Co 2 fixation into carbon nanofibers using electrochemical-thermochemical tandem catalysis

Abstract

A method, comprising electrolyzing a CO2 input and water so as to form a first product comprising CO and H2, the electrolyzing optionally being performed over a Pd / C catalyst or a catalyst that comprises any one or more of gold, silver, iron, cobalt, nickel, copper, or zinc; and thermochemically processing the first product so as to give rise to a second product that comprises carbon nanofibers or nanotubes. A system, comprising: a first reaction zone, the first reaction zone configured to receive CO2 input and water, and the first reaction zone configured for electrolysis of the CO2 input and water to evolve a product that comprises CO; a second reaction zone, the second reaction zone configured to receive a product from the first reaction zone, the second reaction zone configured to support at least one of the Boudouard reaction (R1) and CO + H2 → C(s) + H2O (R2).

Description

CU24182 – 101879.000335 CO2FIXATION INTO CARBON NANOFIBERS USING ELECTROCHEMICAL-THERMOCHEMICAL TANDEM CATALYSIS RELATED APPLICATIONS

[0001] The present application claims priority to and the benefit of United States patent application no.63 / 610,499, “CO2Fixation Into Carbon Nanofibers Using Electrochemical-Thermochemical Tandem Catalysis,” filed December 15, 2023. All foregoing applications are incorporated herein by reference in their entireties for any and all purposes. GOVERNMENT RIGHTS

[0002] This invention was made with government support under DE-SC0012335, DE-FG02-13ER16381, DE-SC0012704, and DE-SC0012653 awarded by the U.S. Department of Energy, and 2036197 awarded by the National Science Foundation. The government has certain rights in the invention. TECHNICAL FIELD

[0003] The present disclosure relates to the field of materials science and to the field of carbon capture technologies. BACKGROUND

[0004] Carbon dioxide (CO2) fixation into value-added solid carbon such as carbon nanofiber (CNF) for longer-term storage represents a promising avenue for achieving net- negative carbon emissions. But directly converting CO2to CNF via thermocatalytic approaches faces thermodynamic constraints, and electrocatalytic methods typically lead to amorphous carbon with limited yields and / or require energy-intensive conditions (>720 °C). Accordingly, there is a long-felt need for methods for carbon dioxide fixation into solid carbon.CU24182 – 101879.000335 SUMMARY

[0005] In meeting the described long-felt needs, the present disclosure provides a method, comprising: electrolyzing a CO2 input and water so as to form a first product comprising CO and H2, the electrolyzing optionally being performed over a Pd / C catalyst or a catalyst that comprises any one or more of gold, silver, iron, cobalt, nickel, copper, or zinc; and thermochemically processing the first product so as to give rise to a second product that comprises carbon nanofibers or nanotubes.

[0006] Also provided is a system, the system comprising: a first reaction zone, the first reaction zone configured to receive a CO2 input and water, and the first reaction zone configured to support electrolysis of the CO2input and water to evolve a product that comprises CO; a second reaction zone, the second reaction zone configured to receive a product from the first reaction zone, the second reaction zone configured to support at least one of the Boudouard reaction (R1) and CO + H2 ^ C(s) + H2O (R2). The at least one of the Boudouard reaction (R1) and CO + H2^ C(s) + H2O (R2) can give rise to a product that comprises carbon nanofibers or

[0007] Also provided is a system, the system configured to perform the method of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In the drawings, which are not necessarily drawn to scale, like numerals may describe similar components in different views. Like numerals having different letter suffixes may represent different instances of similar components. The drawings illustrate generally, by way of example, but not by way of limitation, various aspects discussed in the present document. In the drawings:

[0009] FIGs.1A-1J. Feasibility and demonstration of CNF formation from the CO and H2 reaction. FIG.1A: Gibbs free energy change of primary reactions and potential side- reactions as a function of temperature. FIG.1B: Temperature dependence of equilibrium conversion of CO for individual R1 (2CO=C(s)+CO2) and R2 (CO+H2=C(s)+H2O) reactions and of CO and H2in the presence of both R1 and R2. FIG.1C: The corresponding equilibrium distribution of reactants (CO and H2), products (C(s), CO2, H2O), and balance gas (N2) for the right panel of FIG.1B. FIGs.1D-1F: Effects of Fe / Co molar ratios, CO / H2 / N2flow rates, and reaction temperature on the catalyst weight gain after reaction,CU24182 – 101879.000335 respectively. FIG.1G: Conversion of CO and H2as well as carbon balance over Fe3Co6 / CeO2 along time on stream at 450 ℃. Approximately 100 mg of catalyst are used for each of the experiments shown in Figs. d-g. FIG.1H: Transmission electron microscopy (TEM) imaging of CNF over the spent Fe3Co6 / CeO2 catalyst. FIG.1I: High-resolution imaging of the tip of the resulting CNF. FIG.1J: Scanning transmission electron microscopy (STEM) high-angle annular dark field (HAADF) imaging and (energy-dispersive X-ray spectroscopy) EDS mapping of C, Ce, Fe, and Co over the spent Fe3Co6 / CeO2catalyst. The scale bar in each panel of FIG.1j represents 8 nm.

[0010] FIGs.2A-2G. In situ structural characterizations of CeO2 supported FeCo- based catalysts. FIG.2A: In situ XRD patterns of FexCoy / CeO2catalysts (x / y=0 / 9, 3 / 6, 4.5 / 4.5, 6 / 3, and 9 / 0) under the reduction (H2 only) and reaction (CO and H2 at 450 ℃) conditions. FIG.2B: The Fe K-edge XANES profile of Fe3Co6 / CeO2under the fresh, reduction, and spent conditions as well as the reference standards (bcc Fe foil, FeO, Fe3O4, and Fe2O3). FIGs.2C-2D: The k2-weighted EXAFS of Fe K-edge in k space and R space, respectively, under the reduction and spent conditions as well as the bcc Fe foil. FIGS.2E- 2G: The Co K-edge features of Fe3Co6 / CeO2corresponding to the conditions in FIGs.2B- 2D: Note: Fresh refers to exposure to N2 at 25℃; Red. represents reduction under H2 at 500 ℃; Spent indicates the catalyst was exposed to the reaction stream (CO and H2) at 450 ℃ for about 5 h and then cooled down to room temperature for the XAFS measurement. More details about the peak assignment of XRD peaks can be found in FIG.26.

[0011] FIGs.3A-3E. Theoretical studies on the stability and reactivity of iron-cobalt catalysts. FIG.3A: Ab initio thermodynamic analysis for the stability of bulk phases (Fe, Co, FeCo, Fe3C, Co3C, Fe2CoC) under the reaction conditions (PCO and PH2: 0.1-1.0 bar; T: 325- 525 °C). The corresponding chemical potential changes of C species at the reaction environments (PCO and PH2: 0.1-1.0 bar; T: 325-525 °C) are shown in FIG.32. FIG.3B: DFT-calculated potential energy diagram for CO adsorption (*CO), dissociation (*C + *O) on Fe(110), Co(0001), and FeCo(110) surfaces. Inset: DFT-optimized structures for CO adsorption (*CO), dissociation (*C + *O) and corresponding transition state (TS) on FeCo(110) surfaces. c DFT-calculated potential energy diagram for C-C bond formation (*C + *C ^ *C-C) on Fe(110), Co(0001), and FeCo(110) surfaces. FIG.3D: DFT-calculated potential energy diagram for surface *O species removal in the form of CO2(*CO + *O ^ *CO2^ CO2) on Fe(110), Co(0001), and FeCo(110) surfaces. FIG.3E: DFT-calculatedCU24182 – 101879.000335 potential energy diagram for surface *O hydrogenation (*O + *H ^ *OH) on Fe(110), Co(0001) and FeCo(110) surfaces, followed by *H2O formation via hydrogenation and removal in the form of H2O. Yellow: Fe; Blue: Co; Brown: C; Red: O.

[0012] FIGs.4A-4F. Electrochemical-thermochemical tandem process for producing CNF and renewable H2. FIG.4A: Scheme of the electrochemical-thermochemical tandem process. b-d CO2 conversion and carbon balance (FIG.4B), molar flow rates of CO2, CO, H2, steam, and CH4during the tandem process (FIG.4C), and cell voltage (FIG.4D) at - 150 mA cm-2as a function of time. FIG.4E: Picture showing the fresh and spent Fe3Co6 / CeO2catalysts for the tandem process. FIG.4F: TEM image of the resulting CNF after the ~11 h of tandem process. Reaction conditions: 1 atm, electrolyzer (10%Pd / C, 1 mg cm-2, CO2 / N2=10 / 10 ml min-1, -150 mA cm-2, 25 ℃), and thermal reactor (Fe3Co6 / CeO2, 200 mg, 450 ℃). Notes: Conv._total and Conv._products refer to the CO2 conversion based on total CO2consumption (Xtotal) and carbon-containing gaseous products (Xproducts), respectively; EC and TC represent electrochemical and thermochemical reactors, respectively; given safety concerns, the reaction is stopped at 11 h because of the severe blockage of the thermal reactor caused by CNF growth.

[0013] Figures 5A-5C. Effect of Fe / Co ratios on catalytic performance of different CeO2-supported catalysts.

[0014] FIG.5A: Conversion of CO and H2 as well as carbon balance; FIG.5B: Selectivity in gaseous products; FIG.5C: Selectivity in total products under different reaction temperatures. The panels from top to bottom of each column refer to the results obtained with the Fe / Co ratios of 9 / 0, 6 / 3, 4.5 / 4.5, 3 / 6, and 0 / 9, respectively. Other reaction conditions: 100 mg of catalyst, CO / H2 / N2=15 / 15 / 20 ml min-1, reduction temperature (500 °C), reaction temperature (450 °C), 1 atm, time on stream of 4.5 h. Notes: The Selectivity in gaseous products was calculated based on carbon-containing products in gas phase; the Selectivity in total products was calculated based on all the carbon-containing products in both gas and solid phases.

[0015] FIGs.6A-6C. Effect of metal loadings on catalytic performance of FeCo catalysts.

[0016] FIG.6A: Conversion of CO and H2as well as carbon balance; FIG.6B: Selectivity in gaseous products; FIG.6C: Selectivity in total products under different reaction temperatures. The panels from top to bottom of each column refer to the resultsCU24182 – 101879.000335 obtained with the Fe3Co6 / CeO2and Fe6Co12 / CeO2catalysts, respectively. Other reaction conditions: 100 mg of catalyst, CO / H2 / N2=15 / 15 / 20 ml min-1, reduction temperature (500 °C), reaction temperature (450 °C), 1 atm, time on stream of 4.5 h. Notes: The Selectivity in gaseous products was calculated based on carbon-containing products in gas phase; the Selectivity in total products was calculated based on all the carbon-containing products in both gas and solid phases.

[0017] FIGs.7A-7D. TEM images of spent Fe3Co6 / CeO2after the reaction with CO only at 450 °C.

[0018] As shown in FIGs.7A-7B, no carbon nanofibers were observed over the Fe3Co6 / CeO2catalyst after reaction under the CO atmosphere at 450 °C; instead, the catalyst was encapsulated with thick carbon layers as illustrated by FIGs.7C-7D.

[0019] FIGs.8A-8C: Effect of CO / H2ratios on catalytic performance of Fe3Co6 / CeO2.

[0020] FIG.8A: Conversion of CO and H2as well as carbon balance. FIG.8B: Selectivity in gaseous products. FIG.8C: Selectivity in total products. The panels from top to bottom of each column refer to the results obtained at the CO / H2 / N2flow rates of 3 / 3 / 44, 3 / 15 / 32, 15 / 3 / 32, 15 / 15 / 20, and 15 / 30 / 5 ml min-1, respectively. Other reaction conditions: 100 mg of catalyst, reduction temperature (500 °C), reaction temperature (450 °C), 1 atm, time on stream of 4.5 h. Notes: The selectivity in gaseous products was calculated based on carbon- containing products in gas phase; the Selectivity in total products was calculated based on all the carbon-containing products in both gas and solid phases.

[0021] FIGs.9A-9C: Effect of flow rates on catalytic performance of Fe3Co6 / CeO2.

[0022] FIG.9A: Conversion of CO and H2 as well as carbon balance. FIG.9B: Selectivity in gaseous products. Fig.9C: Selectivity in total products. The panels from top to bottom of each column refer to the results obtained at the CO / H2 / N2 flow rates of 3 / 3 / 4, 9 / 9 / 18, and 15 / 15 / 20 ml min-1, respectively. Other reaction conditions: 100 mg of catalyst, reduction temperature (500 °C), reaction temperature (450 °C), 1 atm, time on stream of 4.5 h. Notes: The Selectivity in gaseous products was calculated based on carbon-containing products in gas phase; the Selectivity in total products was calculated based on all the carbon- containing products in both gas and solid phases. The initial point at t=0 h was not included for case of 3 / 3 / 4 ml min-1because the feeding mixture has not completely purged the gas line due to the relatively low flow rates.CU24182 – 101879.000335

[0023] FIGs.10A-10B. Temperature-programmed reduction profiles of different catalysts. FIG.10A: CeO2 supported catalyst with different Fe / Co ratios; FIG.10B: CeO2, SiO2, and γ-Al2O3supported Fe3Co6catalysts.

[0024] FIG.11: Effect of reduction temperature on catalyst weight gain over the spent Fe3Co6 / CeO2catalyst. Reaction conditions: 200 mg of catalyst, CO / H2 / N2=3 / 3 / 4 ml min-1, reduction temperature (450, 500, or 600 °C), 1 atm, time on stream of 11.5 h.

[0025] FIGs.12A-12C. Effect of reaction temperature on catalytic performance of Fe3Co6 / CeO2 catalysts under different conditions. FIG.12A: Conversion of CO and H2 as well as carbon balance at different reaction temperatures. FIG.12B: Selectivity in gaseous products at different reaction temperatures. FIG.12C: Selectivity in total products at different reaction temperatures. The panels from top to bottom of each column refer to the results obtained at reaction temperature of 370, 400, 450, 500, and 600 °C, respectively. Other reaction conditions: 100 mg of catalyst, CO / H2 / N2=3 / 3 / 4 ml min-1, reduction temperature (500 °C), 1 atm, time on stream of 4.5 h. Notes: The Selectivity in gaseous products was calculated based on carbon-containing products in gas phase; the Selectivity in total products was calculated based on all the carbon-containing products in both gas and solid phases. The initial point at t=0 h was not included because the feeding mixture has not completely purged the gas line due to the relatively low flow rates.

