Utilizing CO₂ as a Reactant for C₃ Oxygenate Production via Tandem Reactions

Abstract

One possible solution to closing the loop on carbon emissions is using CO₂ as the carbon source to generate high-value, multicarbon products. In this Perspective, we describe four tandem reaction strategies for converting CO₂ into C₃ oxygenated hydrocarbon products (i.e., propanal and 1-propanol), using either ethane or water as the hydrogen source: (1) thermocatalytic CO₂-assisted dehydrogenation and reforming of ethane to ethylene, CO, and H₂, followed by heterogeneous hydroformylation, (2) one-pot conversion of CO₂ and ethane using plasma-activated reactions in combination with thermocatalysis, (3) electrochemical CO₂ reduction to ethylene, CO, and H₂, followed by thermocatalytic hydroformylation, and (4) electrochemical CO₂ reduction to CO, followed by electrochemical CO reduction to C₃ oxygenates. We discuss the proof-of-concept results and key challenges for each tandem scheme, and we conduct a comparative analysis of the energy costs and prospects for net CO₂ reduction. The use of tandem reaction systems can provide an alternative approach to traditional catalytic processes, and these concepts can be further extended to other chemical reactions and products, thereby opening new opportunities for innovative CO₂ utilization technologies.

1. Introduction

To prevent the catastrophic, irreversible consequences of climate change, global warming should be limited to below 1.5 °C above preindustrial levels.¹ In addition to deep emission reductions, negative emission technologies that actively remove CO₂ from the atmosphere will also play a critical role. This can be accelerated with economical CO₂ conversion technologies that utilize CO₂ as the carbon source to produce value-added fuels and chemicals, which may result in net-negative CO₂ emissions when considering the avoided emissions associated with replacing fossil-fuel-derived products.

In particular, C₃ oxygenated hydrocarbons (C₃ Oxys), such as propanol and propanal, represent desirable products, since they are used as versatile solvents and feedstocks in many chemical industrial processes.^2,3 Propanol is also an excellent fuel additive because it has a higher octane number and heating value than methanol or ethanol, as well as lower emissions when it is combusted.⁴ However, due to its complex production process, propanol currently has a higher market price and smaller global production volume (0.2 Mt/year) than other mass-produced chemicals such as methanol (150 Mt/year), ethanol (77 Mt/year), and ethylene (140Mt/year).⁵ Therefore, the conversion of CO₂ to C₃ Oxys likely represents one of the shorter-term pathways to economic feasibility by displacing the existing conventional processes and making more immediate progress toward decarbonizing the chemicals industry. Furthermore, the global market usage of C₃ Oxy platform molecules for chemicals and fuels may increase if a cheaper, more sustainable production method emerges.

At present, C₃ Oxys are conventionally produced via hydroformylation (also known as oxo synthesis). However, the industrial high-pressure homogeneous hydroformylation process involves large energy costs due to the high pressures (10–30 atm) needed to obtain a high product yield, in addition to the energy required for product separation and homogeneous catalyst recovery.⁶ Moreover, the feedstocks of ethylene and syngas (CO and H₂) are also generated using energy-intensive and carbon-emitting processes: ethylene is typically produced from thermal steam cracking of naphtha, crude oil, or natural gas, and syngas is primarily derived from natural gas steam reforming or coal gasification. The involvement of multiple independent process steps also leads to operational challenges and safety risks during the transportation and storage of intermediate products (i.e., toxic syngas and flammable ethylene).

Instead, C₃ Oxy synthesis methods that can operate under mild conditions, utilize renewable carbon-free energy sources, and consume CO₂ as a reactant may provide more sustainable alternatives to traditional processes. The direct coupling of multiple reactors in a tandem configuration can also eliminate the separation, transportation, or storage steps that may otherwise be associated with obtaining hydroformylation feedstocks from separate, independent processes. In addition to CO₂ as the carbon source, a hydrogen source is necessary for oxygenate production. However, it is unlikely that the utilization of molecular H₂ can achieve a net reduction of CO₂. At present, ∼95% of H₂ is derived from hydrocarbon-based feedstocks and generates CO₂ as a byproduct.⁷ Furthermore, a previous CO₂ mass balance analysis for methanol synthesis using hybrid electrocatalytic and thermocatalytic schemes revealed that CO₂ conversion to methanol that requires H₂ as a reactant is net CO₂-positive even when using relatively decarbonized energy sources,⁸ suggesting that CO₂ + H₂ to propanol should be even less effective for net CO₂ reduction due to selectivity issues. Alternatively, light alkanes (via thermal or plasma activation) or H₂O (via electrochemical activation) may be used as the hydrogen source for CO₂ conversion. However, CO₂ as well as alkanes/H₂O are thermodynamically stable and difficult (or nearly impossible) to convert directly into C₃ Oxys in a single reactor. Therefore, the application of tandem reaction strategies coupling thermocatalysis, electrocatalysis, or plasma catalysis is necessary for effectively upgrading CO₂ to C₃ Oxys.

In this Perspective, we consider four ambient-pressure tandem reaction pathways (Figure 1) for producing C₃ Oxys from CO₂, using either ethane or water as the hydrogen source: (1) thermocatalytic CO₂-assisted dehydrogenation and reforming of ethane to ethylene, CO, and H₂, followed by thermocatalytic heterogeneous hydroformylation (TC-TC),^9,10 (2) one-pot conversion of CO₂ and ethane using plasma-activated reactions in combination with thermocatalysis (P-TC),¹¹ (3) electrocatalytic CO₂ reduction to ethylene, CO, and H₂, followed by thermocatalytic hydroformylation (EC-TC),¹² and (4) electrocatalytic CO₂ reduction to CO, followed by electrocatalytic CO reduction (EC-EC).¹³ We first introduce the relevant reactions and proof-of-concept results for the four different tandem schemes. Then, we compare the energy costs and prospects for net CO₂ reduction for each process. Finally, we discuss the key challenges and opportunities for tandem reaction systems.

Figure 1.

Schematic of tandem reaction strategies for converting CO₂ into C₃ oxygenated hydrocarbons. In the reaction schemes, C₃ oxygenated hydrocarbons (C₃ Oxys) are produced by reacting CO₂ with ethane via tandem thermocatalytic or one-pot plasma-catalytic conversion (top schemes, blue) or with H₂O via electrocatalytic reduction followed by thermocatalytic hydroformylation or electrocatalytic CO reduction (bottom schemes, yellow).

