Unlocking solar-derived CO2 conversion to hydrocarbons over plasmonic Cu-Fe tandem catalysts under industrially relevant conditions

Imported from: /opt/uploadtmp/s41467-026-75764-5_reference_6a7ec9223442c.pdf

Abstract Summary

This article studies photo-thermal CO₂ hydrogenation to hydrocarbons over a heterogeneous Cu–Fe catalyst, not a molecular photocatalytic CO₂ reduction system. The reported system uses a potassium-promoted Cu₂Fe₁ tandem catalyst under continuous-flow gas-phase conditions with H₂/CO₂ feed and Xe-lamp illumination. The main finding is that under illumination, 250 °C external heating, and elevated pressure, the catalyst gives high selectivity to C₂⁺ hydrocarbons, reaching 77.4% C₂⁺ hydrocarbon selectivity at 20 bar and a C₂–C₅ hydrocarbon yield of 5.04 mmol g⁻¹ h⁻¹. Mechanistic data support a tandem pathway in which Cu promotes the reverse water-gas shift reaction to form CO, while in situ formed χ-Fe₅C₂ drives Fischer–Tropsch-type C–C coupling and hydrocarbon formation.

Advances and Special Progress

The main advance reported is the use of a plasmonic Cu–Fe tandem catalyst that combines photo-thermal activation with tandem catalysis under conditions described as industrially relevant by the article. The catalyst design uses earth-abundant metals and potassium promotion rather than noble-metal plasmonic components.

A second advance is performance. The article states that the catalyst surpasses previously reported photo-thermal CO₂-to-hydrocarbon benchmarks in C₂⁺ production, especially under continuous-flow and elevated-pressure operation. At 20 bar, the system strongly suppresses CO relative to lower pressure and gives high hydrocarbon selectivity.

A third advance is mechanistic insight. The article combines direct versus indirect illumination tests, in situ DRIFTS, in situ XRD, CO-TPR, CO₂-TPSR, UV–vis spectroscopy, transient photocurrent, and time-resolved photoluminescence. These data support two linked conclusions: illumination accelerates formation of reactive surface intermediates, and illumination lowers the temperature required for formation of χ-Fe₅C₂, the iron carbide phase associated with hydrocarbon synthesis.

A fourth advance is operational robustness. After an activation period, the catalyst remains stable during repeated light on/off cycles over 50 h without obvious deactivation in conversion or selectivity.

Additional Remarks

Although the prompt requests a molecular photocatalytic CO₂ reduction entry, the article does not describe a molecular photosensitizer/catalyst pair in solution. Instead, it reports a gas-phase heterogeneous photo-thermal catalytic system using a solid Cu–Fe catalyst, H₂ as reductant, and external heating together with light irradiation.

Chemically, the work is significant because it targets direct formation of multicarbon hydrocarbons from CO₂, which is more demanding than simple CO formation. The study emphasizes that C–C coupling under photo-thermal conditions is difficult because methane formation and the Anderson–Schulz–Flory distribution limit selectivity to desired higher products.

The article also makes clear that the system is not purely photochemical. It is a coupled photo-thermal process in which both heating and light matter. Direct versus indirect illumination experiments indicate that non-thermal light effects are important, but thermal contributions remain present. The chemistry therefore should be understood as plasmon-assisted catalytic hydrogenation rather than classical homogeneous photocatalysis.

Another important point is that the active iron carbide phase is formed in situ during reaction. The need for an activation period of about 14 h before steady state reflects this structural evolution.

Content of the Published Article in Detail

The reported system consists of a potassium-promoted bimetallic Cu–Fe solid catalyst used for gas-phase CO₂ hydrogenation under illumination. The optimized catalyst composition is 2K Cu₂Fe₁, where Cu/Fe = 2 and K loading is 2 wt%. The catalyst is prepared by co-precipitation, calcination, reduction, and passivation before use.

The photocatalytic setup is actually a photo-thermal continuous-flow reactor. A quartz-window reactor is irradiated with a 300 W Xe lamp. The reaction feed is H₂/CO₂ = 4:1, with N₂ as internal standard. Typical tests use 50 mg catalyst, gas hourly space velocity of 12000 mL g⁻¹ h⁻¹, external heating at 250 °C, and pressures from 10 to 20 bar. Thus, the system is not a liquid-phase sacrificial-donor photocatalytic CO₂-to-CO system, but a solid-gas hydrogenation system driven by combined light and heat.