[0026] FIGs.13A-13B. Effect of time on stream on catalyst weight gain over the spent Fe3Co6 / CeO2catalyst. FIG.13A: Catalyst weight gain under the CO / H2 / N2flow rates of 3 / 3 / 4 ml min-1; FIG.13B: Catalyst weight gain under the CO / H2 / N2 flow rates of 15 / 15 / 20 ml min-1; Other reaction conditions: 100 mg of catalyst, CO / H2 / N2=15 / 15 / 20 ml min-1, reduction temperature (500 °C), reaction temperature (450 °C), 1 atm. Note: 5* in Figure 13a referred to that the catalyst loading was 200 mg.

[0027] FIG.14. Large batch synthesis (gram-level) of CNF with the Fe3Co6 / CeO2 catalyst. Reaction conditions: 1 g of fresh catalyst (sample #1) was loaded and subject to the reaction stream of 16.5 h, after which sample #2 (1.52 g, representing a 0.52 g increase) was obtained; Then, 0.5 g of sample #2 was loaded and run for another 20 h to obtain sample #3 (1.65 g, representing a 1.15 g increase); CO / H2 / N2=3 / 3 / 4 ml min-1, reduction temperature (500 °C), reaction temperature (450 °C), 1 atm. The total CWG after the 36.5 h of reaction is 1.67 g.CU24182 – 101879.000335

[0028] FIGs.15A-15C. 15 Effect of support on catalytic performance of FeCo bimetallic catalysts. FIG.15A: Conversion of CO and H2 as well as the carbon balance. FIG. 15B: Selectivity of gaseous products. FIG.15C: Selectivity in total products. The panels from the top to the bottom of each column refer to the results obtained as FeCo was supported over CeO2, γ-Al2O3, SiO2, and MCM-41, respectively. Other reaction conditions: 100 mg of catalyst, CO / H2 / N2=3 / 3 / 4 ml min-1, reduction temperature (450 °C), reaction temperature (400 °C), 1 atm, time on stream of 11.5 h. Notes: The Selectivity in gaseous products was calculated based on carbon-containing products in gas phase; the Selectivity in total products was calculated based on all the carbon-containing products in both gas and solid phases. The initial point at t=0 h was not included because the feeding mixture has not completely purged the gas line due to the relatively low flow rates.

[0029] FIG.16. Structures for CeO2-supported metal clusters before and after interaction with CO. Top and side views of DFT-optimized Fe4 / CeO2(111), Fe2Co2 / CeO2(111), and Co4 / CeO2(111) surfaces before interaction with CO (a), with CO chemically adsorbed (*CO) (b) and C-O bond scission (*C / *O) (c) with. Yellow: Fe; Blue: Co; Green: Ce; Red: O; Brown: C.

[0030] FIG.17 TEM. Images of the fresh Fe3Co6 / CeO2 catalyst.

[0031] FIGs.18A-18G. TEM and STEM images and EDS mapping over spent Fe3Co6 / CeO2 after the reaction of CO and H2 at 450 °C. FIGs.18A-18C: TEM images; FIG. 18D: EDS mapping; FIGs.18E-18G: bright field image, ADF image, and EDS mapping, respectively, for the same corresponding area.

[0032] FIGs.19A-19B. Raman spectra results of resulting solid carbon products as well as the commercial CNF reference standard. FIG.19A: Raman spectra between 900 and 2000 cm-1over different samples. FIG.19B: The relative crystallinity and purity of resulting solid carbon products over different samples compared to the commercial CNF. Note: For the comparison of different samples, all the peak intensities were normalized to the intensity of the D band. Reaction conditions: samples I and II (100 mg of catalyst, 450 °C, 1 atm, 4.5 h); samples III and IV (200 mg of catalyst, 450 °C, 1 atm, 11 h); sample V (commercial CNF).

[0033] FIG.20: EDS mappings over the spent Fe3Co6 / CeO2sample after the thermochemical experiment. Reaction conditions: 100 mg of catalyst, CO / H2 / N2=3 / 3 / 4 ml min-1, reduction temperature (500 °C), reaction temperature (450 °C), 1 atm, time on stream of about 5 h.CU24182 – 101879.000335

[0034] FIGs.21A-21F: STEM images of spent Co9 / CeO2after the reaction of CO and H2 at 450 °C. FIGs.21A and 21C: SEM images; FIGs.21B, D-F: ADF images. Notes: The SEM image of FIG.21A or FIG.21C and corresponding ADF image of FIG.21B (or FIG.21D) were collected in the same area.

[0035] As shown in FIG.18, carbon nanofibers were observed to be dominant over the spent Co9 / CeO2 sample with a tip growth mode as illustrated by panels e and f.

[0036] FIGs.22A-22H. STEM images of spent Fe9 / CeO2after the reaction of CO and H2 at 450 °C. FIGs.22A, FIG.22C, FIG.22E, and FIG.22G: STEM images; FIGs.22B, 22D, 22F, and 22H: ADF images in the same corresponding area of the SEM images.

[0037] Over the spent Fe9 / CeO2sample, as shown in FIG.19, the catalyst was predominantly encapsulated by amorphous carbon species (e.g., FIG.22G and FIG.22H); carbon nanofibers were rarely observed.

[0038] FIG.23. In situ XRD patterns of Fe3Co6 / SiO2 under different conditions. The catalyst was heated to 500 °C in 50 vol% H2in N2(5 / 5 ml min-1) and held for 30 min. Afterwards, the sample was cooled down to 450 °C in 5 min before being exposed to the reaction stream (CO / H2 / N2=3 / 3 / 4 ml min-1) at 450 °C. Each scan took 24 s.

[0039] FIGs.24A-24B. Ex situ XANES results of Fe and Co K-edges of the spent Fe3Co6 / CeO2catalyst after being spent under different reaction streams. Reaction conditions: 100 mg of catalyst, reduction temperature (500 °C), reaction temperature (450 °C), CO / H2 / N2=3 / 15 / 32, 15 / 3 / 32, 15 / 15 / 20, and 15 / 30 / 5 ml min-1, 1 atm, time on stream of 4.5 h.

[0040] In FIG.24A, the Fe K-edge XANES spectra of the spent samples resembled that of the Fe foil, but with a stronger white-line due to reaction-induced carburization. FIG. 24B shows that the Co K-edge XANES spectra were more like that of the Fe foil, rather than the Co foil, indicating the formation of Fe-Co alloy; additionally, a sharper pre-edge peak was observed after the reaction, which is a typical feature of cobalt carbide species formation.

[0041] FIG.24: Fe and Co K-edge EXAFS spectra and corresponding fitting plots in R space over the Fe3Co6 / CeO2 catalysts under different conditions as well as that over the Co and Fe foils.

[0042] FIG.26: Simulated referential XRD patterns of metals, metal oxides, and metal carbides. Notes: COD and MP represent the Crystallography Open Database and the Material Project database, respectively; the number in the parenthesis refer to the index in the corresponding database.CU24182 – 101879.000335

[0043] FIG.27: DFT-optimized bulk structures and properties of mono- and bimetallics. DFT-optimized bulk structures, lattice parameters, and magnetic moments for metallic Fe, FeCo alloy and metallic Co. Yellow: Fe; Blue: Co. Note: The value in parentheses are the experimentally-measured or literature-reported values.

[0044] FIG.28 DFT-optimized surface models for mono- and bimetallics.

[0045] DFT-optimized surface models for Fe(110), FeCo(110), and Co(0001) with top and side views. Yellow: Fe; Blue: Co.

[0046] FIG.29: DFT-optimized bulk structures and properties for metal carbides. DFT-optimized bulk structures, lattice parameters and magnetic moments for Fe3C, Fe2CoC, and Co3C carbides. Yellow: Fe; Blue: Co; Brown: C. Note: The values in parentheses are the experimentally-measured or literature-reported values.

[0047] FIG.30: DFT-optimized surface models for metal carbides. DFT- optimized surface models for Fe3C(001), Fe2CoC(001), and Co3C(001) with top and side views. Yellow: Fe; Blue: Co; Brown: C.

[0048] FIGs.31A-31B. AITD analysis for the stability of mono- / bi-metallics and metal carbides at different reaction environments. AITD analysis for the stability of bulk phases in the reaction environments with CO gas only (FIG.31A) and CO+H2 gas mixture (FIG.31B). Note: The AITD analysis conducted here is used to identify the most possible bulk phases formation for metallics and carbides at different reaction environments on bimetallic FeCo alloy. The identified stable FeCo alloy and Fe2CoC in this case were further used for the AITD analysis in FIG.7A, which compared the stability of bulk FeCo, Fe, Co metallics and Fe2CoC, Fe3C, Co3C carbides in different reaction environments.

[0049] FIGs.32A-32B. AITD analysis for the chemical potential change of C species in the equilibrium at different reaction environments. AITD analysis for the chemical potential change of C species (ΔμCR) at equilibrium in different reaction environments with CO gas only (FIG.32A) and CO+H2 gas mixture (FIG.32B) at different temperatures (325°C, 425°C, 525°C) and pressures (0.1 bar, 0.5 bar, 1.0 bar). Note: The values of ΔμCRare calculated based on the reaction of R-1: CO(g)+CO(g)^CO2(g)+C(s) and R-2: CO(g)+H2(g)^H2O(g)+C(s), which are the extreme cases for CO-only environment and CO+H2 environment, respectively.

[0050] FIGs.33A-33C. DFT-optimized structures for CO dissociation on the surfaces of mono- and bimetallics. Top and side views of the DFT-optimized intermediatesCU24182 – 101879.000335 and transition states for CO dissociation on (FIG.33A) Fe(110), (FIG.33B) FeCo(110), and (FIG.33C) Co(0001) surfaces. Yellow: Fe; Blue: Co; Brown: C; Red: O.

[0051] FIGs.34A-34C. DFT-optimized structures for CO dissociation on the surfaces of metal carbides. Top and side views of the DFT-optimized intermediates and transition states for CO dissociation on (FIG.34A) Fe3C(001), (FIG.34B) Fe2CoC(001), and (FIG.34C) Co3C(001) surfaces. Yellow: Fe; Blue: Co; Brown: C; Red: O.

[0052] FIGs.35A-35C. Perspective view for DFT-optimized structures for C-C bond formation, CO2 formation, O-H bond formation on surfaces of mono- and bimetallics. Perspective views of DFT-optimized structures for (FIG.35A) C-C bond formation, (FIG. 35B) CO2formation, (FIG.35C) O-H bond formation on FeCo(110) and Co(0001) surfaces; Yellow: Fe; Blue: Co; Brown: C; Red: O; White: H.

[0053] FIGs.36A-36C. 36 DFT-optimized structures for C-C bond formation on the surfaces of mono- and bi-metallics.

[0054] Top and side views of the DFT-optimized intermediates and transition states for C-C bond formation on (FIG.36A) Fe(110), (FIG.36B) FeCo(110), and (FIG 36C) Co(0001) surfaces. Yellow: Fe; Blue: Co; Brown: C.

[0055] FIGs.37A-37B. Structures of *C adsorption and diffusion. Top and side views of the DFT-optimized structures for *C adsorption and the corresponding transition states (TS) on (FIG.37A) FeCo(110) and (FIG.37B) Co(0001) surfaces. Yellow: Fe; Bule: Co; Brown: C.

[0056] FIGs.38A-38C. DFT-optimized structures for CO2 formation on the surfaces of mono- and bimetallics. Top and side views of the DFT-optimized intermediates and transition states for CO2 formation on (FIG.38A) Fe(110), (FIG.38B) FeCo(110), and (FIG.38C) Co(0001) surfaces. Yellow: Fe; Blue: Co; Brown: C; Red: O.

[0057] FIGs.39A-39C. DFT-optimized structures for O-H bond formation on surfaces of mono- and bimetallics. Top and side views of the DFT-optimized intermediates and transition states for O-H bond formation on (FIG.39A) Fe(110), (FIG.39B) FeCo(110), and (FIG.39C) Co(0001) surfaces. Yellow: Fe; Blue: Co; Red: O; White: H.

[0058] FIG.40. DFT-calculated potential energy diagram for H2O formation on surfaces of mono- and bimetallics. DFT-calculated potential energy diagram for H2O formation on Fe(110), FeCo(110), and Co(0001) surfaces and selected intermediates and transition states. Yellow: Fe; Blue: Co; Red: O; White: H.CU24182 – 101879.000335

[0059] FIGs.41A-41C. DFT-optimized structures for H2O formation on surfaces of mono- and bimetallics. Top and side views of the DFT-optimized intermediates and transition states for H2O formation on (FIG.41A) Fe(110), (FIG.41B) FeCo(110), and (FIG. 41C) Co(0001) surfaces. Yellow: Fe; Blue: Co; Red: O; White: H.

[0060] FIGs.42A-42C. Illustrations of active sites on mono- and bimetallics. Illustrations of different sites on Fe(110) (FIG.42A), FeCo(110) (FIG.42B), and Co(0001) (FIG.42C) surfaces. Yellow: Fe; Blue: Co; Brown: C.

[0061] FIGs.43A-43C. DFT-optimized structures for the adsorption of gas phase molecules on the surfaces of mono- and bimetallics. Top and side views of the DFT- optimized gas phase adsorption (CO, CO2, H2O) on (FIG.43A) Fe(110), (FIG.43B) FeCo(110), and (FIG.43C Co(0001) surfaces. Yellow: Fe; Blue: Co; Brown: C; Red: O.

[0062] FIGs.44A-44C. DFT-optimized structures for the adsorption of important reaction intermediates on surfaces of mono- and bimetallics. Top and side views of the DFT- optimized intermediates adsorption (*C, *O, *H, *OH) on (FIG.44A) Fe(110), (FIG.44B) FeCo(110), and (FIG.44C) Co(0001) surfaces. Yellow: Fe; Blue: Co; Brown: C; Red: O; White: H.