2. Tandem Reaction Strategies for the Production of C₃ Oxygenated Hydrocarbons

2.1. Upgrading CO₂ with Ethane

The supply of natural gas has increased significantly in recent decades as a result of advances in drilling and fracking techniques, as well as the discovery of large shale gas reserves (currently ∼7.6 quadrillion cubic feet of global recoverable reserves).⁷ The abundant supply, and therefore low prices, of natural gas has motivated efforts to upgrade light alkanes into value-added products. In particular, after methane, ethane represents the second most abundant component in shale gas deposits (3–16%), but it is typically underutilized.¹⁴ Therefore, the abundant ethane in natural gas can be used as a carbon source as well as the hydrogen source to react with CO₂, instead of using CO₂-intensive H₂ derived from natural gas as the hydrogen source. The consumption of CO₂ as a coreactant with ethane obtained as surplus during natural gas extraction can potentially lead to a neutral or negative-emitting process (on the basis of carbon that has already been released from underground storage).¹⁵ In this section, we describe tandem thermocatalytic–thermocatalytic and plasma-catalytic schemes for simultaneously upgrading CO₂ and ethane to C₃ Oxys.

2.1.1. Thermocatalytic–Thermocatalytic Scheme (TC-TC)

In the TC-TC tandem reaction scheme, ethylene and syngas are first coproduced via CO₂-assisted oxidative dehydrogenation of ethane (CO₂–ODHE) and dry reforming of ethane (DRE), as shown in eqs R1 and R2, respectively.

R1

R2

Subsequently, ethylene and syngas are used as the reactants in the downstream hydroformylation reaction to produce propanal, as shown in eq R3; 1-propanol can also be produced by hydrogenation of propanal.

R3

Figure 2 shows a thermodynamic analysis of relevant reactions at ambient pressure. As represented by the purple line, the direct one-step conversion of CO₂ and ethane to C₃ Oxys is thermodynamically unfavorable (ΔG⁰ > 0) across the entire temperature range. It is also evident that CO₂-ODHE (blue line)—which is the sum of the direct dehydrogenation of ethane (DDHE) and reverse water-gas shift (RWGS) reactions (black and green lines, respectively)—and DRE (red line) are favored at higher temperatures, while the hydroformylation reaction (gold line) is favored at lower temperatures. Therefore, a two-step tandem approach with CO₂-ODHE&DRE and hydroformylation reactors operating within their respective favorable temperature regimes can bridge this temperature gap.

Figure 2.

Diagram of standard Gibbs free energy change (ΔG⁰) as a function of temperature for the relevant reactions of CO₂ and ethane. Reprinted with permission from Xie et al.⁹ This work is licensed under a Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/).

In comparison with the nonoxidative DDHE process, CO₂-ODHE favors coke elimination (via the reverse Boudouard reaction: CO₂ + C → 2CO) and forward-shifts the reaction (via the RWGS reaction: CO₂ + H₂ → CO + H₂O).¹⁶⁻¹⁸ The use of CO₂ as a soft oxidant, as opposed to O₂, also mitigates overoxidation of hydrocarbons as well as the safety issues associated with the strongly exothermic O₂-assisted dehydrogenation (O₂-ODHE).¹⁹ By tuning the relative contributions of the CO₂-ODHE and DRE reactions, the optimal mixture of C₂H₄, CO, and H₂ can be produced and fed to the downstream hydroformylation reactor. However, the DRE reaction typically outperforms the CO₂-ODHE reaction, as cleavage of the ethane C–C bond (368 kJ mol^–1) is more thermodynamically favorable than C–H bond scission (415 kJ mol^–1); thus, it remains challenging to enhance ethylene selectivity. Using a nonprecious Fe₃Ni₁/CeO₂ bimetallic catalyst, Xie et al. observed much lower C₂H₄ production relative to syngas production below 700 °C; however, incorporating the homogeneous contribution to ethylene production at 750 °C and above, the C₂H₄/CO/H₂ product ratio was similar to the 1/1/1 stoichiometric feed ratio required for the hydroformylation reaction (Figure 3a).⁹

Figure 3.

TC-TC tandem performance. (a) Amount of products formed during the first reaction step of CO₂ and C₂H₆ over a Fe₃Ni₁/CeO₂ catalyst at different temperatures. (b) C₂H₄-based product yields in a second reactor over the MCM-41-supported RhCo_x catalysts at 200 °C. The values within the bars of Figure 3a,b indicate product selectivity. (c) C₂H₆-based yield in the tandem configuration (first reactor, 600–800 °C, Fe₃Ni₁/CeO₂ catalyst; second reactor, 200 °C, RhCo_x catalysts). Reprinted with permission from Xie et al.⁹ This work is licensed under a Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/). (d, e) C₂H₆-based selectivity, and (f, g) C₂H₆-based yield and productivity for the tandem reactor system with reduced temperatures in the first reactor (first reactor, 550–600 °C, PtSn₃/γ-Al₂O₃ catalyst; second reactor, 200 °C, RhCo₃ catalyst). The upper and lower panels in Figure 3d–g represent the results obtained for the first reactor at 550 and 600 °C, respectively. The pink dashed line in Figure 3g indicates the productivity of total C₃ oxygenates obtained in the tandem system containing the Fe₃Ni₁/CeO₂ catalyst at 750 °C. Reprinted with permission from Xie et al.¹⁰ Copyright 2022, American Chemical Society.