The article first optimizes composition. Monometallic Fe shows some CO₂ hydrogenation activity and some C₂⁺ selectivity. Monometallic Cu is almost exclusively selective to CO, showing strong reverse water-gas shift activity but poor hydrocarbon formation. Bimetallic CuFe materials perform better, and Cu₂Fe₁ gives the best balance of CO₂ conversion and hydrocarbon selectivity among the tested Cu/Fe ratios. Potassium further improves performance by strengthening CO₂ adsorption, increasing CO adsorption capacity and binding strength, and increasing chain-growth probability toward longer hydrocarbons.

Under 4 W cm⁻² illumination, 250 °C, 20 bar, and 12000 mL g⁻¹ h⁻¹, the optimized catalyst gives 53.5% CO₂ conversion, 23.5% CO selectivity, and 77.4% selectivity to C₂⁺ hydrocarbons on a CO-free hydrocarbon basis. The C₂–C₅ hydrocarbon yield is 5.04 mmol g⁻¹ h⁻¹, of which 4.1 mmol g⁻¹ h⁻¹ is C₂–C₅⁼ olefins. Lower gas hourly space velocity increases conversion further but also increases hydrogenation and methane/paraffin formation.

Several characterization methods identify the active phases before and after reaction. Before reaction, reduced 2K Cu₂Fe₁ contains metallic Cu and metallic Fe, with some surface oxidation from passivation. After reaction, powder X-ray diffraction, Mössbauer spectroscopy, TEM, XPS, and TGA-TPH show that the Fe phase is largely converted into χ-Fe₅C₂. The article interprets this iron carbide as the active Fischer–Tropsch phase responsible for C–C coupling and hydrocarbon formation. Metallic Cu remains present and is linked to reverse water-gas shift activity.

The mechanism is described as a tandem pathway. In situ DRIFTS under dark conditions shows bicarbonate and carbonate species at lower temperature, then formate species as temperature rises, then adsorbed CO species, and finally hydrocarbon-related bands at 250 °C. Based on these observations, the article proposes the following sequence: CO₂ adsorbs as bicarbonate/carbonate species, these are hydrogenated to formate, formate decomposes to CO, and the resulting CO is further hydrogenated over iron carbide-derived sites to hydrocarbons through a Fischer–Tropsch-type pathway. Methane and olefin/paraffin products are both observed.

Comparison of monometallic catalysts supports division of labor between the two metals. On Cu, DRIFTS detects carbonate, formate, and adsorbed CO features but no significant methane signal, supporting a primary reverse water-gas shift role. On Fe, carbonate, formate, CO-related species, and weak hydrocarbon bands appear, supporting the ability of iron-based sites to perform downstream CO hydrogenation.

The role of light is investigated carefully. Direct versus indirect illumination experiments use a Ti₂O₃ overlayer to block direct photochemical excitation while preserving similar heating. Under direct illumination, CO₂ conversion increases linearly with light intensity from 2.8 to 4.0 W cm⁻², which the article interprets as evidence for dominant non-thermal light effects with some thermal contribution. Indirect illumination gives much lower conversion, supporting a beneficial effect of direct photoexcitation. At high intensity, the non-thermal contribution appears to level off.

At 200 °C, light makes a major chemical difference. Under illumination, the catalyst gives 17% CO₂ conversion and more than 20% hydrocarbon selectivity; in the dark, CO₂ conversion is only 11.5% and CO selectivity is 95.6%. XRD shows that χ-Fe₅C₂ forms under illuminated conditions at 200 °C, whereas metallic Fe remains predominant in the dark. This supports the article’s proposal that illumination facilitates iron carburization and thereby enables hydrocarbon synthesis at lower temperature.

In situ DRIFTS under illumination shows the same types of intermediates as in the dark, so the article concludes that light does not change the overall reaction pathway. However, the same intermediates appear at lower temperatures and with stronger intensities under illumination. The data therefore support faster generation of bicarbonate/carbonate, adsorbed CO, and hydrocarbon-related intermediates when light is present.

In situ XRD and CO-TPR give further mechanistic support. During reaction, the metallic Fe diffraction peak disappears at lower temperature under illumination than in the dark, and χ-Fe₅C₂ reflections emerge earlier with light. In CO-TPR, initial CO uptake begins at 130 °C under illumination versus 190 °C in the dark, and Fe reduction/carburization associated with χ-Fe₅C₂ formation occurs around 260 °C under illumination, nearly 50 °C lower than in the dark. CO₂-TPSR also shows formation of CO, CH₄, and hydrocarbon fragments at lower temperature under illumination.