[0063] FIG.45 DFT-calculated electronic states for clean surface of mono- and bimetallics. Partial density of states (PDOS) for Fe 3d in Fe(110) and FeCo(110) surfaces, and Co 3d in FeCo(110) and Co(0001) surfaces. Note: Upon alloying, the Fe 3d spin-up states shifted to the left while Fe 3d spin-down states shifted to the right, pointing to the change of spin states. The Co 3d spin-up states became more localized while Co 3d spin- down states only changed slightly.

[0064] FIGs.46A-46B. DFT-calculated electronic states for mono- and bimetallics before and after CO adsorption. Partial density of states (PDOS) for (FIG.46A) Fe 3d in Fe(110) and FeCo(110) surfaces before and after CO adsorption, and (FIG.46B) Co 3d in FeCo(110) and Co(0001) surfaces before and after CO adsorption. Note: The alloying of monometallic Fe and Co hindered the donation of electrons from Fe 3d spin-down states to CO while helped the donation of electrons from Co 3d spin-down states to CO.

[0065] FIGs.47A-47B. DFT-calculated electronic states for mono- and bimetallics before and after H2O adsorption. Partial density of states (PDOS) for (FIG.47A) Fe 3d in Fe(110) and FeCo(110) surfaces before and after H2O adsorption, and (FIG.47B) Co 3d in FeCo(110) and Co(0001) surfaces before and after H2O adsorption. Note: The electronicCU24182 – 101879.000335 states for either monometallic or bimetallic does not vary much upon the adsorption of H2O, pointing to a limited effect of electronic structure for H2O adsorption.

[0066] FIGs.48A-48C. Structures for the C-C bond formation and *O fragment removal on Fe3C surface models. Top and side views of the DFT-optimized intermediates for C-C bond formation (FIG.48A) and *O fragment removal via Boudouard reaction (FIG. 48B) or H2O formation (FIG.48C) on Fe3C(001) surface. Yellow: Fe; Brown: C; Red: O.

[0067] FIGs.49A-49B. Effect of current densities on the catalytic performance of cathodic 10%Pd / C catalyst. FIG.49A: CO2 conversion and carbon balance (upper panel); molar flow rates of CO2, CO, H2, steam, and CH4 during the tandem process (lower panel); FIG.49B: Cell voltage at -60, -100, -150, and -200 mA cm-2as a function of time. Reaction conditions: 1 atm, electrolyzer (10%Pd / C, 1 mg cm-2, CO2 / N2=10 / 10 ml min-1, 25 ℃), and thermal reactor (Fe3Co6 / CeO2, 200 mg, 450 ℃). Note: Conv._total and Conv._products refer to the CO2 conversion based on total CO2 consumption (Xtotal) and carbon-containing gaseous products (Xproducts), respectively

[0068] FIGs.50A-50B. Effect of current densities on the catalytic performance of cathodic 20%Pd / C catalyst. FIG.50A: O2conversion and carbon balance (upper panel); molar flow rates of CO2, CO, H2, steam, and CH4 during the tandem process (lower panel); FIG.50B: Cell voltage at -60, -100, -150, and -200 mA cm-2as a function of time. Reaction conditions: 1 atm, electrolyzer (10%Pd / C, 1 mg cm-2, CO2 / N2=10 / 10 ml min-1, 25 ℃), and thermal reactor (Fe3Co6 / CeO2, 200 mg, 450 ℃). Note: Conv._total and Conv._products refer to the CO2 conversion based on total CO2 consumption (Xtotal) and carbon-containing gaseous products (Xproducts), respectively

[0069] FIGs.51A-51B: Effect of current densities on the catalytic performance of cathodic 40%Pd / C catalyst. FIG.51A: CO2conversion and carbon balance (upper panel); molar flow rates of CO2, CO, H2, steam, and CH4 during the tandem process (lower panel); FIG.51B: Cell voltage at -60, -100, -150, and -200 mA cm-2as a function of time. Reaction conditions: 1 atm, electrolyzer (10%Pd / C, 1 mg cm-2, CO2 / N2=10 / 10 ml min-1, 25 ℃), and thermal reactor (Fe3Co6 / CeO2, 200 mg, 450 ℃). Note: Conv._total and Conv._products refer to the CO2conversion based on total CO2consumption (Xtotal) and carbon-containing gaseous products (Xproducts), respectively

[0070] FIGs.52A-52C: Electrochemical-thermochemical tandem process using the 10%Pd / C catalyst at -200 mA cm-2and Fe3Co6 / CeO2 at 450 ℃. FIG.52A: CO2 conversionCU24182 – 101879.000335 and carbon balance; FIG.52B: Molar flow rates of CO2, CO, H2, steam, and CH4during the tandem process; FIG, 52C: Cell voltage at -200 mA cm-2as a function of time. Reaction conditions: 1 atm, electrolyzer (10%Pd / C, 1 mg cm-2, CO2 / N2=10 / 10 ml min-1, -200 mA cm-2, 25 ℃), and thermal reactor (Fe3Co6 / CeO2, 200 mg, 450 ℃). Note: Conv._total and Conv._products refer to the CO2conversion based on total CO2consumption (Xtotal) and carbon-containing gaseous products (Xproducts), respectively; EC and TC represent electrolyzer and thermal reactor, respectively; Given safety concerns, the reaction is stopped at 11 h because of the severe blockage of the thermal reactor

[0071] FIGs.53A-53G. EDS mappings over the spent Fe3Co6 / CeO2 sample after the electrochemical-thermochemical tandem experiment

[0072] FIG.54. Schematic illustration of the separation of CNF and regeneration of spent catalysts. Notes: Mδ+denotes the metal ions that are dissolved in the solution, while the small blue cubes in Impregnation represent new CeO2. More details about the procedures can be referred to in Methods in the manuscript; The dashed lines represent that new CeO2or a portion of the resulting CNF can be added and used as the support material for the re- synthesis of the recycled catalyst for subsequent reaction studies.

[0073] FIGs.55A-55P: SEM and TEM images over the solid products after the separation process. The catalyst used for separation was the FeCo catalyst that was spent after the EC-TC tandem reaction.

[0074] FIGs.56A1-56J2: TEM images and corresponding SAED (or fast Fourier transformation (FFT)) over the solid products after the separation process. Note: The catalyst used for separation was the FeCo catalyst that was spent after the EC-TC tandem reaction. Yellow dashed circles indicate the locations of the selected area apertures used for acquiring SAEDs.

[0075] FIGs.57A-57D. Catalytic performance over the regenerated FexCoy+CeO2 sample after CNF separation as well TEM images after CO-H2 reaction. FIG.57A: Conversion of CO and H2 as a function of time on stream. FIG.57B: Catalyst weight changes over the initial Fe6Co12 / CO2 catalyst and the regenerated FexCoy+CeO2 catalyst after the CO-H2reaction. FIGs.57C and 57D: TEM images over the spent FexCoy+CeO2catalyst. Reaction conditions: 1 atm, reduction (H2 / N2=20 / 20 ml min-1) at 500oC, reaction (CO / H2 / N2=3 / 3 / 4 ml min-1) at 450oC for 4.5 h. Note: I and W referred to the initial catalyst mass and the weight gain after the reaction, respectively.CU24182 – 101879.000335

[0076] FIGs.58A-58B: Catalytic performance over the regenerated FexCoy+solids sample after CNF separation as well TEM images after CO-H2 reaction. FIG.57A: Conversion of CO and H2as a function of time on stream. FIG.57B: Catalyst weight changes over the solid products after separation and the regenerated FexCoy+solids catalyst after the CO-H2reaction. Note: I and W referred to the initial catalyst mass and the weight gain after the reaction, respectively; the shadow region represented a mass decrease.

[0077] FIGs.59A-59C: Effect of co-feeding with water vapor on catalytic performance of Fe3Co6 / CeO2.

[0078] FIG.59A: Conversion of CO and H2 as well as carbon balance. FIG.59B: Selectivity of gaseous products. FIG.59C: Selectivity of all products. The panels from the top to the bottom of each column referred to the results obtained at the CO / H2 / N2 flow rates of 15 / 3 / 32, 15 / 3 / 32 (hydrated), 15 / 15 / 20, 15 / 15 / 20 (hydrated), 15 / 30 / 5, and 15 / 30 / 5 (hydrated) ml min-1, respectively. Other reaction conditions: 100 mg of catalyst, reduction temperature (500 °C), reaction temperature (450 °C), 1 atm, time on stream of about 5 h. Notes: The Selectivity in gaseous products was calculated based on carbon-containing products in gas phase; the Selectivity in total products was calculated based on all the carbon- containing products in both gas and solid phases. The water vapor was fed by passing the CO / H2 / N2mixture through a bubbler at room temperature.

[0079] FIG.60 Effect of CO2 on catalyst weight gain over the spent Fe3Co6 / CeO2 catalyst. Reaction conditions: 100 mg of catalyst, CO / H2 / CO2 / N2=15 / 03 / 00 / 32, 15 / 03 / 15 / 17, 15 / 15 / 00 / 20, 15 / 15 / 05 / 15, and 15 / 15 / 15 / 05 ml min-1, reduction temperature (500 °C), reaction temperature (450 °C), 1 atm, time on stream of about 5 h.

[0080] FIGs.61A-61B: Energy cost and net CO2 emission per kilogram CNF products based on the electrocatalytic-thermocatalytic tandem scheme. FIG.61A: Energy cost per kilogram of CNF. FIG.61B: Net CO2 emission per kilogram of CNF. Reaction conditions: 1 atm, electrolyzer (10%Pd / C with a mass loading of 1 mg cm-2, 5 cm2cathode, - 150 mA cm-2, CO2 / N2=10 / 10 ml min-1, 25 ℃), and thermocatalytic reactor (Fe3Co6 / CeO2, 200 mg, 450 ℃), 11 h. DETAILED DESCRIPTION OF ILLUSTRATIVE EMBODIMENTSCU24182 – 101879.000335

[0081] The present disclosure may be understood more readily by reference to the following detailed description of desired embodiments and the examples included therein.

[0082] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art. In case of conflict, the present document, including definitions, will control. Preferred methods and materials are described below, although methods and materials similar or equivalent to those described herein can be used in practice or testing. All publications, patent applications, patents and other references mentioned herein are incorporated by reference in their entirety. The materials, methods, and examples disclosed herein are illustrative only and not intended to be limiting.

[0083] The singular forms “a,” “an,” and “the” include plural referents unless the context clearly dictates otherwise.

[0084] As used in the specification and in the claims, the term "comprising" can include the embodiments "consisting of" and "consisting essentially of.” The terms “comprise(s),” “include(s),” “having,” “has,” “can,” “contain(s),” and variants thereof, as used herein, are intended to be open-ended transitional phrases, terms, or words that require the presence of the named ingredients / steps and permit the presence of other ingredients / steps. However, such description should be construed as also describing compositions or processes as "consisting of" and "consisting essentially of" the enumerated ingredients / steps, which allows the presence of only the named ingredients / steps, along with any impurities that might result therefrom, and excludes other ingredients / steps.

[0085] As used herein, the terms “about” and “at or about” mean that the amount or value in question can be the value designated some other value approximately or about the same. It is generally understood, as used herein, that it is the nominal value indicated ±10% variation unless otherwise indicated or inferred. The term is intended to convey that similar values promote equivalent results or effects recited in the claims. That is, it is understood that amounts, sizes, formulations, parameters, and other quantities and characteristics are not and need not be exact, but can be approximate and / or larger or smaller, as desired, reflecting tolerances, conversion factors, rounding off, measurement error and the like, and other factors known to those of skill in the art. In general, an amount, size, formulation, parameter or other quantity or characteristic is “about” or “approximate” whether or not expressly stated to beCU24182 – 101879.000335 such. It is understood that where “about” is used before a quantitative value, the parameter also includes the specific quantitative value itself, unless specifically stated otherwise.

[0086] Unless indicated to the contrary, the numerical values should be understood to include numerical values which are the same when reduced to the same number of significant figures and numerical values which differ from the stated value by less than the experimental error of conventional measurement technique of the type described in the present application to determine the value.

[0087] All ranges disclosed herein are inclusive of the recited endpoint and independently of the endpoints. The endpoints of the ranges and any values disclosed herein are not limited to the precise range or value; they are sufficiently imprecise to include values approximating these ranges and / or values.

[0088] As used herein, approximating language can be applied to modify any quantitative representation that can vary without resulting in a change in the basic function to which it is related. Accordingly, a value modified by a term or terms, such as “about” and “substantially,” may not be limited to the precise value specified, in some cases. In at least some instances, the approximating language can correspond to the precision of an instrument for measuring the value. The modifier “about” should also be considered as disclosing the range defined by the absolute values of the two endpoints. For example, the expression “from about 2 to about 4” also discloses the range “from 2 to 4.” The term “about” can refer to plus or minus 10% of the indicated number. For example, “about 10%” can indicate a range of 9% to 11%, and “about 1” can mean from 0.9-1.1. Other meanings of “about” can be apparent from the context, such as rounding off, so, for example “about 1” can also mean from 0.5 to 1.4.

[0089] Further, the term “comprising” should be understood as having its open- ended meaning of “including,” but the term also includes the closed meaning of the term “consisting.” For example, a composition that comprises components A and B can be a composition that includes A, B, and other components, but can also be a composition made of A and B only. Any documents cited herein are incorporated by reference in their entireties for any and all purposes.

[0090] Any embodiment or aspect provided herein is illustrative only and does not limit the scope of the present disclosure or the appended claims. Any part or parts of any oneCU24182 – 101879.000335 or more embodiments or aspects can be combined with any part or parts of any one or more other embodiments or aspects.

[0091] Carbon dioxide (CO2) fixation into value-added solid carbon such as carbon nanofiber (CNF) for longer-term storage represents a promising avenue for achieving net- negative carbon emissions. However, directly converting CO2to CNF via thermocatalytic approaches faces thermodynamic constraints, while electrocatalytic methods typically lead to amorphous carbon with limited yields or require energy-intensive conditions (>720 °C).