For the conventional homogeneous hydroformylation of ethylene, high pressures are typically required in order to increase the concentration of reactants (i.e., olefins, CO, and H₂) in the solvent. Industrial phosphine-modified rhodium (Rh) catalysts have demonstrated very high activities (i.e., turnover frequencies up to 600 min^–1 for α-olefins) with very low byproduct formation and alkene isomerization.²⁰ However, the homogeneous process deals with separation challenges, which may lead to precious-metal leaching and the generation of phosphorus-containing waste.²¹ In contrast, in the heterogeneous hydroformylation of ethylene, an appreciable surface coverage of such reactants can be obtained over oxide-supported metal catalysts,²²⁻²⁵ thereby allowing for reduced reaction pressures. As a result, heterogeneous supported metal catalysts can even be used at ambient pressure, as opposed to the typical homogeneous ligand-modified metal complexes used in industrial practice.⁶ In general, unmodified Rh metal is more active than other monometallic catalysts and enables the hydroformylation reaction under milder conditions. Liu et al. summarized the performance of various heterogeneous Rh-based hydroformylation catalysts, including the use of organic phosphines, inorganic phosphides, and second metal modifications.²⁵

As shown in Figure 3b, Xie et al. demonstrated that increasing the Co/Rh ratio in a RhCo bimetallic catalyst increased the C₃ Oxy yield as well as the alcohol/aldehyde product ratio.⁹ The presence of Co was found to increase Rh particle dispersion and favor hydrogenation to 1-propanol, due to the stronger binding of key oxygenate intermediates (*CH₃CH₂CHO and *CH₃CH₂CH₂O) on the RhCo bimetallic surface compared to the pure Rh surface, based on DFT calculations.⁹ In the directly coupled tandem TC-TC configuration, the highest ethane-based yield of C₃ Oxys (4.7%) was obtained with Fe₃Ni₁/CeO₂ at 800 °C in the first reactor and Rh₁Co₃/MCM-41 at 200 °C in the second reactor (Figure 3c).⁹ As expected, higher temperatures in the first CO₂-ODHE/DRE reactor and a higher Co/Rh ratio in the second hydroformylation reactor led to the optimal production of C₃ Oxys.

With the homogeneous contribution to ethylene formation at high temperatures, the control over selectivity was compromised. Thus, it is critical to reduce the temperature in the first reactor to enable a completely catalytic tandem process, in which the C₂H₄/CO/H₂ ratios can be better tuned via the selection of appropriate catalysts. Xie et al. recently analyzed and identified the desired C₂H₄/CO/H₂ ratios by considering different combinations of main and side reactions of CO₂ and ethane.¹⁰ To obtain a C₂H₄/CO/H₂ ratio that is close to the typical feed ratio (1/1/1) for the subsequent hydroformylation, the analysis showed that the desired catalysts for the first reactor should enable multiple simultaneous reactions of direct dehydrogenation of ethane (DDHE), ODHE, DRE, and RWGS: i.e., either ODHE + DDHE + DRE (theoretical product ratio of 1/1/0.88) or DDHE + RWGS + DRE (theoretical product ratio of 1/1.15/1). Accordingly, a typical dehydrogenation catalyst, i.e., PtSn₃/γ-Al₂O₃, was identified to be promising to promote the simultaneous DDHE (major), ODHE (minor), and DRE reactions at a lower temperature (i.e., 550–600 °C) and, in turn, supply a mixture of C₂H₄, CO, and H₂ (1/0.9/0.4) for the second reactor (Figure 3d–g). It should be noted that the stoichiometric ratio (1/1/1) is a typical benchmark for single-site catalysts (e.g., RhCo_x/MCM-41 in Figure 3) in conventional hydroformylation processes. Yet, the preferred ratio could be different in the case of catalysts with bifunctional active sites, as demonstrated by the recently reported Rh-WO_x pair sites, where a higher fraction (or partial pressure) of CO is favorable for propanal formation.²⁴ In the tandem configuration with PtSn₃/γ-Al₂O₃ catalyst at 600 °C in the first reactor, the obtained C₃ Oxy productivity was comparable to that of the aforementioned Fe₃Ni₁/CeO₂ catalyst at 750 °C (Figure 3g).¹⁰

2.1.2. Plasma–Thermocatalytic Scheme (P-TC)

As shown in the thermodynamic analysis in Figure 2, the one-step conversion of CO₂ and ethane to C₃ Oxys is not feasible; a gap exists between the high temperatures required to activate the stable C=O and C–H bonds and the low temperatures that favor the exothermic oxygenate production. Instead of using a two-step TC-TC approach, however, nonthermal plasma can be used to overcome the thermodynamic limitations and produce C₃ Oxys in one step under mild conditions. In nonthermal plasma, a large voltage difference between two electrodes is used to activate the gaseous reactants and form a nonequilibrium phase containing electrons, ions, radicals, atoms, molecules, and other excited species, which can subsequently react and recombine into desirable products. Despite the bulk gas remaining at or near ambient conditions, the highly energized electrons (typically 1–10 eV) within the plasma can overcome the energy barriers associated with chemical ionization and bond dissociation (i.e., E_diss = 4.4 eV for C=O, E_diss = 5.5 eV for C–H).²⁶ Plasma also offers other distinct advantages over thermocatalytic processes, including fast reaction rates that allow for rapid process startup/shutdown and easier integration with intermittent renewable electricity sources.²⁷

Plasma-activated reactions between CO₂ and ethane have been demonstrated to generate C₃ Oxys.¹¹ The complexity of the production and destruction pathways occurring within the plasma (Figure 4a) has been described for the full list of reactions as determined by plasma chemical kinetic modeling by Biswas et al.¹¹ The formation of C₃ Oxys, among other products, from a plasma-activated reaction of CO₂ and ethane was achieved using a dielectric barrier discharge (DBD) reactor at atmospheric pressure and 200 °C. In addition to CO and C₁–C₅₊ hydrocarbons, a variety of oxygenate products were generated, including formaldehyde, acetic acid, methanol, ethanol, 1-propanol, 2-propanol, and propanal (Figure 4b,c). Increasing the plasma power was found to increase CO₂ and ethane reactant conversions, but higher total oxygenate selectivities were obtained at lower powers. Additionally, a greater CO₂/ethane feed ratio reduced the formation of hydrocarbons in favor of CO and oxygenated species, with a maximum oxygenate selectivity of 12% obtained at a 4/1 CO₂/ethane feed ratio.

Figure 4.

P-TC tandem performance. (a) Schematic of the most important CO₂ + ethane plasma reaction pathways, where the thickness of arrows and frames indicates the relative importance of the corresponding pathways and product densities, respectively. Dotted lines indicate very low rates and densities. (b) Flow rates of CO and hydrocarbon products and (c) flow rates of oxygenate products for the plasma-only reaction of CO₂ and ethane at 10.0 kV and 9 kHz under ambient pressure and 200 °C. (d) Effect of adding RhCo₃/MCM-41 catalyst on oxygenate selectivity. Reprinted with permission from Biswas et al.¹¹ Copyright 2022, American Chemical Society.