To explain these effects, the article considers three possible origins for faster carburization and concludes that the third is most plausible: plasmonic Cu absorbs light, generates hot carriers, accelerates reverse water-gas shift chemistry, and promotes interfacial charge transfer and/or localized heating at the Cu–Fe interface. This, in turn, lowers the apparent barrier for χ-Fe₅C₂ formation. UV–vis spectroscopy shows a plasmon band for metallic Cu at 505 nm. The wavelength dependence of activity also peaks around 500 nm under quasi-monochromatic irradiation, supporting involvement of Cu-associated localized surface plasmon resonance. Transient photocurrent and time-resolved photoluminescence further support improved charge separation and rapid electron transfer in the bimetallic catalyst relative to monometallic Cu or Fe.

No molecular photosensitizer, reductive or oxidative quenching pathway, sacrificial electron donor, or reduced molecular catalyst state is reported. The reduction equivalents come from H₂ in the gas feed, and the chemistry proceeds on illuminated solid catalyst surfaces.

Catalyst

The catalyst is a heterogeneous potassium-promoted bimetallic Cu–Fe catalyst, optimized as 2K Cu₂Fe₁. It is not a molecular catalyst. Before reaction, the reduced material contains metallic Cu and metallic Fe, with minor surface oxidation attributed to passivation. During reaction, the Fe phase is converted in situ into χ-Fe₅C₂.

Functionally, the catalyst is a tandem system. Cu is associated mainly with the reverse water-gas shift step that converts CO₂ to CO, while χ-Fe₅C₂ is associated with Fischer–Tropsch-type hydrogenation and C–C coupling to produce hydrocarbons. Potassium promotion enhances CO₂ adsorption, strengthens CO binding, and shifts selectivity toward longer hydrocarbons. Excess potassium decreases activity, which the article attributes to over-coverage of active sites.

The catalyst is also described as plasmonic because metallic Cu shows a localized surface plasmon absorption band near 505 nm. This plasmonic response is tied to non-thermal light effects and to accelerated iron carburization at the Cu–Fe interface.

Photosensitizer

No molecular photosensitizer is reported. The light-harvesting function is performed by the solid catalyst itself, especially the metallic Cu component, which the article identifies as plasmonic.

The article supports this assignment through UV–vis spectroscopy showing a Cu plasmon absorption band at 505 nm, photocurrent measurements showing the strongest response for the bimetallic catalyst, time-resolved photoluminescence indicating faster charge-transfer dynamics in the bimetallic material, and wavelength-dependent catalysis showing highest activity near 500 nm. Thus, the light-absorbing component is the heterogeneous Cu-containing catalyst rather than a separate molecular photosensitizer.