[0092] Herein, we present an electrocatalytic-thermocatalytic tandem strategy for CNF production, which circumvents the aforementioned thermodynamic limitations by integrating the co-electrolysis of CO2and water into syngas (CO and H2) with a subsequent thermochemical process at comparatively mild conditions (370-450 ℃, 1 atm), yielding CNF at a high production rate (avg.2.5 gcarbongmetals-1h-1). The coordinated actions of FeCo alloy and extra metallic Co were ascertained to enhance the dissociative activation of syngas and favor the carbon-carbon bond formation to produce CNF. This tandem strategy opens a door to leverage renewable energy for decarbonizing CO2 into valuable solid carbon products while producing renewable H2.

[0093] The global climate crisis has become a major concern due to the emissions of anthropogenic carbon dioxide (CO2).1,2CO2mitigation is strongly demanded in hard-to- abate sectors, such as the cement, steel, and chemical industries. Carbon capture, utilization, and storage (CCUS) has been identified as one promising strategy to achieve net zero- emissions.3Capturing CO2 for underground storage has been discussed for decades, however, engineering challenges and risks of leakage have hampered implementation. In parallel, various approaches to CO2 conversion, including thermochemical, electrochemical, photochemical, plasma, biological, and hybrid processes, have been advanced to produce a variety of value-added chemicals (e.g., syngas, alkanes, olefins, aromatics, and oxygenates) or fuels (e.g., gasoline, diesel, and jet fuels). But most of the resulting gaseous or volatile products are eventually released to atmosphere as CO2 after consumption, thereby making it difficult to achieve negative CO2 emissions. Moreover, the vast scale of carbon emissions, estimated to be 40 Gt CO2annually, presents challenges in discovering breakthrough technologies to effectively mitigate CO2 emissions.

[0094] CO2valorization into non-volatile or solid products that are suitable for longer-term storage is expected to play a crucial role in achieving negative emissions as theCU24182 – 101879.000335 world shifts toward renewable energy sources. Solid carbon materials such as carbon nanotubes (CNT), carbon nanofibers (CNF), graphene, and carbon black have gained interest in a wide range of industrial sectors with a > $30 billion market size due to their unique structural, mechanical, electrical, thermal, optical, and chemical properties. But thermodynamic constraints pose challenges for direct CO2-to-solid carbon conversion via thermocatalytic methods. At present, there have been only a limited number of reports on CO2valorization into solid carbon products: the electrochemical reduction of CO2with Galinstan-based liquid metals, which produces solid carbons, but in amorphous morphology and with low yield and / or low current density (<10 mA cm-2); alternatively, CO2 electrolysis in molten lithium (Li) bicarbonate was introduced to synthesize CNT or CNF, but it requires high temperatures (>720 ℃) and competes for the limited Li supplies with the battery industry. As such, approaches that allow for large-scale conversion of CO2into value-added solid carbon products at mild conditions, in facile integration with renewable energy and good compatibility with existing infrastructure, will open a door for negative carbon emissions and economically and technically facilitate commercial decarbonization.

[0095] To achieve this, herein we propose an electrochemical-thermochemical tandem strategy that circumvents thermodynamic limitations (ΔGeq). This process firstly electrolyzes CO2and water into syngas (CO and H2), followed by thermochemical upgrading into CNF at relatively low temperatures (370-450 ℃) and ambient pressure. Moreover, H2 can be obtained as a renewable byproduct. Experimental and theoretical insights reveal that the synergistic actions between FeCo alloy and extra metallic Co serve as a suitable optimal catalyst for the CNF production. Furthermore, proof-of-principle demonstrations of product separation involving metal recycling and assessment of energy costs and CO2 emissions reveal the promising applicability of the tandem strategy.

[0096] Results

[0097] CNF production from thermal conversion of syngas at mild conditions

[0098] In addition to the wide range of applications in power generation and storage, electronic devices and gas storage, CNF have been shown to reinforce composites such as cements by preventing nanocracks and filling pores in the cementitious matrix, thus presenting an opportunity for permanent fixation of large amounts of carbon into worldwide building materials. However, the conventional synthesis methods of pyrolysis or catalytic decomposition of carbon precursors typically require high temperatures (>700 ℃) and oftenCU24182 – 101879.000335 at elevated pressures, representing an energy-intensive process. Thus, a challenge is to produce appreciable amounts of CNF from CO2 with lower energy consumption, preferably powered by renewable energy.

[0099] The reactions of 2CO ^ C(s) + CO2(i.e., Boudouard reaction, R1) and CO + H2^ C(s) + H2O (R2) have been identified for carbon formation. However, they are typically considered undesired coking reactions that are responsible for catalyst deactivation blockage in commercial processes such as Fischer-Tropsch (F-T) synthesis and steam methane reforming. Intriguingly, both R1 and R2 are highly exothermic (FIG.1a) and therefore thermodynamically favorable at low temperatures. R1 is energetically more preferred than R2 with a higher equilibrium conversion of CO (e.g., 94.1% vs.75.3% at 400 ℃, FIG.1b), while R2 exhibits a higher C(s) yield than R1 (e.g., 75.3% vs.47.3% at 400 ℃). In the presence of both pathways, CO conversion can be retained at above 90% up to 450 ℃, followed by rapid decreases at higher temperatures; in parallel, the C(s) amount or yield gradually decreases (e.g., 66.7% at 400 ℃) and becomes nearly negligible at above 600 ℃ (FIG.1c). It is noted that the practical equilibrium C(s) yield is likely lower considering the potentially competitive reactions of CO methanation (CO + 3H2 ^ CH4 + H2O), water-gas shift (WGS, CO + H2O ^ H2+ CO2), and / or carbon gasification (C(s) + 2H2^ CH4). Thus, catalysts with high selectivity to C(s) can be employed to maximize the solid carbon products; moreover, to balance the thermodynamics and kinetics of solid carbon production, optimal temperatures appear to be between 400-450 ℃.

[0100] Although CNF formation is a complex process, it is primarily governed by three successive steps: dissociative adsorption of carbon precursors at the gas / metal interface, carbon diffusion through the bulk of metal catalysts or migration along the catalyst surface, and carbon precipitation and growth at the catalyst / solid carbon interface. Among potential catalysts, Fe, Co, Ni, and Ru have been recognized as highly effective metals for the dissociative adsorption of CO and H2in traditional F-T synthesis. However, Fe exhibits higher WGS activity, Ni favors methanation more than Co, and Ru is limited by its high cost. Bimetallic catalysts have been shown to optimize reactant activation and alter selectivity.6,33,34Therefore, supported non-precious FeCo bimetallic catalysts are employed in this study to achieve CNF formation at relatively mild temperatures and ambient pressure.

[0101] Catalyst weight gain (CWG) relative to the initial catalyst loading is employed as an initial indicator of solid carbon formation after reaction. The elementalCU24182 – 101879.000335 composition effect (FIG.1D, FIGs.5-6) is assessed by varying the Fe / Co ratio from 9 / 0 to 6 / 3, 4.5 / 4.5, 3 / 6, and 0 / 9 while keeping the total moles of metal elements constant. The highest CWG (265 mg) is obtained at the Fe / Co ratio of 3 / 6, showing the synergistic effect of alloying Fe and Co for catalyzing solid carbon production. Solid carbon production can be further enhanced to 350 mg with higher metal loading (Fe6Co12) while keeping the same Fe / Co atomic ratio. It is noted that with CO feed only (CO / H2 / N2=3 / 0 / 7 ml min-1), rapid deactivation of Fe3Co6 / CeO2is observed with a small CWG (40 mg) due to carbon encapsulation (FIG.7). The use of H2 in CNF formation is corroborated by an increased CWG when co-feeding H2 (3 / 3 / 4 ml min-1in FIG.1E, FIG.8). Further, the observation of CWG under various feed composition ratios highlights the tolerance of the FeCo catalysts for CNF formation using H2 / CO mixtures with a relatively wide range of feed compositions.

[0102] The CWG also increases with the flow rates of CO / H2 / N2(FIG.9). A comparison of reduction and reaction temperatures (FIGs.10-12) reveal optimal carbon formation at around 450 ℃ (FIG.1F, FIG.12), in line with the thermodynamic analysis of the reaction temperature window. Although both CO and H2 show a gradual decrease in conversion after being on stream for 30 min, primarily due to the diminishing methanation reaction toward CH4, the carbon fixation rate remains stable within about 5 h (FIG.1g, FIG. 13) as evidenced by the constant carbon balance. A gram-level CWG (1.7 g) can be obtained within ~37 h (FIG.14). Additionally, the role of the CeO2 support in promoting solid carbon formation is highlighted by the lower CWG values observed over its γ-Al2O3, SiO2, and MCM-41 supported counterparts (FIG.16). It is noted that the metal-CeO2 interface in activating reactants is limited according to the density function theory (DFT) calculations (see Supplementary Note 1, FIG.16, and Supplementary Table 1). The enhanced performance of CeO2-supported catalysts toward solid carbon formation is primarily due to the more facile reducibility of Fe and Co oxides supported on CeO2 than the other supports, as corroborated by the temperature-programmed reduction (TPR) measurements (FIG.10).

[0103] Transmission electron microscopy (TEM) imaging of the spent Fe3Co6 / CeO2 catalyst reveals the formation of CNF with diameters of 20-30 nm and lengths of hundreds of nm to a few μm. The formation of CNF primarily follows a tip-growth mechanism, which pushes the metal nanoparticles away from the support and places them within the tip of CNF (FIG.1i), as further confirmed by the combined high-angle annular dark field (HAADF) imaging and energy-dispersive X-ray spectroscopy (EDS) mapping ofCU24182 – 101879.000335 C, Fe, Co, and Ce (FIG.1j, Supplementary Tables 2-3). Analysis using Raman spectroscopy (FIG.19 and Supplementary Table 4) suggests a CNF purity of ~90% and crystallinity of ~80% relative to the commercial CNF reference, consistent with the presence of a minor fraction of amorphous carbon agglomerates seen in the TEM image (FIG.1h). In addition, the EDS measurements (FIG.20, Supplementary Table 5) conducted over the resulting CNFs reveal a high elemental purity as evidenced by the atomic compositions (C: ~99.0%, O: <1.0%, Fe, Co, and Ce: <0.2%). Co9 / CeO2follows a similar tip-growth mechanism of CNF (FIG.21). In contrast, the spent Fe9 / CeO2 catalyst shows predominantly amorphous carbon species (FIG.22), confirming the synergistic effect of the Fe-Co bimetallic catalysts.

[0104] Structural properties of catalysts responsible for CNF formation

[0105] To unravel the structures that are responsible for the synergistic interaction of Fe and Co in CO and H2dissociative activation and C-C backbone formation, in situ powder X-ray diffraction (XRD) and in situ X-ray absorption fine structure (XAFS) spectroscopy are utilized. Co3O4(2θ(311)=14.62°, 2θ(222)=15.28°, and 2θ(400)=17.66°, FIG.2a) is observed to be dominant over the fresh Co9 / CeO2 catalyst, which is reduced to CoO (2θ(111)=14.42° and 2θ(200)=16.67°) first before being fully reduced to metallic Co with the co- existence of fcc (2θ(111)=17.37°) and hcp (2θ(100)=16.39°) phases. The reduction behavior of Co3O4to CoO is observed similarly in the FexCoy / CeO2bimetallic catalysts but with diminished peak intensity as the Fe / Co ratio increases from 3 / 6 to 4.5 / 4.5 and 6 / 3. At 500 ℃ reduction, Fe3Co6 / CeO2shows the dominance of a bcc Fe-Co alloy phase (2θ(110)=17.67°) with a minor amount of fcc Co phase (2θ(111)=17.35°). As for Fe9 / CeO2, a bcc Fe phase (2θ(110)=17.48°) forms after 700 ℃ reduction. Upon exposure to the CO and H2stream at 450 ℃, the metallic Fe phase vanishes instantly, which is attributed to fragmentation, a well- documented phenomenon caused by carbon supersaturation within the bulk of metal particles. A similar phenomenon was observed over Fe3Co6 / SiO2 (FIG.23). In contrast, for Fe3Co6 / CeO2 the alloy phase remains stable under reaction conditions, indicating the stabilizing effect of Co on the alloy phase. Furthermore, the presence of additional metallic Co in Fe3Co6 / CeO2, which results in a higher CWG (265 mg) than Fe4.5Co4.5 / CeO2 (190 mg) and Co9 / CeO2(120 mg), suggests that the co-existence of FeCo alloy and extra Co is likely the optimal combination for CO and H2 dissociative activation and C-C backbone formation as supported by the DFT calculations.CU24182 – 101879.000335

[0106] The in situ Fe and Co K-edge X-ray absorption near-edge spectroscopy (XANES) results (Figs.2b, 2e) indicate that Fe2O3 and Co3O4 over Fe3Co6 / CeO2 are reduced to metallic state at 500 ℃. The Fe-Co alloy formation in Fe3Co6 / CeO2is further validated by the Fe and Co K-edge extended X-ray absorption fine structure (EXAFS) features in both k- space (Figs.2c, 2f) and R space (Figs.2d, 2g), which resemble those of a bcc Fe foil rather than an fcc Co foil, consistent with the in situ XRD results. After exposure to CO and H2 at 450 ℃ for about 5 h, both Fe and Co are slightly oxidized but with different oxidative features, in particular the pre-edge peak (Figs.2b, 2e, FIG.24), from their oxide standards (FeO, Fe3O4, Fe2O3, CoO, and Co3O4). This is likely due to metal-C bond formation,40as revealed by the EXAFS fittings (FIG.25 and Supplementary Table 6) where the average coordination number (CN) of Co-C bond is 0.8 with a bond distance of 1.80 Å. Although the formation of a minor amount of metal carbides is not observed by XRD, likely due to the lack of long-range ordering, they may affect the activation of CO and H2 and the stability of catalysts as supported by the DFT calculations.

[0107] Mechanistic insights into syngas activation and C-C bond formation

[0108] DFT studies (FIG.3, FIGs.27-48, Supplementary Tables 7-14) were performed to enhance insights into the structures and catalytic behaviors observed experimentally, where bcc FeCo was selected to model the Fe-Co alloy according to the in situ XRD measurement (FIG.2a). The thermodynamic analysis (FIG.3a) demonstrates that under the reaction conditions of both R1 and R2, bcc Fe is highly unstable and prone to form Fe3C, while hcp Co prefers to stay as metallic. By combining Co and Fe, the FeCo alloy displays a stability as high as Co with the metallic phase being favored. Introducing H2along with CO gas also helps stabilize the metallic phase for Fe, Co and FeCo due to the H2- induced increase in chemical potential change of C species (ΔμCR) at the equilibrium states (FIG.32).