The primary issue limiting the plasma-only reaction of CO₂ and ethane is low product selectivity. The highest C₃ Oxy selectivity achieved by Biswas et al. was 2.9%,¹¹ necessitating energy-intensive separations to isolate these products from the numerous other oxygenate and hydrocarbon species generated. One possible approach to improving selectivity involves coupling the plasma-activated reaction with suitable thermocatalysts. Catalysts can be placed within the plasma zone to interact with short-lived plasma species (such as ions, radicals, and vibrationally excited molecules) or downstream of the plasma zone where long-lived quasi-stable intermediates can reach the catalyst bed. Given the formation of ethylene and syngas in the plasma-activated reaction, the heterogeneous Rh₁Co₃/MCM-41 hydroformylation catalyst (as discussed in section 2.1.1) was initially considered to be a promising candidate to enhance C₃ Oxy production. However, Biswas et al. found the inclusion of Rh₁Co₃/MCM-41 only increased total oxygenate selectivity at early time scales (<100 min of time on stream) and had a negligible effect on the C₃ Oxy production (Figure 4d). Instead, the formation of formaldehyde and acids was favored, indicating that the presence of a catalyst altered the reaction pathways and product distribution.¹¹ The desired catalytic effect to enhance C₃ Oxy production may not have been observed if excess H₂ promoted ethylene hydrogenation more than the hydroformylation reaction. Additionally, the lack of catalytic effect may also be related to material shielding effects, where the plasma discharge may be unable to penetrate into the mesopores of the porous support (i.e., MCM-41) where the active catalyst metals were deposited.²⁸ The plasma-activated conversion of CO₂ and ethane was also investigated by Gomez-Ramirez et al.²⁹ While formaldehyde was the only oxygenate product detected, the incorporation of a vanadia/alumina catalyst dispersed on a BaTiO₃ ferroelectric was found to significantly enhance formaldehyde production, further demonstrating the possibilities for combining catalysts with plasma excitations to modify the product selectivity.

2.2. Upgrading CO₂ with H₂O

The electrocatalytic conversion of CO₂ to C₃ Oxys with water as the hydrogen source is an alternative to the ethane-based thermocatalytic and plasma-catalytic processes. Electrochemical processes are often advantageous because of their operation under ambient temperature and pressure, scalability, and ease of integration with renewable electricity sources. When it is employed for CO₂ conversion, water is often introduced in the form of a humidified CO₂ feed and/or an aqueous electrolyte solution (e.g., H₂O, KHCO₃, KOH) in low-temperature electrolysis. Unlike the temperature-gap limitations in the thermochemical CO₂ + ethane reactions, CO₂ can be electrochemically reduced to C₃ Oxys in one step. However, due to the large overpotentials and multitude of competing reactions, the direct CO₂ reduction to 1-propanol in flow cells has only been demonstrated with very low selectivities, limited to Faradaic efficiencies (FEs) of <5%.^30,31 Furthermore, the separation of C₃ Oxys from other dilute liquid products (such as formic acid, acetic acid, and ethanol) within an aqueous electrolyte is challenging and energy-intensive. Instead, a two-step tandem approach can break down the process into simpler components and improve the overall energy efficiency and C₃ Oxy production rate. In this section, we discuss the tandem electrocatalytic–thermocatalytic and electrocatalytic–electrocatalytic schemes for CO₂ conversion to C₃ Oxys using water as the hydrogen source.

2.2.1. Electrocatalytic–Thermocatalytic Scheme (EC-TC)

In the EC-TC tandem reaction scheme, the first reactor step of the TC-TC process involving CO₂-ODHE and DRE reactions (eqs R1 and R2, respectively) is replaced with the electrochemical CO₂ reduction reaction (CO₂RR) to produce C₂H₄ and CO (eqs R4 and R5, respectively) and the hydrogen evolution reaction (HER) to produce H₂ (eq R6). The products of these electrochemical reactions, C₂H₄ and syngas, are then directly fed as the reactants to a downstream thermochemical heterogeneous hydroformylation reactor (eq R3), as in the second step of the TC-TC scheme.

R4

R5

R6

These three reactions are typically considered to be problematic competing reactions, and most electrocatalytic research has traditionally focused on minimizing side reactions in favor of producing a single desired species. In particular, suppressing excess H₂ formation via the HER remains a challenge due to the presence of water-based electrolytes and the highly negative overpotentials required to drive the CO₂RR. Despite this issue, several recent studies have demonstrated the selective production of C₂H₄, with FEs as high as 83%.³²⁻³⁴ In contrast, the EC-TC approach seeks to leverage these competing reactions in order to coproduce C₂H₄, CO, and H₂ to be used directly as the feed for the subsequent thermocatalytic hydroformylation reaction.

Due to its intermediate binding energies of H* and CO*, copper (Cu) is the only monometallic catalyst that is able to electrochemically convert CO₂ into C₂H₄, as well as any other product with more than two electron transfers.^35,36 Extensive research has been conducted on the CO₂RR over Cu electrocatalysts, including the effects of surface structure, morphology, composition, electrolyte, pH, and cell design, as summarized in a comprehensive review by Nitopi et al.³⁷ In a demonstration of the tandem EC-TC system, Biswas et al. used Cu nanoparticle electrocatalysts with different oxidation states (i.e., Cu and oxide-derived Cu catalysts), as well as modifications to the gas diffusion layer (GDL) hydrophobicity (i.e., fluorinated ethylene propylene (FEP) reinforced carbon) in a zero-gap, vapor-fed flow electrolyzer with a Sustainion anion exchange membrane.¹² Relative to the benchmark Cu on carbon GDL, the oxide-derived Cu and FEP-reinforced carbon GDL cathodes enhanced the ethylene FE from 50% to 60% and lowered the H₂ FE from 27% to 19% at a cell current of −220 mA cm^–2, as shown in Figure 5a. The FE values describe the distribution of charge transferred and are not directly comparable to the thermocatalytic definition of selectivity; therefore, the relative molar production rates are also presented in Figure 5b. The C₂H₄/CO/H₂ product ratios obtained from the electrolyzer were 1/1.1/3.2 (for Cu/C), 1/0.5/2.2 (for oxide-derived Cu/C), and 1/1.3/2.0 (for Cu/FEP-reinforced C), which were then fed to the thermocatalytic hydroformylation reactor.

