor than TFE, resulted in forming H₂ as a major product and a decrease in CH₄ production. Further, the use of higher concentrations of TFE (greater than 0.1 M) reversed the catalytic selectivity toward H₂ formation (Table 1 entries 10 and 11). Thus, these results showed that 0.1 M TFE provided proper acidity and concentration to maximize CH₄ production and suppress H₂ evolution (see below for a more detailed mechanistic discussion).

Figure 1.

Generation of CO (black squares), H₂ (red circles) and CH₄ (blue diamonds) with time upon visible light irradiation (λ > 435 nm) of a CO₂-saturated DMF solution containing 10 μM Fe-p-TMA, 1 mM Phen2 and 0.1 M TEA (open symbols); addition of 0.1 M TFE (2,2,2-trifluoroethanol) as an external acid is indicated by “filled symbols”. Arrows indicate increase in product formation upon addition of an external acid.

Table 1.

Turnover Number (TON) of Gaseous Products Measured after 47 or 102 h^a of Visible Light (λ > 435 nm) Irradiation of DMF Solution Containing 10 μM Fe-p-TMA, 1 mM Phen2 and Various Components (SD, CO₂/CO, acid)

				TON
entry	gas	SD (M)	acid (M)	H2	CO	CH4
1	CO2	TEA (0.1)	none	8	50	8
2	CO2	TEA (0.1)	TFE (0.1)	10	71	14
3	CO2	TEA (0.1)	TFE (0.1)	23a	140a	29a
4	CO	TEA (0.1)	none	21	–	10
5	CO	TEA (0.1)	TFE (0.1)	7	–	45
6	CO	TEOA (0.1)	TFE (0.1)	9	–	21
7	CO	DIPEA (0.1)	TFE (0.1)	7	–	39
8	CO	BIH (0.1)	TFE (0.1)	8	–	46
9	CO	TEA (0.1)	TFE (0.1)	14a	–	80a
10	CO	TEA (0.1)	TFE (0.25)	17	–	27
11	CO	TEA (0.1)	TFE (0.5)	37	–	17
12	CO	TEA (0.1)	H2O (0.1)	5	–	10
13	CO	TEA (0.1)	H2O (0.5)	6	–	12
14	CO	TEA (0.1)	PhOH (0.1)	44	–	26

Figure 2.

CO (black squares), H₂ (red circles) and CH₄ (blue diamonds) generation with time upon visible light irradiation (λ > 435 nm) of a CO_2⁻ (filled symbols) or CO-saturated (open symbols) DMF solution containing 10 μM Fe-p-TMA, 1 mM Phen2, 0.1 M TEA and 0.1 M TFE.

Figure 3.

Catalytic turnovers in H₂ (red) and CH₄ (blue) measured after 47 h of visible light irradiation (>435 nm) of a CO-saturated DMF solution containing 10 μM Fe-p-TMA, 1 mM Phen2, in the presence of various SDs and added acids.

By employing 0.1 M TFE in a CO-saturated DMF solution irradiated for 102 h, we were able to produce CH₄ in 80 TON and 85% selectivity (Table 1, entry 9). The corresponding quantum yield is 0.47% based on the chemical actinometer method.²⁸ As a comparison, the noble metal Ir(ppy)₃ catalyzed CH₄ production in 159 TON, 81% selectivity and 0.18% quantum yield under similar reaction conditions (with 0.1 M TFE at optimized conditions).¹¹ We note that the Fe catalyst concentration in the Ir(ppy)₃ case was 5 times less as compared to this study, and thus the absolute mol number of CH₄ produced by the catalytic system comprising Fe-p-TMA + Phen2 is in fact ~2 times larger. Consequently, given the higher quantum yield and more CH₄ produced, the system employing the organic dye Phen2 is significantly more efficient. Table 1 summarizes key results from Figure 1, 2 and 3.

Notably, replacing Phen2 by the less reducing Phen1 led to the exclusive formation of CO (TON 60, selectivity 90%) and H₂ (TON 6) upon irradiation of a CO₂-saturated solution. However, irradiation of a CO-saturated solution with Phenl only furnished H₂ as the detectable product. These results indicate that the slightly lower reducing ability of Phenl led to the inability for further reduction beyond CO.

Investigating the mechanism, emission quenching experiments showed efficient quenching between Fe-p-TMA and the photoexcited Phen2 (or Phen2*) with a second order rate constant of k_q ≈ (1.60 ± 0.07) × 10⁸ M⁻¹ s⁻¹ (see Figures S4–S6), which supports an oxidative electron transfer from Phen2* to Fe-p-TMA. This result is in line with the fact that E⁰(²Phen2^•+/³Phen2*) = −2.20 V vs Fc⁺/Fc²⁷ is more negative than all three redox couples related to the Fe porphyrin (Fe^III/Fe^II, Fe^II/Fe^I and Fe^I/Fe⁰),^29–31 and thus allowing the generation of potential catalytically active Fe^II, Fe^I and Fe⁰ species upon light irradiation.

Scheme 2 highlights our proposed mechanism. As already observed and demonstrated in our previous electrochemical and photochemical studies, CO₂ first complexes to the triply reduced Fe⁰ species (to form Fe^IICO₂), which upon protonations and elimination of water, generates a Fe^IICO intermediate.^29–31 This intermediate has been detected by UV–vis absorption spectroscopy in a previous study.³¹ We note that the electron-rich Fe⁰ species can react with 2H⁺ to form Fe^II and the undesired H₂ byproduct, although it remains a minor pathway in our optimized conditions. It is only for higher concentration of the acid or in the presence of a stronger acid that protonation at the metal (and additionally at the ligand) may outcompete CO₂ insertion and favor H₂ production.^29,30 Note also that a high concentration of Fe⁰ active species in solution, which would be obtained upon highly efficient electron transfers from the sensitizer, may also favor H₂ evolution. The Fe^IICO intermediate can eliminate the CO product and form Fe^II. Alternatively, it may participate in further 6e⁻/6H⁺ reductions to produce the ultimate CH₄ product. DFT calculations are in progress to get hints on the reaction pathway from CO to methane and will be reported in due time. Note also that transposing the 8e⁻ catalysis of CO₂ to electrochemical conditions is hampered by the fact that the potential would be set at values negative enough to generate the Fe⁰ species that are reacting with CO₂. In the reaction–diffusion layer close to the electrode surface, only Fe⁰ and Fe^I are present in sizable amounts, with no Fe^II accumulation for further reduction of CO to CH₄.

Scheme 2.

Proposed Mechanism for the 8e⁻ /8H⁺ Reduction of CO₂ to CH₄ by Tandem Catalysis of Phen2 and Fe-p-TMA

In conclusion, we have successfully demonstrated the first noble metal free molecular system for visible light-driven 8e⁻/ 8H⁺ reduction of CO₂ to CH₄ employing an organic phenoxazine-based photosensitizer (Phen2) and an earth-abundant iron porphyrin catalyst (Fe-p-TMA) at ambient conditions. In a CO₂-saturated DMF solution, CO (TON of 140) and CH₄ (TON of 29) were produced after 102 h of light irradiation; whereas in a CO-saturated solution, CH₄ was produced with TON of 80, a selectivity of 85% and a quantum yield of 0.47%. Remarkably, Phen2 was significantly more efficient than Ir(ppy)3, producing ~2 times the amount of CH₄ with ~3 times higher quantum yield under similar reaction conditions. We envision that this work will open up new perspectives toward the development of integrated (photo)-electrochemical catalytic systems where the multielectron and -proton conversion of the CO₂ will be coupled to oxidation of water, biomass or organic compounds for sustainable solar fuels production.