[0109] To compare the reaction pathways, the most stable orientation is considered for the monometallic and bimetallic surfaces, i.e., Fe(110), Co(0001) and FeCo(110). Upon exposure to CO only (FIG.3B, FIG.33), the Fe(110) surface is highly reactive in term of adsorption (adsorption energy, Eads=-2.17 eV) and dissociation (reaction energy, ΔE=-0.89 eV, activation barrier, Ea=1.18 eV). In comparison, the Co(0001) surface binds CO more weakly (Eads=-1.86 eV) with a higher barrier for C-O bond cleavage (ΔE=0.82 eV, Ea=2.28 eV). The FeCo(110) alloy surface exhibits a comparable reactivity to Fe(110) towardCU24182 – 101879.000335 adsorption (Eads=-2.09 eV) and dissociation (ΔE=-0.19 eV, Ea=1.64 eV) of CO (FIG.3b), where both Fe and Co sites exposed on the surface participate in CO dissociation to stabilize the dissociated *C and *O fragments.

[0110] Once *C is formed from CO dissociation, the initial C-C bond formation toward the growth of CNF is unfavorable on Fe(110) (ΔE= 0.65 eV, Ea=1.43 eV) (FIG.3c, Supplementary Table 7). Instead, the conversion of Fe to Fe3C is preferred under reaction conditions (FIG.3a, Supplementary Note 2), which likely reduces activity due to difficulty in C-O bond activation (Ea = 2.49 eV, FIG.34, and Supplementary Table 8). In comparison, the C-C bond formation is more preferred or at least comparable to the reverse decomposition on both Co(0001) (ΔE=-0.62 eV, Ea=0.96 eV) and FeCo(110) (ΔE=0.26 eV, Ea=1.13 eV) surfaces (FIG.3c, FIGs.35-36). Furthermore, compared to Co alone, Co alloyed with Fe likely facilitates the CO dissociation (FeCo: ^E = -0.19 eV, Ea= 1.64 eV; Co: ^E = 0.82 eV, Ea= 2.28 eV), but with a reduced rate for C-C bond formation. Thus, the coexistence of Co with FeCo likely takes advantage of both systems, being able to allow CO dissociation to produce *C on the FeCo alloy, while facilitating the C-C coupling of the produced *C species and CNF growth on Co by the facile diffusion of *C species from FeCo to Co (see FIG.37 and Supplementary Table 9). This is consistent with experimental observation of higher CWG values on FeCo catalysts with higher Co / Fe ratios (FIG.1d) and the corresponding detection of the co-existence of FeCo and metallic Co by in situ XRD (FIG.2).

[0111] As observed experimentally, a trend going from Fe < Co ≈ FeCo in increasing activity during CNF formation was observed for the removal of *O fragments from *CO dissociation, which can occur via the formation of either CO2by reacting with neighboring *CO via the Boudouard reaction (FIG.3d, FIG.38) or H2O by reacting with *H from H2dissociation (FIG.3e, FIGs.39-41, Supplementary Table 7). In both cases, FeCo(110) and Co(0001) can enable more facile removal of *O than Fe. Note that the hydrogenation of *OH species to *H2O on both FeCo (Ea=1.74 eV) and Co (Ea=1.50 eV) requires overcoming a higher energy barrier than the formation of *CO2 (FIG.3, Supplementary Table 7), likely indicating a dominant Boudouard reaction for CNF formation as observed experimentally (FIG.1). In addition, the *H-assisted conversion of CO to *HCO or *COH is thermodynamically less favorable compared to the direct CO dissociation (FIG. 3b, Supplementary Table 10), which also ensures the preference for R1 over R2.CU24182 – 101879.000335

[0112] Overall, the formation of the FeCo alloy is advantageous over Fe and Co alone in CO / H2 activation, being as active as that of Fe to facilitate the direct CO dissociation, but still moderately enough to enable facile *O removal and C-C bond formation as that of Co toward the growth of CNF. In addition, the FeCo metallic phase, which is essential for initiating the CNF growth, is stable enough as that of Co to prevent carbide formation under reaction conditions. The catalytic promoting effect of FeCo strongly depends on interplay between the ensemble and electronic effects (see Supplementary Note 4, FIGs.42-46, Supplementary Tables 11-14), which effectively promotes the Boudouard reaction and initiates the CNF growth.

[0113] CO2to CNF via the electrochemical-thermochemical tandem process

[0114] The successful synthesis of CNF from CO and H2 over the FeCo catalysts suggests that, in principle, any process capable of producing syngas with appropriate H2 / CO ratios can potentially be used to generate the feedstock to produce CNF (FIG.1e, FIGs.8-9). One promising method is the electrochemical CO2reduction reaction (CO2RR) because it easily integrates with renewable electricity and can produce H2 / CO in controllable ratios at industrially relevant current densities (>1 A cm-2). We demonstrate an electrochemical- thermochemical tandem process for CNF production as illustrated in FIG.4a. The tandem process starts with the co-electrolysis of CO2and H2O at the cathode to produce CO via CO2RR and H2 via the hydrogen evolution reaction (HER), respectively, balanced by the oxygen evolution reaction (OER) at the anode. The resulting cathodic products (CO and H2) are subsequently fed into the thermochemical reactor to produce CNF via the R1 and / or R2 pathways.

[0115] HER is typically considered as an undesired competing reaction to CO2RR; however, the FeCo thermocatalyst is tolerant to—and in fact benefits from—the presence of H2 (FIG.8), which allows for more flexibility in selecting CO2RR catalysts and optimization of electrolyzer conditions, including the overpotential, electrolyte composition, ion exchange membrane, and electrolyte pH. We previously demonstrated that the in situ formation of β- phase Pd hydride over a Pd / C catalyst can provide a H2 / CO ratio between 1 and 2, which can be fed into the thermochemical reactor for CNF production. We employ a 5 cm2zero-gap membrane electrode assembly (MEA) electrolyzer equipped with an anion exchange membrane and adopt commercially available 10%, 20%, and 40%Pd / C catalysts immobilized on a gas diffusion layer. Under galvanostatic conditions, the current density is raised from -60CU24182 – 101879.000335 to -100, -150, and -200 mA cm-2. As summarized in FIGs.49-52, electrolysis using 10%Pd / C at -150 mA cm-2represents the optimal combination (see Supplementary Note 5) to integrate with the thermochemical reactor operating at 450 ℃. FIG.4b shows that a stable conversion of CO2, in terms of both Xtotal (~78.5%) and Xproducts (~15.2%), is obtained at -150 mA cm-2within the 11 h of tandem process. As indicated by the carbon balance, the significant difference between Xtotal and Xproducts during the tandem process should not only stem from carbonation at the cathode and transport of (bi)carbonates toward the anode, but also from carbon fixation as a solid. H2 is the dominant product (~0.4 mol gPd / C-1h-1) in the gas phase followed by CO, while the side hydrogenation product, CH4, is largely suppressed. The resulting spent Fe3Co6 / CeO2catalyst shows 240 mg of CWG (FIG.4e), corresponding to a production rate of avg.2.5 gcarbon gmetals-1h-1, and the TEM imaging (FIG.4f) reveals the predominant formation of CNF. Accordingly, the Raman spectra (FIG.19 and Supplementary Table 4) suggest a nearly 100% purity and ~80% crystallinity relative to the commercial CNF reference; moreover, the EDS analysis (FIG.53 and Supplementary Table 15) confirms the prevalence of carbon, with impurities (e.g., O, Fe, Co, and Ce) being less than 1.0 atomic %. At a higher current density of -200 mA cm-2(FIG.52), a similar CWG of 250 mg was obtained with a higher H2 productivity of ~0.7 mol gPd / C-1h-1.

[0116] FIG.4d shows that the CNF products are connected to the FeCo catalysts. For practical applications, one can separate the CNF from the catalysts and recycle the metals for subsequent reactions. As illustrated in FIG.54, we demonstrate a separation process including a sequence of steps involving ultrasonication, nitric acid leaching, centrifugation, and filtration, after which the separated catalysts are regenerated and reused for subsequent reactions. The spent FeCo sample is treated with nitric acid for 2, 5, and 12 h. The ICP analysis (Supplementary Table 16) indicates that ~70% of metals could be leached out after 2 h of treatment, while 5 and 12 h could leach out most of the metals (80-90%). Thus, a 5 h nitric acid treatment is used for subsequent regeneration experiments. Collected across various regions and scales, the TEM, scanning electron microscopy (SEM), and selected area electron diffraction (SAED) images (see Supplementary Note 6, FIGs.55 and 56) reveal the presence of high-crystallinity CNF and the open CNF tips after the separation process, indicating efficient removal of metal ensembles within the tip through treatments like ultrasonication and nitric acid digestion. For catalyst regeneration, we either added new CeO2(FexCoy+CeO2) or used some of the obtained solid products after separation (FexCoy+solids)CU24182 – 101879.000335 to a concentrated solution containing iron and cobalt ions to achieve a similar metal loading as catalysts in the initial experiments. FIGs.56A-56B show that the FexCoy+CeO2 sample exhibits stable and similar CO and H2conversion compared to the initial Fe6Co12 / CeO2catalyst within about 5 h, along with a comparable CWG (170 vs.160 mg). TEM analysis (FIGs.57C-57D) reveals the formation of CNFs, similar to the fresh catalysts. The FexCoy+solids sample (Fig.58) shows relatively steady and lower conversion, in line with a lower CWG (114 mg). These results demonstrate the proof-of-principle of recycling and reusing the FeCo catalysts; future follow-up studies will further optimize the recycling and regeneration procedures.

[0117] As shown in FIG.4b and FIG.52, the effluents from the tandem reactors mainly contain H2, CO, CO2, with a minor presence of H2O, and negligible CH4. Separating the produced CNF product from the gaseous effluents is straightforward due to the distinct phases of solid and gas, respectively. While separating H2 from the gas stream, which primarily contains CO and CO2, may pose a challenge, several commercialized separation techniques can be readily incorporated, including pressure swing adsorption, cryogenic distillation, and membrane separation. The choice of the most suitable separation method for practical applications depends on various factors, such as the gas volume, required H2 purity, energy efficiency, and cost considerations. Following the separation process, as postulated by the dashed lines in FIG.4a, the remaining CO and CO2, along with a minor amount of H2O if any, can be directly recycled to the first electrolyzer for subsequent reactions. Moreover, control experiments where syngas is co-fed with water vapor or CO2 into the thermal reactor show the potentially positive or neutral effects of recycling CO2and H2O on the subsequent formation of CNF (see Supplementary Note 7, FIGs.59-60).

[0118] Furthermore, we analyze energy consumption and net CO2emissions for our EC-TC system (FIG.61A). Based on the laboratory-scale setup with a 10%Pd / C cathode in the electrolyzer at room temperature and the subsequent thermocatalytic reactor at 450 °C under ambient pressure, the total energy cost of the current EC-TC system is ~162 kWh / kgCNF, lower than the reported value (~418 kWh / kgCNT) for an electrochemical system operating at 750 °C for CO2fixation. Moreover, the EC-TC system produces an appreciable amount of H2 and CO in addition to the solid carbon products, as well as O2 reported in the precious electrochemical approaches. The current energy cost corresponds to net CO2emissions of ~56 kgCO2 / kgCNF, which can be reduced to about 17 kgCO2 / kgCNF with fullCU24182 – 101879.000335 integration of renewable energy for the electrochemical process. In theory, a negative emission of -3.67 kgCO2 / kgCNF (FIG.61B) can be achieved if all EC-TC tandem processes are driven by renewable energy.

[0119] Conclusions

[0120] Numerous carbon dioxide (CO2) transformation technologies generate end chemical products that re-emit CO2 in their life cycle, making the CO2 footprint of these processes at best carbon neutral. In contrast, results from the current study demonstrate a tandem electrochemical-thermochemical process to enable the fixation of CO2 as valuable CNF at relatively mild conditions. By integrating a Pd / C electrocatalyst with a non-precious iron-cobalt (FeCo) thermocatalyst, we have achieved high quality CNF with an appreciable production rate (avg.2.5 gcarbon gmetals-1h-1) at ambient pressure. Combined reactor studies, in situ characterization and DFT calculations reveal the coordinated action of FeCo alloy with extra metallic Co as the highly active and selective catalyst for CO and H2 dissociation, *O removal, and C-C bond formation in the CNF growth; moreover, the stable FeCo alloy demonstrates enhanced resistance to the formation of less active carbide species. Overall, by harnessing renewable energy and knowledge in thermocatalytic reactions, together with the implementation of product separation with metal recycling and an analysis of energy cost and CO2emissions, the tandem strategy represents a viable pathway to achieve negative carbon emissions while generating value-added solid carbon products and renewable H2.

[0121] Methods

[0122] Catalysts Preparation for Thermochemical Reaction

[0123] All the catalysts for the thermochemical reaction were synthesized via a slurry impregnation method. For supported bimetallic catalysts, desired amounts of iron and cobalt nitrate precursors were dissolved in 30 ml of deionized (DI) water under stirring, followed by ultrasonication for 10 min. Afterwards, the support was added to the precursor solution under stirring with another 10 min of ultrasonication. The resulting liquid suspension was constantly stirred and dried at 70 °C overnight. The dried sample was ground into powder before being subject to calcination at 400 °C in static air for 2 h with a heating ramp rate of 1 °C min-1. Following the same procedure, the monometallic catalysts were synthesized with the same total molar amount of metal elements as the bimetallic catalysts. Details about the precursors, metal loadings, atomic ratios, and support identity were specified in Supplementary Table 1.CU24182 – 101879.000335

[0124] Electrodes Preparation for Electrochemical Reaction

[0125] Commercial 10%Pd / C, 20%Pd / C, and 40%Pd / C (Vulcan carbon XC-72) catalysts were purchased from Fuel Cell Store and used without further purification. For cathode preparation, 60 mg of Pd / C powder was dispersed in a mixture containing 3 ml of isopropanol (Fisher Chemical), 1 ml of DI water, and 145 μl of 5% Nafion solution (Sigma Aldrich), followed by ultrasonication for 30 min. The resulting cathode ink was hand sprayed with an airbrush onto a carbon gas diffusion layer with a microporous layer (Sigracet 39BB, Fuel Cell Store) to obtain a mass loading of 1.0±0.1 mg cm-2after drying. Likewise, the anode was prepared by spraying the ink of IrO2 (Alfa Aesar, Premion®, 99.99% (metals basis)) onto a platinized titanium fiber felt (Fuel Cell Store) with a mass loading of 3.0±0.1 mg cm-2.