Figure 5.

EC-TC tandem performance. (a) Faradaic efficiencies for CO₂RR on commercial Cu on carbon, oxide-derived Cu on carbon, and Cu on reinforced carbon GDL at −220 mA cm^–2. (b) Electrochemical product selectivity represented in terms of relative molar ratios. (c) CO₂-based selectivities and (d) yields of products in the tandem configuration for Cu/C, oxide-derived Cu/C, and Cu/reinforced C cathodes at −220 mA cm^–2 in combination with 160 and 200 °C hydroformylation temperatures. Reprinted with permission from Biswas et al.¹² Copyright 2022, American Chemical Society.

Biswas et al. used the same Rh₁Co₃/MCM-41 hydroformylation catalyst in the EC-TC system as was previously employed in the TC-TC and P-TC schemes.¹² In the directly coupled tandem configuration, the highest C₃ Oxy selectivities obtained were 18.4% for the oxide-derived Cu/C electrode coupled with a 160 °C hydroformylation temperature and 18.0% for the Cu/FEP-reinforced C electrode coupled with a 200 °C hydroformylation temperature (Figure 5c), which corresponded to C₃ Oxy yields of 2.7% and 2.5%, respectively (Figure 5d). The remainder of the carbon-containing products consisted primarily of CO and ethylene (CO₂RR products that were unconverted in the thermocatalytic reactor) or ethane (undesired product from C₂H₄ hydrogenation), which further highlights the need for inhibiting the HER in the upstream CO₂ electrolyzer and developing more active and selective hydroformylation thermocatalysts.

2.2.2. Electrocatalytic–Electrocatalytic Scheme (EC-EC)

As discussed above, the direct electrochemical conversion of CO₂ to multicarbon products (such as ethylene, ethanol, and 1-propanol) remains challenging due to low selectivities and high overpotentials. It has been well-established, however, that CO is the key reaction intermediate during the CO₂RR to C₂₊ species over Cu electrodes.³⁸ Therefore, feeding pure CO can enable an increased surface coverage of CO and C–C coupling and consequently enhance C₂₊ product formation.³⁹ As a result, CO₂ electrolysis can be conducted in a tandem EC-EC scheme, in which CO₂ is first electrochemically reduced to CO (eq R5), followed by the CO reduction reaction (CORR) to propanol, as shown in eq R9.

R9

For CO₂-to-CO electrolysis, a variety of catalytic materials, including precious and earth-abundant metals, transition-metal chalcogenides, and carbon-based catalysts, have effectively converted CO₂ to CO, with FEs reaching >95%.⁴⁰ Additionally, several studies have reported CO electrolysis to C₂₊ products over Cu-based catalysts with high FEs (>75%) at current densities of up to 1 A cm^–2,⁴¹⁻⁴³ representing a significant performance improvement over the CO₂RR.⁴⁴⁻⁴⁶ At present, tandem CO₂/CO electrolysis strategies have primarily demonstrated enhanced production of acetate, ethylene, and ethanol,⁴⁷⁻⁵⁰ including the development of tandem catalyst architectures within a single device.⁵¹⁻⁵³ However, the focus of the current Perspective is mainly on the implementation of a two-step EC-EC system that produced C₃ Oxys.

Romero Cuellar et al. used two flow electrolyzers in series, with an Ag gas diffusion electrode in the first electrolyzer and Cu nanoparticles on a carbon-based gas diffusion electrode in the second electrolyzer.¹³ Although >90% FE to CO was obtained in the first electrolyzer, the outlet contained large amounts of unconverted CO₂. Therefore, the CO₂ feed rate and applied voltage had to be carefully controlled in order to maximize CO₂ conversion without compromising CO selectivity relative to the competing HER. The second CO electrolyzer was then tested with varying CO/CO₂ feed ratios to simulate the outlet from the upstream reaction. As shown in Figure 6a, a larger CO/CO₂ feed ratio resulted in an enhanced production of 1-propanol and C₂ products, again illustrating the clear advantage of the CORR to multicarbon products versus the CO₂RR.

Figure 6.

EC-EC tandem performance. (a) Cumulative Faradaic efficiencies at a current density of −300 mA cm^–2 from single-step electrolysis with pure CO, an 80% CO/20% CO₂ mixture, and pure CO₂ compared to the two-step tandem configuration. (b) Faradaic efficiencies in the tandem configuration at −470 mA cm^–2 with and without the NaOH absorber. Reprinted with permission from Romero Cuellar et al.¹³ Copyright 2020, Elsevier.

Due to the observed benefits of feeding pure CO to the second electrolyzer, Romero Cuellar et al. also integrated a CO₂ absorption column filled with 5 M NaOH between the two electrolyzers.¹³ While the NaOH absorber purified the CO feed and enhanced the overall selectivity to C₂₊ products (Figure 6b), it also sacrificed CO₂ conversion and would require additional energy input to recover the captured CO₂. Improving CO₂ single-pass conversion would minimize the energy costs of separating unreacted CO₂ and potentially eliminate the need for any intermediate separation altogether. With the coupled electrolyzers (−100 mA cm^–2 in the first cell, −200 mA cm^–2 in the second cell) along with the NaOH separation step, Romero Cuellar et al. achieved a combined FE of 62% for C₂₊ products, including 7.5% FE to 1-propanol.¹³ It is important to note that this work was not specifically aimed at maximizing 1-propanol production. Therefore, more targeted electrocatalysts that have previously been shown to favor the CORR to 1-propanol could improve the overall tandem performance for C₃ Oxy production.⁵⁴⁻⁵⁶