Investigation

	cat	cat conc [µM]	PS	PS conc [mM]	e-D	e-D conc [M]	.	.	solvent A	.	.	.	additives	.	.	.	.	λexc [nm]	.	TON CO	.	.	TON CH4	TON H2	.	TON HCOOH	.	.
1.	2K Cu2Fe1	not reported	not reported	not reported	H2O	not reported			not reported				Molecule:100493					not reported		not reported			not reported	not reported		not reported			"not reported" is not a number. "wt%" is not declared as a valid unit of measurement for this property. "H2/CO2 = 4:1" is not in the list (nitrogen, argon, CO2, CO, O2, H2, none) of allowed values for the "Feedstock gas" property.
2.	2K Cu2Fe1	not reported	not reported	not reported	H2O	not reported			not reported				Molecule:100493					not reported		not reported			not reported	not reported		not reported			"not reported" is not a number. "wt%" is not declared as a valid unit of measurement for this property. "H2/CO2 = 4:1" is not in the list (nitrogen, argon, CO2, CO, O2, H2, none) of allowed values for the "Feedstock gas" property.
3.	2K Cu2Fe1	not reported	not reported	not reported	H2O	not reported			not reported				Molecule:100493					not reported		not reported			not reported	not reported		not reported			"not reported" is not a number. "wt%" is not declared as a valid unit of measurement for this property. "H2/CO2 = 4:1" is not in the list (nitrogen, argon, CO2, CO, O2, H2, none) of allowed values for the "Feedstock gas" property.
4.	2K Cu2Fe1	not reported	not reported	not reported	H2O	not reported			not reported				Molecule:100493					not reported		not reported			not reported	not reported		not reported			"not reported" is not a number. "wt%" is not declared as a valid unit of measurement for this property. "H2/CO2 = 4:1" is not in the list (nitrogen, argon, CO2, CO, O2, H2, none) of allowed values for the "Feedstock gas" property.
5.	2K Cu2Fe1	not reported	not reported	not reported	H2O	not reported			not reported				Molecule:100493					not reported		not reported			not reported	not reported		not reported			"not reported" is not a number. "wt%" is not declared as a valid unit of measurement for this property. "H2/CO2 = 4:1" is not in the list (nitrogen, argon, CO2, CO, O2, H2, none) of allowed values for the "Feedstock gas" property.
6.	2K Cu2Fe1	not reported	not reported	not reported	H2O	not reported			not reported				Molecule:100493					not reported		not reported			not reported	not reported		not reported			"not reported" is not a number. "wt%" is not declared as a valid unit of measurement for this property. "H2/CO2 = 4:1" is not in the list (nitrogen, argon, CO2, CO, O2, H2, none) of allowed values for the "Feedstock gas" property.
7.	2K Cu2Fe1	not reported	not reported	not reported	H2O	not reported			not reported				Molecule:100493					not reported		not reported			not reported	not reported		not reported			"not reported" is not a number. "wt%" is not declared as a valid unit of measurement for this property. "H2/CO2 = 4:1" is not in the list (nitrogen, argon, CO2, CO, O2, H2, none) of allowed values for the "Feedstock gas" property.
8.	2K Cu2Fe1 with Ti2O3 cover	not reported	not reported	not reported	H2O	not reported			not reported				K/Ti2O3					not reported		not reported			not reported	not reported		not reported			"not reported" is not a number. "wt%/notreported" is not declared as a valid unit of measurement for this property. "H2/CO2 = 4:1" is not in the list (nitrogen, argon, CO2, CO, O2, H2, none) of allowed values for the "Feedstock gas" property.
9.	2K Cu2Fe1	not reported	not reported	not reported	H2O	not reported			not reported				Molecule:100493					not reported		not reported			not reported	not reported		not reported			"not reported" is not a number. "wt%" is not declared as a valid unit of measurement for this property. "H2/CO2 = 4:1" is not in the list (nitrogen, argon, CO2, CO, O2, H2, none) of allowed values for the "Feedstock gas" property.
10.	2K Cu2Fe1	not reported	not reported	not reported	H2O	not reported			not reported				Molecule:100493					not reported		not reported			not reported	not reported		not reported			"not reported" is not a number. "wt%" is not declared as a valid unit of measurement for this property. "H2/CO2 = 4:1" is not in the list (nitrogen, argon, CO2, CO, O2, H2, none) of allowed values for the "Feedstock gas" property.
11.	2K Cu2Fe1	not reported	not reported	not reported	H2O	not reported			not reported				Molecule:100493					not reported		not reported			not reported	not reported		not reported			"not reported" is not a number. "wt%" is not declared as a valid unit of measurement for this property. "H2/CO2 = 4:1" is not in the list (nitrogen, argon, CO2, CO, O2, H2, none) of allowed values for the "Feedstock gas" property.
12.	2K Cu2Fe1	not reported	not reported	not reported	H2O	not reported			not reported				Molecule:100493					not reported		not reported			not reported	not reported		not reported			"not reported" is not a number. "wt%" is not declared as a valid unit of measurement for this property. "H2/CO2 = 4:1" is not in the list (nitrogen, argon, CO2, CO, O2, H2, none) of allowed values for the "Feedstock gas" property.
13.	2K Cu2Fe1 with Ti2O3 cover	not reported	not reported	not reported	H2O	not reported			not reported				K/Ti2O3					not reported		not reported			not reported	not reported		not reported			"not reported" is not a number. "wt%/notreported" is not declared as a valid unit of measurement for this property. "H2/CO2 = 4:1" is not in the list (nitrogen, argon, CO2, CO, O2, H2, none) of allowed values for the "Feedstock gas" property.
14.	2K Cu2Fe1 with Ti2O3 cover	not reported	not reported	not reported	H2O	not reported			not reported				K/Ti2O3					not reported		not reported			not reported	not reported		not reported			"not reported" is not a number. "wt%/notreported" is not declared as a valid unit of measurement for this property. "H2/CO2 = 4:1" is not in the list (nitrogen, argon, CO2, CO, O2, H2, none) of allowed values for the "Feedstock gas" property.
15.	2K Cu2Fe1 with Ti2O3 cover	not reported	not reported	not reported	H2O	not reported			not reported				K/Ti2O3					not reported		not reported			not reported	not reported		not reported			"not reported" is not a number. "wt%/notreported" is not declared as a valid unit of measurement for this property. "H2/CO2 = 4:1" is not in the list (nitrogen, argon, CO2, CO, O2, H2, none) of allowed values for the "Feedstock gas" property.
16.	2K Cu2Fe1 with Ti2O3 cover	not reported	not reported	not reported	H2O	not reported			not reported				K/Ti2O3					not reported		not reported			not reported	not reported		not reported			"not reported" is not a number. "wt%/notreported" is not declared as a valid unit of measurement for this property. "H2/CO2 = 4:1" is not in the list (nitrogen, argon, CO2, CO, O2, H2, none) of allowed values for the "Feedstock gas" property.

Investigation-Name: inv0ExportRefresh