[0126] In situ X-Ray Diffraction (XRD)

[0127] The in situ time-resolved XRD measurements were performed at beamline 7-BM (QAS, 1012ph s-1@ 10 keV) of the National Synchrotron Light Source II (NSLS-II) at Brookhaven National Laboratory (BNL). A Clausen cell equipped with a quartz capillary (1.2 mm O.D., 1.0 mm I.D.) was employed. For each in situ experiment, the sample (60-80 mesh) was fixed by two pieces of quartz wool and heated in a H2 / N2 mixture (5 / 5 ml min-1) from room temperature to 500 °C (or 700 °C) with a heating ramp rate of 10 °C min-1and held for 30 min. Afterwards, the sample was cooled down in the same atmosphere to 450 °C in 5 min before being exposed to the reaction stream (CO / H2 / N2=3 / 3 / 4 ml min-1) at 450 °C. For safety concerns, the reaction time was determined by the pressure build-up within the reactor. The energy of the incident X-ray was fixed and calibrated to the K-edge (20000 eV) of Mo foil, corresponding to a wavelength of 0.6199 Ǻ. The two-dimensional (2D) diffraction images were continuously collected using a PerkinElmer 1621 area detector at a speed of one scan per 24 s. The sample-to-detector distance was calibrated by a CeO2 standard. The 2D images were processed and integrated using the Dioptas software to obtain the XRD profiles after background subtraction.

[0128] In situ and ex situ X-Ray Absorption Spectroscopy (XAS)

[0129] The in situ and ex situ XAS measurements at the Fe K (7112 eV) and Co K (7709 eV) edges were conducted at beamlines 8-ID (ISS, 5×1013ph s-1@ 10 keV) and 7-BM (QAS), respectively, at NSLS-II at BNL. Given the relatively low X-ray penetration depth at the 3d-metal energy, further attenuated by the CeO2 support, the in situ XAS measurementsCU24182 – 101879.000335 were performed using a Nashner-Adler (N-A) reaction cell sealed by a Kapton window. For each in situ measurement, an appropriate amount of fresh catalyst was pressed into a wafer (13 mm O.D.) and loaded into the N-A cell. The sample was reduced by 50 vol% H2(H2 / N2=20 / 20 ml min-1) at 500 °C for 1 h. Afterwards, the sample was exposed to reaction stream (CO / H2 / N2=15 / 15 / 20 ml min-1) at 450 °C and held at 450 °C for 60 min. The XAS spectra were collected successively at a speed of 2 scans min-1. Fe and Co foils were measured to calibrate the energy shifts as well as to obtain the passive factor (S02) for the extended X-ray absorption fine structure (EXAFS) fittings. Other metal oxides (FeO, Fe3O4, Fe2O3, Co3O4, and CoO) were also measured for the sake of comparison with the X-ray absorption near edge structure (XANES) of different catalysts. For ex situ measurements, the spent catalysts were also pressed into a wafer (13 mm O.D.) and sealed by Kapton tapes. All data processing was performed using the IFEFFIT package.

[0130] Temperature-Programmed Reduction (TPR)

[0131] The TPR measurements were performed using an AMI-300ip (Altamira) instrument. For each test, approximately 50 mg of fresh catalyst was loaded in a U-shape quartz tube and pre-treated with He (50 ml min-1) at 120 °C for 30 min in prior to cooling to 50 °C. Next, the sample was temperature-programmed (10 °C min-1) heated to 850 °C in 10 vol% H2in Ar (50 ml min-1). The hydrogen consumption profile as a function of reduction temperature was recorded via a thermal conductivity detector (TCD), which allowed to compare the reducibility of active metals in different catalysts.

[0132] Raman Spectroscopy

[0133] The structure of the carbon species on the spent catalysts was characterized using a Renishaw inVia Confocal Raman microscope with a 50× objective lens. The spent samples were excited by a 532 nm laser over the range of 270-2015 cm-1. The spectral acquisition consisted of 10 scans with an exposure time of 10 s at 10% of maximum laser power. For the comparison of different samples, all the peak intensities were normalized to the intensity of the D band. To provide a reference for high-purity CNF, a commercial CNF product (C-GR-01-NFB, American Elements) was measured as well. More details can be referred to FIG.19 and Supplementary Table 4.

[0134] Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES)CU24182 – 101879.000335

[0135] The ICP-OES analysis was conducted using an Agilent 5100 ICP-OES instrument on the liquid samples obtained after the separation process to determine the quantities of Fe and Co ions that had been leached out from the spent FeCo / CeO2catalysts.

[0136] Electron Microscopy

[0137] All the electron microscopy measurements were conducted using the facilities at the Center for Functional Nanomaterials (CFN) of BNL. Transmission electron microscopy (TEM) measurements were carried out using a JEOL 2100F (accelerating voltage of 200 kV) instrument. High-angle annular dark-field imaging (HAADF) imaging, energy- dispersive X-ray spectroscopy (EDS) mapping, and selected area electron diffraction (SAED) were carried out using an FEI Talos F200X S / TEM facility (accelerating voltage of 200 kV). Simultaneous high-resolution scanning electron microscopy (SEM) and ADF imaging were measured with a Hitachi 2700C STEM at an accelerating voltage of 200 kV. To prepare the sample, the catalyst powder was ultrasonically dispersed in ethanol for 3 min, after which a droplet was dripped onto a Lacey carbon film (300 mesh) supported on copper grids. After being dried at room temperature, the sample was used for measurements.

[0138] Thermochemical Evaluation

[0139] All the thermochemical experiments were conducted using a fixed-bed flow reactor (quartz tube, 7 mm ID, 9.6 mm OD) and heated by a Thermo Scientific Lindberg / Blue M furnace at ambient pressure. In each experiment, a desired amount of catalyst (60-80 mesh) was loaded into the quartz tube and secured by quartz wool on both sides. The catalyst was pre-reduced in 50 vol% H2 in N2 (total 40 ml min-1) at 500 °C for 1 h before being cooled down to 450 °C in 15 min under the same atmosphere. Subsequently, the catalyst was exposed to the reaction stream of CO / H2 / N2 with desired flow rates (3 / 3 / 4 ml min-1) at 450 °C for different time on stream. The gas line connecting the reactor outlet to the gas chromatography (GC) inlet was maintained at 150 °C using heating tape to prevent the condensation of water vapor. To calibrate water vapor, a stable PtCo3 / CeO2 catalyst was used to perform the reverse water-gas shift (RWGS) reaction at temperatures ranging from 150- 350 °C.54,55This produced a set of water vapor peak areas that were then correlated with their molar fractions, which were calculated based on the oxygen balance. The reactor effluents were analyzed using an Agilent 7890B GC equipped with PLOT Q and MOLESEIVE columns, as well as a TCD and a flame ionized detector (FID). N2was used as a balance gas to tune the inlet fraction of reactants and as an internal standard gas to correct for the volume-CU24182 – 101879.000335 change effect due to the reaction. After the reaction, the catalyst was recollected from the tube reactor and weighed again. Reactant conversion (X), product selectivity (S), catalyst weight gain (CWG), and carbon balance (CB) were calculated as follows: ^ X୧ൌ^,^^ି^^,^౫౪^^,^^ ൈ 100% (1)(2)(3)(4)(5)(6)where F and Fi,outthe reactor inlet andcat,freshspent and fresh catalysts, respectively; ni indicated the number of carbon in gaseous species i; Si in gasand Si in totalcorresponded to the selectivity based on the gaseous C containing products (i.e., CO2 and CH4) and the total conversion of CO, respectively.

[0140] Electrochemical Evaluation

[0141] The CO2 reduction reaction (CO2RR) experiments were carried out using a zero-gap membrane electrode assembly (MEA) electrolyzer (5 cm2, Dioxide Materials) with a stainless-steel cathode flow field and a titanium anode flow field. The cathodic flow field was fed with a CO2 / N2mixture (10 / 10 ml min-1), while the anodic flow field was fed with a50 mM KHCO3 electrolyte at a flow rate of 5 ml min-1using a peristaltic pump. The cathode and anode compartments were separated by an anion exchange membrane (Sustainion X37- 50 Grade RT, Dioxide Materials). The Pd / C cathode was preteated by being scanned between -0.05 V and -1.20 V (50 mV s-1) for 20 cycles. Using linear sweep voltammetry (LSV), the cathode subsequently underwent scanning from 0 V to -3.00 V (20 mV s-1) for two or three times to obtain a stable LSV. The chronopotentiometry tests under different constant current densities (-60, -100, -150, and -200 mA cm-2) were controlled by a BioLogic VSP potentiostat complemented with a BioLogic VMP3B-20 (20 A / 20 V) current booster, and the corresponding evolution of cell voltage was recorded. The cathodic effluents were directly analyzed by the same Agilent 7890B GC. The CO2 conversion, H2 / CO ratio, and carbon balance were calculated based on equations (1) and (2), respectively.CU24182 – 101879.000335 ^ిో ,^ି^ X^୭^ୟ୪ൌమ ^ ిోమ,^౫౪^ిో , ൈ 100% (7)మ ^^(8)(9)where the Xtotal, Xproduct, the CO2 converted intocrossover as well as liquid products, respectively.

[0142] Tandem Electrochemical-Thermochemical Evaluation

[0143] Evaluation of the tandem system was conducted using the same setups and procedures discussed earlier, except that the Fe3Co6 / CeO2 loading for the thermochemical reactor was 200 mg. The electrochemical reactor was carried out at -60, -100, -150, and -200 mA cm-2and the thermochemical reactor was maintained at 450 °C for about 11 h. The gaseous effluents from the MEA electrolyzer outlet were first passed through an ice trap and then directed to the inlet of the thermochemical reactor. Subsequently, all these gaseous effluents were analyzed using the Agilent 7890B GC. The total CO2 conversion, fraction of CO2converted to gaseous products, CWG, and carbon balance were calculated based on equations (7), (8), (5), and (11), respectively. It should be noted that SC(s) in total was not quantified due to the difficulty in simultaneously determining the CO2lost to (bi)carbonate formation and crossover and the CO2 fixed in the solid carbon product.

[0144] CNF Separation and Catalyst Regeneration

[0145] The spent catalyst containing CNF was subject to ultrasonication in 5 mL of concentrated nitric acid (70%, 87.8 mmol) at room temperature for 30 min. The resulting suspension was refluxed under stirring at 120-130 °C for 2, 5, or 12 h. Afterwards, the suspension was cooled to room temperature and centrifuged at 4476 rcf (g) for 10 min; the transparent solution was decanted, while the wet powder was washed with 15 mL of DI water and sonicated for 5 min. The centrifugation and washing cycles were repeated three times. Subsequently, the wet powder was washed with 500 mL DI water and filtered using a PTFE membrane (pore size 0.1 μm), after which it was dried in vacuum at 80 °C overnight. All the solution containing iron and cobalt ions during centrifugation, washing, and filtration was collected and then concentrated at 80 °C for the next cycle of catalyst synthesis. For theCU24182 – 101879.000335 synthesis of regenerated catalysts, an appropriate amount of new CeO2or dried solid products obtained after separation was added to the concentrated solution. The regenerated catalysts were subject to the same reduction treatment and reaction conditions as the fresh catalysts in the initial reactions. Similar methods can be referred to in Refs.56 and 57.

[0146] Density Functional Theory (DFT) Calculations

[0147] Spin polarized DFT calculations were performed by using Vienna ab initio simulation package (VASP). A 400eV kinetic energy cutoff and the projector augmented wave method (PAW) together with GGA exchange-correlation functional plus the PBE functional63were employed. The Van der Waals (VDW) force was addressed by using the DFT-D3 method developed by Grimme et al. Monkhorst-Pack meshes were used to sample the Brillouin zone for all the surface calculations while gamma point was employed for all gas-phase species. The criteria for total energies and forces on all atoms were set as 10−6eV and 0.02 eV Å−1for convergence, respectively. Methfessel-Paxton order one smearing with width 0.2 eV was used to improve the convergence. The Fe (3p, 3d, 4s), Co (3p, 3d, 4p), C (2s, 2p), O (2s, 2p) and H (1s) electrons were treated as valence states, while the remaining electrons were kept frozen as core states. The climbing image nudged elastic band method (CI-NEB) for each elementary reaction intermediates were conducted to derive the transition states. All the transition states have been verified with only one imaginary frequency by using harmonic approximation.

[0148] For the metallic bulk phases, body centered cubic (bcc) Fe and FeCo bulk phases and hexagonal close packed (hcp) Co bulk phase were used for bulk DFT calculations, where the optimized lattice parameters and the average magnetic moments are shown in FIG. 27, consisting with the experimental values. The most stable surfaces for bcc and hcp structures were adopted accordingly. Specifically, 3×3 Fe(110), 3×3 FeCo(110), and 4×4 Co(0001) surfaces with four layers (FIG.28) were used for surface DFT calculations with Monkhorst-Pack meshes of 3×2×1, 3×2×1, and 3×3×1, respectively. As for the metal carbides, the Cementite-derived structures with orthorhombic system were adopted for Fe3C, Co3C, Fe2CoC bulk phases, where the optimized lattice parameters and the magnetic moments per unit are also provided in FIG.29, consisting with the literature reports. According to previous studies on the stable surface and termination for metal carbides with Cementite-derived structures, 2×1 Fe3C(001), 2×1 Co3C(001) and 2×1 Fe2CoC(001) surfaces were adopted with Monkhorst-Pack meshes of 2×3×1 (FIG.30). Fe4 / CeO2(111),CU24182 – 101879.000335 Fe2Co2 / CeO2(111), and Co4 / CeO2(111) were used as simplified models to capture the metal- CeO2 interaction at the interface (FIG.16). The Hubbard-like U term with a value of 4.5 was applied for highly correlated Ce 4f states78. A Gaussian smearing with a width of 0.05 eV was used to improve the convergence for all the CeO2(111)-based models. The bottom O-Ce-O repetitive unit was fixed for all the CeO2(111)-based calculations. The bottom two layers were fixed for the metallic surfaces and the bottom two Carbon-Metal-Metal-Carbon (C-M- M-C) repetitive units were fixed for the metal carbides while the rest of the atoms were allowed for full relaxation. The adsorption energy of adsorbate on each model surface was calculated as: Eads = E(Adsorbate / Surface) − E(Surface) − E(Adsorbate).