3. Comparison of Energy Costs and Prospects for Net CO₂ Reduction

Valuable C₃ Oxys may be produced more sustainably using the alternative ambient-pressure tandem processes described above, by reacting CO₂ either with surplus ethane in underutilized shale gas fractions via TC-TC or P-TC schemes or with water via EC-TC and EC-EC schemes. Evaluating these alternative approaches requires comparing their energy costs and prospects for achieving a net consumption of CO₂, but it is difficult to directly compare the various tandem strategies due to discrepancies in the as-published results. For instance, electrocatalytic studies typically report Faradaic efficiencies and partial current densities (and infrequently CO₂ conversions), while thermocatalytic studies report conversions, selectivities, and yields. Moreover, even within the thermocatalysis literature, there are often inconsistencies in how these metrics are defined. Comparing across these metrics is even more challenging in tandem reaction schemes, in which different types of reactors are combined in series. Therefore, in order to provide a fair comparison, we calculate the C₃ Oxy yield on a CO₂ basis using the equation

1

where F_C3Oxy,out is the total C₃ Oxy formation rate and F_CO2,in is the CO₂ feed rate. The number of carbon atoms from CO₂ in C₃ Oxys equals 3 for the CO₂ + H₂O schemes (since CO₂ is the only carbon source) and 1 for the CO₂ + ethane schemes (since the remaining 2 carbon atoms are derived from ethane). We note that the exact number of carbon atoms from CO₂ may be different for the plasma-activated reaction due to more complex reaction pathways. The tandem C₃ Oxy production rates are either directly taken from the results cited above or back-calculated from reported conversions and selectivities (or currents and FEs), as summarized in Table S1. In addition, the energy input per mole of C₃ Oxys produced is calculated based on the driving force of each reaction type: i.e., temperature for thermochemistry and electrical energy for electrochemistry and plasma.

Figure 7a shows a comparison of the yields and energy costs of the laboratory-scale tandem schemes reported in the literature (see the Supporting Information for calculation details). In this comparison, we include the state-of-the-art demonstrated tandem results with the largest C₃ Oxy yields and lowest energy costs per mole of C₃ Oxy produced. The TC-TC scheme demonstrated the highest C₃ Oxy yield of 6.3%, while the P-TC scheme had the lowest yield of 0.4%. The higher yields of the TC-TC approach are consistent with the issues associated with CO₂ loss (and hence, low conversion) in the electrochemical-based systems and the low selectivity of the plasma-activated reaction. The EC-TC scheme has the lowest energy cost per mole of C₃ Oxy, as a result of replacing the high-temperature CO₂-ODHE/DRE thermocatalytic reactor step with CO₂ electrolysis at ambient temperature. The EC-TC system also utilizes a low-temperature hydroformylation reactor in the second step as opposed to the more difficult electrochemical conversion of CO/CO₂ to C₃ Oxy products. The energy cost in the P-TC scheme is especially high, due to the large power input in combination with the low C₃ Oxy production rate.

Figure 7.

Comparison of energy costs and net CO₂ emissions of the alternative ambient-pressure tandem schemes for the production of C₃ oxygenated hydrocarbons (C₃ Oxys). (a) C₃ Oxy yields (left) and energy costs per mole of C₃ Oxy produced (right) for each of the tandem schemes. The red dashed line represents the energy cost for conventional 1-propanol production. (b) Net CO₂ emissions as a function of CO₂ emissions per unit energy based on demonstrated laboratory-scale results. (c) Theoretical net CO₂ emissions calculated for the ideal scenarios based on minimum energy requirements. In (b) and (c), the vertical black dashed line represents the average CO₂ emissions associated with U.S. grid electricity and the red data point represents the CO₂ footprint for conventional propanol production.

Any viable CO₂ utilization technology must consume more CO₂ than the process emits, or at a minimum, emit less CO₂ than the conventional process it replaces. Therefore, it is essential to consider the CO₂ emissions associated with these tandem processes, where the energy input typically accounts for the majority of the CO₂ footprint. Figure 7b plots the net CO₂ emissions per mole of C₃ Oxy produced as a function of the CO₂ emissions per unit energy for the four tandem reaction strategies (see the Supporting Information for more details regarding emissions calculations). Based on the currently demonstrated laboratory-scale results, the net CO₂ emissions are largely positive and in the trend of EC-TC < TC-TC < EC-EC < P-TC, following the trends in energy cost per mole of C₃ Oxy produced. As we shift from the average CO₂ emissions associated with the U.S. grid electricity (∼0.39 kg-CO₂/kWh)⁵⁷ to less carbon-intensive energy sources (i.e., moving leftwards on the x axis), the overall net CO₂ emissions can be reduced. Unsurprisingly, net negative emissions can only be realized for any of the laboratory-scale processes if a significant portion of the energy input is obtained from renewable sources (<0.06 kg-CO₂/kWh). Therefore, compatibility with renewable energy must be a key consideration in the potential commercial-scale implementation of any tandem CO₂ conversion strategy. With completely carbon-free energy (i.e., 0 kg-CO₂/kWh), the EC-EC and EC-TC schemes can theoretically reduce 3 mol of CO₂ into 1 mol of C₃ Oxy, while the TC-TC and P-TC schemes only reduce 1 mol of CO₂ into 1 mol of C₃ Oxy due to the carbon contribution from ethane.

Based on Figure 7a,b using the currently available experimental results, all of the tandem processes have a greater energy cost and CO₂ footprint than those of the existing 1-propanol production process (energy cost of ∼1.32 kWh/mol-C₃ Oxy; CO₂ footprint of ∼4.23 mol-CO₂/mol-C₃ Oxy).⁵⁸ Therefore, the conventional 1-propanol production process is unlikely to be replaced by the state-of-the-art tandem schemes without an improvement in catalytic performance and/or utilization of low-carbon energy sources. It should also be noted that the calculations in Figure 7a,b were performed based on currently demonstrated laboratory-scale results, since none of the tandem processes have been implemented at scale and therefore reliable parameters for plant-level process simulation are unavailable. As a result, the analysis for the tandem schemes does not consider the CO₂ emissions associated with obtaining feedstocks: namely, CO₂ capture, ethane extraction and separation, and water purification. Energy savings via heat integration were also not accounted for in the energy cost calculations (e.g., heat integrating the first endothermic reactor and second exothermic reactor in the TC-TC scheme).