[0149] Ab-initio Thermodynamic (AITD) Analysis

[0150] To explore the stabilities of metallic Fe, Co, FeCo alloy, and the corresponding carbides formation under the experimental reaction conditions, ab initio thermodynamics analysis was utilized, where two reactions (R1 and R2) were considered as the extreme cases for the experimental processes without or with H2gas in the CO gas environment. Detailed derivations for the AITD analysis on monometallic, bimetallic, and carbides catalysts can be found elsewhere herein.

[0151] Supplementary Methods

[0152] Ab-initio Thermodynamic (AITD) Analysis

[0153] To explore the stabilities of metallic Fe, Co, FeCo alloy, and the corresponding carbides formation under the experimental reaction conditions, ab initio thermodynamics analysis was utilized.

[0154] Two reactions were considered as the extreme cases for the experimental processes without or with H2 gas in the CO gas environment. Reaction one (Supplementary Equation 1) is Boudouard Reaction. Reaction two (Supplementary Equation 2) is hydrogen- assisted reaction.

[0155] CO(g) + CO(g) ^ CO2(g) + C(s) (Supplementary Equation 1)

[0156] CO(g) + H2(g) ^ H2O(g) + C(s) (Supplementary Equation 2)

[0157] During the Boudouard Reaction (Supplementary Equation 1), the Gibbs reaction energy can be described as (Supplementary Equation 3):CU24182 – 101879.000335

[00158] ∆G^T, P^ ൌ μ^େ ^ μ ^େ^మ െ μ ^େ^ െ μ^େ^(Supplementary Equation 3)

[0159] Where μ^is thepotential of solid C; μ^େ^is the chemical of gas phase CO; μ^is the chemical potential of gas phase The chemical potentialgas phase CO can be described as (Supplementary 4) and Equation 5),

[0160] μ^େ^ ^T, P^ ൌ μ^େ^ ^ ∆μେ^^T, P^ (Supplementary Equation 4)

[0161] in which μ^େ^ is the standard chemical potential of CO and ∆μେ^^T, P^ isthe chemical potential change with respect to the corresponding temperature (T) and pressure (P) variations.

[0162] μ^େ^మ ^T, P^ ൌ μ ^େ^మ ^ ∆μେ^మ^T, P^ (Supplementary Equation5)

[0163] in which μ^େ^మ is the standard chemical potential of CO2 and ∆μେ^మ^T, P^ isthe chemical potential change of CO2with respect to the corresponding T and P variations.

[0164] Note that the Gibbs reaction energy equals zero (Supplementary Equation 6) under the equilibrium state:

[00165] ∆G^T, P^ ൌ 0 (Supplementary Equation 6)

[0166] Thus, the chemical potential of solid C at the equilibrium of reaction one can be derived from (Supplementary Equation 3) and (Supplementary Equation 6) as described in (Supplementary Equation 7):

[0167] μ^ ൌ μ ^ ^ μ ^ െ μ^େେ^ େ^ େ^మ(Supplementary Equation 7)

[0168] Bypotential of gas phase CO (Supplementary Equation 4) and CO2 (Supplementary Equation 5) into Supplementary Equation 7, the chemical potential of solid C at the equilibrium of reaction one can be further described as (Supplementary Equation 8):

[0169] μ^ ^େ ൌ 2 ൈ μେ^ െ μ ^େ^మ ^ 2 ൈ ∆μେ^^T, P^ െ ∆μେ^మ^T, P^(Supplementary Equation 8)

[0170] In consideration of the fact that the standard chemical potential can be obtained from DFT calculations (μ0= EDFT) and the chemical potential change of can be acquired from the NIST-JANAF Thermochemical Tables,4the chemical potential of solid C at the equilibrium of reaction one can be rewritten as (Supplementary Equation 9):CU24182 – 101879.000335

[0171] μ^େ ൌ μ ^େ^భ ^ ∆μୖ^େ ^T, P^ (Supplementary Equation 9)

[0172] ^ 2 ൈ μ^େ^ െ μ^େ^మ, the standard chemical potential ofCO and CO2in and ∆μୖ^େ ^T, P^ represents 2 ൈ ∆μେ^^T, P^ െ∆μେ^మ^T, P^, the chemical potential change of CO and CO2 in the reaction environments withrespect to the variations of T and P.

[0173] Similarly, the chemical potential of solid carbon at the equilibrium of reaction two can be described as (Supplementary Equation 10):

[0174] μ^ ൌ μ ^ ^ μ ^ െ^େେ^ ୌమ μୌమ^(Supplementary Equation 10)

[0175] ^ ^the chemical potential of gas phase H2and H2O,respectively.

[0176] The chemical potential of H2 can be described as (Supplementary Equation 11):

[0177] μ^ ^T, P^ ^ୌమ ൌ μୌమ ^ ∆μୌమ^T, P^ (Supplementary Equation 11)

[0178] Where μ^ୌమ is the standard chemical potential of H2 and ∆μୌమ^T, P^ is thechemical potential change of H2with respect to the variations of T and P.

[0179] And the chemical potential of H2O can be described as (Supplementary Equation 12)

[0180] μ^ୌమ^ ^T, P^ ൌ μ ^ୌమ^ ^ ∆μୌమ^^T, P^ (SupplementaryEquation 12)

[0181] Where μ^ୌమ^ is the standard chemical potential of H2O and ∆μୌమ^^T, P^ isthe chemical potential change of H2O with respect to the variations of T and P.

[0182] By substituting (Supplementary Equation 4), (Supplementary Equation 11), and (Supplementary Equation 12) into (Supplementary Equation 10), the chemical potential of solid C at the equilibrium of reaction two can be derived as (Supplementary Equation 13)

[0183] μ^େ ൌ μ^େ^ ^ μ ^ୌమ െ μ ^ୌమ^ ^ ∆μେ^^T, P^ ^ ∆μୌమ^T, P^ െ ∆μୌమ^^T, P^(Supplementary

[0184] Furthermore, (Supplementary Equation 13) can be simplified as (Supplementary Equation 14):

[0185] μ^ ^ ୖଶେ ൌ μେ^మ ^ ∆μେ ^T, P^ (Supplementary Equation 14)CU24182 – 101879.000335

[0186] Where μ^represent^ ^^େ^మs μେ^ ^ μୌమ െ μୌమ^, the standard chemical potential of CO, H2, and H2O in the reaction environments^T, P^ represents ∆μେ^^T, P^ ^∆μ ^T, P^ െ ∆μ ^T, P^, the chemical potentiୌమ ୌమ^ al CO, H2, and H2O in the reactionenvironments with respect to the variations of T and P.

[0187] In addition, the chemical potential changes for CO, CO2, H2, and H2O in (Supplementary Equation 4), (Supplementary Equation 5), (Supplementary Equation 11), and (Supplementary Equation 12) with respect to the variations of T and P can be described as (Supplementary Equation 15), (Supplementary Equation 16), (Supplementary Equation 17), and (Supplementary Equation 18) respectively.

[00188] ∆μେ^^T, P^ ൌ ∆μେ^^T, P^^ ^ RT ln^Pେ^⁄ P^େ^ ^ (SupplementaryEquation 15)

[00189] ൌ Hେ^^T, P^^ െ Hେ^^0K, P^^ െ T^Sେ^^T, P^^ െ Sେ^^0K, P^^^ ^RT ln^Pେ^⁄ P^େ^ ^

[0190] ∆μ ^T, P^ ൌ ∆μ ^T, P^^ ⁄ ^େ^మ େ^మ ^ RT ln^Pେ^మ Pେ^మ ^(Supplementary Equation 16)

[00191] ൌ H ^େ^మ^T, P ^ െ Hେ^మ^0K, P^^ െ T^Sେ^మ^T, P^^ െ Sେ^మ^0K, P^^൧ ^RT ln൫Pେ^మ⁄ P ^େ^మ ൯

[0192] ∆μ ^T, P^ ൌ 1⁄ 2 ∆ ^ ^^ ⁄ ^ୌ μୌమ^T, P^ ൌ 1 / 2^∆μୌమ T, P ^ RT ln^Pୌమ P ^^

[0193] (Supplementary Equation 17)

[00194] ൌ 1⁄ 2 ^Hୌమ^T, P^^ െ Hୌమ^0K, P^^ െ T^Sୌమ^T, P^^ െ Sୌమ^0K, P^^൧ ^RT ln൫Pୌమ⁄ P ^ୌమ ൯൧

[0195] ∆μୌమ^^T, P^ ൌ ∆μୌమ^^T, P^^ ^ RT ln^Pୌమ^⁄ P ^ୌమ^ ^ (SupplementaryEquation 18)

[00196] ൌ Hୌమ^^T, P^^ െ Hୌమ^^0K, P^^ െ T^Sୌమ^^T, P^^ െ Sୌమ^^0K, P^^൧ ^RT ln

[00197] Where the Hେ^^T, P^^ and Sେ^^T, P^^ represent for the enthalpy andentropy of CO at the standard pressure (1.0 bar) and temperature of T, respectively;Hେ^^0K, P^^ and Sେ^^0K, P^^ represent for the enthalpy and entropy of CO at 0K,respectively; Pେ^is the pressure of CO in the reaction environments while Pେ^^ is the standard pressure (1.0 bar) of CO; Similar meanings of enthalpy and entropy at different P and T canCU24182 – 101879.000335 also apply for CO2, H2and H2O in (Supplementary Equation 16), (Supplementary Equation 17), and (Supplementary Equation 18); R is the ideal gas constant.

[0198] The metallic bulks of Fe, Co, and FeCo alloy and the carbide bulks of Fe3C, Co3C and Fe2CoC and FeCo2C were considered for their stability. The following reactions and methods were adopted for studying the formation process of carbide bulks from metallic bulk under the reaction conditions.

[0199] The formation of Fe3C bulk from metallic Fe bulk and its corresponding Gibbs reaction energy can be described as (Supplementary Equation 19) and (Supplementary Equation 20):

[0200] 3Fe(s) + C(s) ^ Fe3C(s) (Supplementary Equation 19)

[00201] ∆G^FeଷC, T, P^ ൌ μ ^^^యେ െ 3 ൈ μ^^^ െ μ^େ (Supplementary Equation 20)

[0202] Where μ^^^యେis the chemical potential of Fe3C bulk, approximated by the standard chemical potential of Fe3C, μ^^^^యେ; μ^^is the chemical potential of Fe bulk, approximated by the standard of Fe, μ^^^; μ^is the chemical potential of Cେin the reaction environments, which can be calculated from (Supplementary Equation 9) or (Supplementary Equation 14) for the reaction conditions without (Supplementary Equation 20A) or with H2(Supplementary Equation 20B) in the CO gas environment.

[00203] ∆G^FeଷC, T, P^ ൌ μ ^^^యେ െ 3 ൈ μ^^^ െ μ ^େ^భ െ ∆μେ^భ^T, P^

[00205] ∆G^FeଷC, T, P^ ൌ μ ^^^యେ െ 3 ൈ μ^^^ െ μ ^େ^మ െ ∆μେ^మ^T, P^

[0206] (Supplementary Equation 20B)

[0207] Similarly, the formation of Co3C bulk from metallic Co bulk and its corresponding Gibbs reaction energy can be described as (Supplementary Equation 21) and (Supplementary Equation 22):

[0208] 3Co(s) + C(s) ^ Co3C(s) (Supplementary Equation 21)

[00209] ∆G^CoଷC, T, P^ ൌ μ ^ ^େ୭^యେ െ 3 ൈ μେ୭ െ μେ(Supplementary Equation 22)CU24182 – 101879.000335

[0210] Where μ^େ୭యେis the chemical potential of Co3C bulk, approximated by the standard chemical potential of Co3C, μ^େ୭^యେ; μେ୭is the chemical potential of Co bulk, approximated by the standard chemical potential of Co, μ^େ୭; By substituting (Supplementary Equation 9) or (Supplementary Equation 14) into (Supplementary Equation 22), the Gibbs reaction energy for the formation of Co3C bulk from metallic Co bulk in the two different reaction environments can be described as (Supplementary Equation 22A) and (Supplementary Equation 22B).

[00211] ∆G^Co ^ଷC, T, P^ ൌ μ ^ ^େ୭యେ െ 3 ൈ μେ୭ െ μେ^భ െ ∆μେ^భ^T, P^μେ୭యେ େ୭ μେ^మ P^

[0214] (Supplementary Equation 22B)

[0215] The formation of Fe3C bulk form FeCo alloy bulk and its corresponding Gibbs reaction energy can be described as (Supplementary Equation 23) and (Supplementary Equation 24):

[0216] 3FeCo(s) + C(s) ^ Fe3C(s) +3Co(s) (Supplementary Equation 23)

[00217] ∆G^FeଷC, T, P^ ൌ μ ^^^యେ ^ 3 ൈ μ^େ୭ െ 3 ൈ μ ^^^େ୭ െ μ^େ (Supplementary Equation 24)[0 Where μ^^େ୭is the chemical potential of FeCo alloy bulk, approximated by the standard chemical potential of FeCo alloy, μ^^^యେ; By substituting (Supplementary Equation 9) or (Supplementary Equation 14) into (Supplementary Equation 24), the Gibbs reaction energy for the formation of Fe3C bulk from FeCo alloy bulk in the two different reaction environments can be described as (Supplementary Equation 24A) and (Supplementary Equation 24B).