Moreover, downstream separation processes were not considered, although these may ultimately constitute a large fraction of the process energy consumption depending on the final C₃ Oxy concentration supplied by the tandem reactors. Among the four tandem processes, the TC-TC and EC-TC tandem schemes would likely have significant separation advantages due to the ease of separating targeted liquid oxygenates from the gaseous outlet stream of a thermocatalytic reactor. Both TC-TC and EC-TC should have similar separations following the hydroformylation reactor; however, the EC-TC scheme may have additional water content originating from the electrolyte in the upstream electrolyzer. In contrast, P-TC (which produces a wide range of liquid products) and EC-EC (which requires liquid–liquid separation of dilute products from an aqueous electrolyte) would likely involve much more energy-intensive separations.

Although the results presented in this Perspective serve as a useful proof of concept, they remain far from optimized. Therefore, in Figure 7c, we include a comparison of net CO₂ emissions as calculated from theoretical limits, thus representing an upper bound for performance. As given in Table S2, the minimum energy requirements were determined from the enthalpy of reactions for thermochemistry, reversible cell potentials for electrochemistry, and vibrational excitation energies for plasma chemistry and by assuming the absence of any competing reactions. More details regarding the theoretical minimum energy calculations are provided in the Supporting Information. Although these assumptions are highly optimistic, this method allows for a comparison of the best-case scenarios. Figure 7c shows a trend of P-TC > EC-EC > EC-TC > TC-TC based on the crossover points for reaching net negative emissions. Even though the P-TC and TC-TC pathways have a lower thermodynamic energy requirement, the electrocatalytic-based tandem schemes have the advantage of converting 3 mol of CO₂ into C₃ Oxys. More importantly, for all four tandem strategies, there are significant opportunities for reducing energy input and CO₂ emissions relative to the currently demonstrated performance. Future advances in catalyst development and reactor design will enable progress toward these theoretical limits. Engineering optimizations, such as heat integration and reactant recycling, can also help minimize energy costs and thereby reduce CO₂ emissions.

4. Challenges and Opportunities

In this Perspective, we present four tandem reaction strategies, TC-TC, P-TC, EC-TC, and EC-EC, to produce C₃ Oxys using CO₂ as a reactant at atmospheric pressure. The tandem catalytic strategies may be a promising approach to reducing atmospheric CO₂ concentrations while simultaneously producing high-value C₃ oxygenates. However, these tandem schemes have been reported in the literature only as proof-of-concept demonstrations; thus, opportunities to improve and optimize these processes are significant. Currently, these systems operate far from the theoretical limits shown in Figure 7c and will require significant improvements before becoming competitive with existing C₃ Oxy synthesis methods. We summarize the main challenges and opportunities for tandem reactions in this section.

4.1. Thermocatalytic Schemes

A major challenge for the tandem TC-TC and EC-TC schemes is the development of efficient heterogeneous hydroformylation catalysts to increase the C₃ Oxy selectivity relative to ethane, which is produced via the competitive ethylene hydrogenation reaction. As shown in Figures 3b and 5c,d, ethane is often observed to be the primary product. This can be attributed to the common C₂H₅* intermediate involved in the competing reaction pathways, which can more easily undergo hydrogenation than CO insertion—the typical rate-determining step of hydroformylation. Moreover, strong CO binding leads to significant competition with ethylene over single active sites, as suggested by the negative reaction order with respect to the CO partial pressure.²² Thus, a selective hydroformylation catalyst should favor ethylene adsorption and CO insertion while making H₂ activation kinetically slower, which can be achieved via bifunctional active sites.²⁴ For example, the recently reported Rh-WO_x pair sites by Ro et al. were demonstrated to (1) facilitate W⁶⁺ reduction to form the active Rh–W site, (2) alleviate the typical CO poisoning effect by transferring adsorbed C₂H₄ from W to Rh, and (3) promote CO insertion and propanal formation via H₂ activation at the Rh–W interface to form a bridging hydride that facilitates adsorption of a third CO on Rh before acylation.²⁴ The pair sites exhibited extremely high selectivity (>95%) and production rate (0.1 g cm^–3 h^–1) of propanal in heterogeneous ethylene hydroformylation. Therefore, the employment of bifunctional active sites with facile site cooperation during the catalytic cycle is an important approach to circumventing the undesired ethylene hydrogenation. In a practical system, the ethane byproduct can be easily separated from the liquid C₃ Oxy products and recycled as a reactant to the first reactor, reducing the natural-gas-derived ethane feed requirement and increasing the overall conversion efficiency.

4.2. Electrocatalytic Schemes

Suppressing the overproduction of H₂ via the HER remains a key challenge in aqueous-based electrochemical systems. This is especially relevant in the EC-TC scheme, where the excessive production of H₂ (e.g., H₂/C₂H₄ ratio >2 as shown in Figure 5c) promotes undesired ethylene hydrogenation over hydroformylation in the downstream thermocatalytic reactor. Various strategies, including the use of nanostructured catalysts, highly alkaline electrolytes, ionic liquids, and hydrophobic additives, have shown promise in inhibiting HER kinetics; however, further systematic studies are required to tune the relative product ratios. Alternatively, high-temperature solid oxide electrolyzer cells (SOECs) may be employed to avoid liquid electrolytes altogether, where cofeeding varying amounts of steam can potentially enable better control over the H₂ production rates.

The low product yields in the EC-TC system were also attributed to large CO₂ losses to carbonate formation and crossover within the electrolyzer, representing another significant challenge for low-temperature CO₂ electrolysis.^59,60 Under the highly alkaline conditions near the electrode surface (due to the consumption of protons and the formation of equivalent amounts of OH^– in CO₂RR and HER), CO₂ can rapidly react with OH^– to produce (bi)carbonates, as shown in eqs R7 and R8.

R7

R8

These (bi)carbonates either accumulate within the electrolyte or migrate across the anion exchange membrane, where they are protonated and released as CO₂. This issue is further exacerbated by the large overpotentials required for the CO₂RR to C₂H₄ that increase the OH^– concentration, as well as the excess CO₂ feed needed to support high current densities. Therefore, novel electrolyzer cell designs or strategies that can efficiently regenerate and recycle CO₂ should be employed to improve the overall carbon conversion to C₃ Oxys in the EC-TC tandem scheme.

While CO₂ losses also affect the performance of the CO₂RR step in the EC-EC scheme, these losses occur to a lesser extent than in the EC-TC scheme. The maximum carbon efficiency is 25% for CO₂-to-C₂H₄ electrolysis (three CO₂ molecules converted to carbonate per CO₂ to C₂H₄), whereas CO₂-to-CO has a maximum carbon efficiency of 50% (one CO₂ to carbonate per CO₂ to CO).⁵⁹ Despite the relative advantage of converting CO₂ to CO instead of C₂H₄, state-of-the-art low-temperature CO₂-to-CO flow electrolyzers have only demonstrated single-pass conversions of up to 33%.⁶¹ Alternative cell configurations, such as the use of bipolar membranes (BPMs)⁶²⁻⁶⁴ or SOECs,^65,66 can potentially circumvent issues with carbonate formation; however, their respective tradeoffs must also be considered (i.e., more negative overpotentials for water dissociation and lower CO FEs in BPMs and elevated operating temperatures in SOECs). Unlike CO₂RR, CO electrolysis can avoid CO₂ losses to carbonate and has demonstrated higher single-pass conversions.⁶⁷ However, previous experimental trends have shown 1-propanol selectivities to be inversely related to partial currents, with FEs dropping below 10% at partial current densities above 40 mA cm^–2.⁴⁴ More recently, though, Wang et al. reported 36% FE to 1-propanol at 300 mA cm^–2 and 85% CO conversion with Ag–Ru–Cu catalysts, demonstrating the possibility for more active and selective electrocatalysts for C₃ Oxy production via the EC-EC scheme.⁶⁸

4.3. Plasma-Catalytic Scheme

While the P-TC scheme can accomplish one-pot C₃ Oxys production, the high energy consumption and wide product distribution present significant challenges. Additional plasma-catalysis studies are essential to improve the selective conversion of CO₂ and ethane to C₃ Oxys. Although well-established protocols for product prediction and catalyst screening do not currently exist, inspiration can be drawn from the plasma-activated reaction between CO₂ and methane, in which DBD reactors packed with catalysts (e.g., reducible metal oxides and copper) and high-dielectric-constant additives (e.g., BaTiO₃) have been found to enhance the production of methanol and other oxygenates.⁶⁹ Furthermore, the use of plasma kinetic modeling can provide valuable insights into the dominant reaction pathways. For instance, Biswas et al. used plasma kinetic modeling, in combination with ¹²CO₂/¹³CO₂ isotopic labeling experiments, to reveal a CO₂ + ethane plasma reaction mechanism that proceeded via oxidation of ethane-derived species by oxygen from CO₂, which is different from the thermocatalytic alcohol formation pathway that involves either CO₂ hydrogenation or CO insertion.¹¹ Therefore, catalysts that promote conventional thermochemical reaction pathways to oxygenates may not necessarily be the ideal candidates for the corresponding plasma-activated reactions. The effects of various catalysts and support materials on the plasma, and vice versa, should be further studied. Approaches that aim to tune both the plasma conditions as well as the catalytic surface reactions will likely be necessary to achieve sufficient control over product selectivity. However, for applications that require high-purity C₃ Oxys, plasma-catalytic schemes are unlikely to be competitive due to the energy-intensive postreaction separation steps that would be required to separate products with similar physicochemical properties, such as mixed alcohols or liquids with similar boiling points. Alternatively, plasma-catalytic schemes may be more appropriate for applications that can tolerate a wider distribution of products among the liquid fraction, such as blended alcohol or oxygenate product streams for fuel applications.

4.4. Future Directions

Although many research efforts have been dedicated to the individual reaction steps, more systematic studies should be conducted with a tandem configuration in mind. In some cases, this will require new approaches to catalyst and reactor design. For example, as opposed to simply maximizing selectivity toward a single product, the tandem processes must control product formation while considering the amounts of unconverted reactants. Improving the tandem processes also requires tuning key competing reactions, i.e., CO₂-ODHE vs DRE and hydroformylation vs hydrogenation in the TC-based cases, CO₂RR to C₂H₄ and CO vs HER in the EC-based cases, and C–C vs C–O recombination reactions involving C₂H₅ in the P-TC case. The downstream reactors and catalysts must also be designed for compatibility with the outlet stream of the first reactor: for example, considering the potential effects of unconverted reactants, undesired byproducts, and moisture (from CO₂-ODHE or CO₂RR) or plasma–catalyst interactions (in the P-TC scheme).

Furthermore, the CO₂-based tandem strategy for C₃ Oxys can be extended to a range of other multicarbon, value-added liquid products. The effluent from the first reactors is essentially a mixture of unconverted reactants (e.g., CO₂, C₂H₆, and/or H₂O) and products (e.g., C₂H₄, CO, H₂, and/or H₂O); thus, there remain many opportunities for incorporating other reaction chemistries in the downstream reactors. For example, carboxylic acids can be obtained via a direct or CO₂-mediated hydrocarboxylation process (i.e., C₂H₄ + CO + H₂O → CH₃CH₂COOH or C₂H₄ + CO₂ + H₂ → CH₃CH₂COOH). Despite the similar thermodynamic characteristics with hydroformylation, successful demonstrations of heterogeneous hydrocarboxylation have been rarely reported. Future efforts should focus on mechanistic studies for the kinetically relevant step(s) and, in turn, develop proof-of-concept catalysts. Aromatics can also be obtained via a downstream aromatization reaction of ethylene (i.e., 3 C₂H₄ → C₆H₆ + 3 H₂) by virtue of the synergistic Brønsted acid–Lewis acid functions of zeolite-based catalysts. Such types of catalysts should be advanced with regard to coking resistance and hydrothermal stability for the tandem processes. Electrochemical CO reduction to acetate, ethanol, and other C₂₊ oxygenates also remains a promising approach to tandem EC-EC CO₂ electrolysis; however, the desired liquid products must be generated with much higher concentrations and purities to reduce the energy consumption of product separation. Equally important to the two-step tandem strategies, it is worthwhile to investigate the coupling of thermo/electro/plasma/bio/photochemical processes within a single reactor to bridge the so-called “thermodynamic gap” and achieve a one-pot conversion process. We hope that this Perspective will stimulate future research efforts exploring the vast possibilities of tandem CO₂ conversion strategies.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacsau.2c00533.

• Calculations of C₃ oxygenate yields, practical and theoretical energy costs, and net CO₂ emissions (PDF)

Author Contributions

A.N.B., L.R.W., and Z.X. contributed equally.

The authors declare no competing financial interest.