[00219] ∆G^FeଷC, T, P^ ൌ μ ^ ^ ^ ^^^యେ ^ 3 ൈ μେ୭ െ 3 ൈ μ^^େ୭ െ μେ^భ െ ∆μେ^భ^T, P^

[00221] ∆G^Fe ^ ^ ^ଷC, T, P^ ൌ μ^^యେ ^ 3 ൈ μ ^େ୭ െ 3 ൈ μ^^େ୭ െ μେ^మ െ ∆μେ^మ^T, P^

[0222] (Supplementary Equation 24B)

[0223] The formation of Co3C bulk from FeCo alloy bulk and its corresponding Gibbs reaction energy can be described as (Supplementary Equation 25) and (Supplementary Equation 26):CU24182 – 101879.000335

[0224] 3FeCo(s) + C(s) ^ Co3C(s) +3Fe(s) (Supplementary Equation 25)

[00225] ∆G^Co ^ ^ ^ଷC, T, P^ ൌ μେ୭^యେ ^ 3 ൈ μ^^ െ 3 ൈ μ^^େ୭ െ μେ(Supplementary Equation 26)

[0226] Equation 26A) and (Supplementary Equation 26B) in the two different reaction environments by substituting (Supplementary Equation 9) or (Supplementary Equation 14) into (Supplementary Equation 26).

[00227] ∆G^CoଷC, T, P^ ൌ μ ^ ^ ^ ^େ୭యେ ^ 3 ൈ μ^^ െ 3 ൈ μ^^େ୭ െ μେ^భ െ ∆μେ^భ^T, P^

[0228] (Supplementary Equation 26A)

[00229] ∆G^CoଷC, T, P^ ൌ μ ^େ୭యେ ^ 3 ൈ μ^^^ െ 3 ൈ μ ^^^େ୭ െ μ ^େ^మ െ ∆μେ^మ^T, P^

[0230] (Supplementary Equation 26B)

[0231] The formation of Fe2CoC bulk from FeCo alloy bulk and its corresponding Gibbs reaction energy can be described as (Supplementary Equation 27) and (Supplementary Equation 28):

[0232] 2FeCo(s) + C(s) ^ Fe2CoC(s) +Co(s) (Supplementary Equation 27)

[00233] ∆G^FeଶCoC, T, P^ ൌ μ ^^^మେ୭େ ^ μ^େ୭ െ 2 ൈ μ ^^^େ୭ െ μ^େ (Supplementary Equation 28)

[0234] Where μ^^^మେ୭େis the chemical potential of Fe2CoC bulk, approximated by the standard chemical potential of Fe2CoC, μ^^^మେ୭େ; By substituting (Supplementary Equation 9) or (Supplementary EquationEquation 28), the Gibbs reaction energy for the formation of Fe2CoC bulk from FeCo alloy bulk in the two different reaction environments can be described as (Supplementary Equation 28A) and (Supplementary Equation 28B).

[00235] ∆G^FeଶCoC, T, P^ ൌ μ ^ ^ ^ ^^^మେ୭େ ^ μେ୭ െ 2 ൈ μ^^େ୭ െ μେ^భ െ ∆μେ^భ^T, P^

[0236] (Supplementary Equation 28A)

[00237] ∆G^FeଶCoC, T, P^ ൌ μ ^ ^ ^ ^^^మେ୭େ ^ μେ୭ െ 2 ൈ μ^^େ୭ െ μେ^మ െ ∆μେ^మ^T, P^

[0238] (Supplementary Equation 28B)CU24182 – 101879.000335

[0239] The formation of FeCo2C bulk from FeCo alloy bulk and its corresponding Gibbs reaction energy can be described as (Supplementary Equation 29) and (Supplementary Equation 30):

[0240] 2FeCo(s) + C(s) ^ FeCo2C(s) +Fe(s) (Supplementary Equation 29)

[00241] ∆G^FeCoଶC, T, P^ ൌ μ ^^^େ୭మେ ^ μ^^^ െ 2 ൈ μ ^^^େ୭ െ μ^େ (Supplementary Equation 30)

[0242] Where μ^^^େ୭మେis the chemical potential of FeCo2C bulk, approximated by the standard chemical potential of FeCo2C, μ^^^େ୭మେ; By substituting (Supplementary Equation 9) or (Supplementary Equation 14) into (Supplementary Equation 30), the Gibbs reaction energy for the formation of FeCo2C bulk from FeCo alloy bulk in the two different reaction environments can be described as (Supplementary Equation 30A) and (Supplementary Equation 30B).

[00243] ∆G^FeCoଶC, T, P^ ൌ μ ^^^େ୭మେ ^ μ^^^ െ 2 ൈ μ ^^^େ୭ െ μ ^େ^భ െ ∆μେ^భ^T, P^

[00245] ∆G FeCoଶC, T, P^ ൌ μ^^େ୭మେ ^ μ^^ െ 2 ൈ μ^^େ୭ െ μେ^మ െ ∆μେ^మ^T, P^

[0246] (Supplementary Equation 30B)

[0247] The following reactions were considered for the potential to stay as metallic states of Fe (Supplementary Equation 31 and Supplementary Equation 32), Co (Supplementary Equation 33 and Supplementary Equation 34), and FeCo alloy (Supplementary Equation 35 and Supplementary Equation 36) bulks, where the corresponding Gibbs reaction energy equals zero.

[0248] Fe(s) + C(s) ^ Fe(s) +C(s) (Supplementary Equation 31)

[00249] ∆G^Fe, T, P^ ൌ 0.00 (Supplementary Equation 32)

[0250] Co(s) + C(s) ^ Co(s) +C(s) (Supplementary Equation 33)

[00251] ∆G^Co, T, P^ ൌ 0.00 (Supplementary Equation 34)

[0252] FeCo(s) + C(s) ^ FeCo(s) +C(s) (Supplementary Equation 35)

[00253] ∆G^FeCo, T, P^ ൌ 0.00 (Supplementary Equation 36)CU24182 – 101879.000335

[0254] In addition, the following reaction (Supplementary Equation 37) was considered for the formation of metallic Fe and Co bulks from FeCo alloy bulk and its Gibbs reaction energy (Supplementary Equation 38).

[0255] FeCo(s) + C(s) ^ Fe(s) + Co(s) +C(s) (Supplementary Equation 37)

[00256] ∆G^FeCo, T, P^ ൌ μ^ ^^^ ^ μେ୭ െ μ^^^େ୭(Supplementary Equation 38)

[0257] Energy Cost and CO2 Emission Calculations

[0258] In the context of the electrocatalytic (EC)-thermocatalytic (TC) tandem scheme, we separately assessed the energy requirements for the electrocatalytic and thermocatalytic steps. To exemplify this, we determined the energy consumption for a laboratory-scale EC-TC system, which employed a 10%Pd / C cathode in the electrolyzer and operated the reaction at 450 °C in the subsequent thermocatalytic reactor under ambient pressure.

[0259] Energy calculation for the electrocatalytic step

[0260] The electrochemical energy cost was calculated as

[00261] Q^େ ൌ I ൈ V ൈ t ൌ 0.75 A ൈ 3.07 V ൈ 11 h ൌ 2.30 W ൈ 11 h ൌ91.18 kJ

[0262] (Supplementary Equation 39)

[0263] where I, V, and t referred to the current, voltage, and time, respectively.

[0264] Energy calculation for the thermocatalytic step

[0265] The thermochemical energy cost (QTC) was calculated based on an isothermal reactor model, which took into account of several contributions: 1) the heat used to preheat the gaseous effluents from the upstream electrolyzer (QGas) and the catalyst bed in the downstream thermocatalytic reactor (QCatal) to the desired reaction temperature, 2) the heat required to compensate heat loss through the reactor walls (QFurnace) to the environment, and 3) the reaction heat (QRxn) associated with the reactions taking place within the thermocatalytic reactor. It should be noted that the furnace or reactor was assumed to be preheated (with the associated energy not considered) and treated as a heat reservoir, but with a gradual heat loss. As a result, QTC was calculated as follows:

[00266] Q^େ ൌ Qୋୟ^ ^ Qେୟ^ୟ୪ ^ Q^^୰୬ୟୡ^ ^ Qୖ^୬ (SupplementaryEquation 40)CU24182 – 101879.000335

[0267] The energy required to preheat the gaseous effluents from the cathode of the electrolyzer was calculated as

[00268] Qୋୟ^ ൌ ∑୧ Q୧ ൌ ∑^ ୧C୧m୧∆T ൌୖ^ൈ ൫Cେ^Vେ^Mେ^ ^ Cେ^మVେ^మMେ^మ ^^ ^ ^flow rate, and molar weight of gaseous species i, respectively; P, R, and T were the pressure (1 atm), universal gas constant (0.0821 L atm K-1mol-1), and temperature (298 K), respectively; ΔT (425 K) was the difference between the reaction temperature (723 K) and room temperature (298 K).

[0270] The specific heat capacity (Ci) and molar weight (Mi) of each species used for the calculation are as below:

[0271] CO (846 J kg-1K-1; 44.0 g mol-1)

[0272] CO2(1040 J kg-1K-1; 28.0 g mol-1)

[0273] H2 (14310 J kg-1K-1; 2.0 g mol-1)

[0274] H2O (4182 J kg-1K-1; 18.0 g mol-1)

[0275] CH4(2226 J kg-1K-1; 16.0 g mol-1)

[0276] N2 (1040 J kg-1K-1; 28.0 g mol-1)

[0277] The energy required to preheat the thermocatalyst bed to the reaction temperature was calculated as follows:

[00278] Q ି^ ି^େୟ^ୟ୪ ൌ Cେୟ^ୟ୪mେୟ^ୟ୪∆T ൌ 390 J kg K ൈ 200 mg ൈ 425 K ൌ 0.03 kJ

[0279] (Supplementary Equation 42)

[0280] where CCatal and mCatal were the specific heat capacity (assumed being dominated by CeO2, 390 J kg-1K-1) and mass (200 mg) of the thermocatalyst bed, respectively.

[0281] The energy needed to offset heat loss through the walls of the cylindrical reactor or furnace to the environment was calculated based on the Fourier’s Law of heat conduction.

[00282] Q ଶ^୩^^^భି^మ^^^୰୬ୟୡ^ ൌ Qୌ^ୟ^ ୪୭^^ ୧୬ ^ୟ୪୪^ ൌ ൈ t ൌଶ^ൈ^.ଶ^ ^ ୫షభ^షభൈ^.^^ ୫ൈ^^ଶଷ ^ିଶଽ଼ ^^ ୪୬^^ ୫ / ^.^^ ୫^ൈ 11 h(SupplementaryEquation 43)CU24182 – 101879.000335

[0283] where the k and L referred to the heat conductivity of the insulation material (assuming fiber hardboard, 0.20 W m-1K-1) and the length of the catalyst bed (0.01 m), respectively; r1(assuming 1 m)5and r2(assuming 0.01 m) represented the radius of the inner and outer surface of the cylinder, respectively.

[0284] The energy released or consumed by the reactions (as listed below) occurring in the thermocatalytic reactor was calculated as follows:

[00285] Qୖ^୬ ൌ ∑୨ Q୨ ^ Qୖ^ ൌ nେ^^^ ൈ ∆Hୖ^୫ి^౩^^,^ଶଷ^ൌ^^ి ൈ ∆Hୖ^,^ଶଷ^ൌ ଶସ^ ^ଶ ^mC(s) was the mass of the solid carbon products, which wasweight gain (240 mg). Note: Given the limited presence of generated water vapor, the Boudouard reaction (R1) should be the dominant reaction within the thermocatalytic reactor. Thus, the energy release within the reactor was estimated based on R1.

[0287] R1: 2CO(g)→CO2(g)+C(s), ΔH0R1,723K= -159.9 kJ mol-1

[0288] R2: CO(g)+H2(g)→H2O(g)+C(s), ΔH0R2,723K= -122.1 kJ mol-1

[0289] R3: CO(g)+3H2(g)→H2O(g)+CH4(g), ΔH0R3,723K= -220.3 kJ mol-1

[0290] R4: CO2(g)+H2(g)→H2O(g)+CO(g), ΔH0R4,723K= +37.7 kJ mol-1

[0291] Based on the above calculations, the total energy cost to produce one kilogram of CNF can be calculated as follows:

[0292] E ^ుిା^^ి ଽ^.^଼ ୩^ା^ହ.^ହା^.^ଷାସହ.ଽ^ିଷ.ଶ^^ ୩^ ଽ^.^଼ ୩^ାସ଼.ସ଼ ୩^େ^^ൌ ୫ిొూ ൌଶସ^ ୫^ൌଶସ^ ୫^ ൌ

[0293] The corresponding gross CO2emission associated with the energy cost can be calculated based on the current CO2 emission factor (ε) of 0.37 kgCO2 kWh-1. ^ 00294] γ ൌ ^ ା ^ [ుి^^ిൈகൌ 161.63 kWh kg ି^ ൈ 0.3 ି^େ^ଶ,^୰୭^^ େ^^ 7 kgେ^ଶ kWh ൌ59.80

[0295] Furthermore, the net CO2 emission can be calculated by subtracting the amount of carbon fixed into the CNF products.

[00296] γ ൌ γ െ n ൈସସ.^ ^ిోమ୫୭୪షభ^ γ^ ୫୭୪ େ ୧୬ ^ ୫୭୪ େ^^ ^ଶ.^ ^ిొూ୫୭୪ 47)CU24182 – 101879.000335

[0297] Given that the electrolyzer can be directly integrated with renewable sources of electricity such as solar, wind, hydropower, and geothermal energy, the net CO2 emission was also calculated specifically for the downstream thermocatalytic reactor.

[0298] γᇱେ^ଶ,୬^^ ൌ ^^ిൈக୫ిొూ െ ସସ.^^ଶ.^ൌ 17.16 kgେ^ଶ kgିେ^^^ (Supplementary Equation 48)

[0299] Assumingall the EC and TC processes can be directly driven by renewable energy sources, the theoretically ideal net CO2emissions per kilogram of CNF should be:

[00300] γେ^ଶ,୬^^,୧^^ୟ୪ ൌ െସସ.^ ^ଶ.^ൌ െ3.67 kgେ^ଶ kgିେ^^^ (Supplementary Equation 49)

[0301] Based on the energy cost, to achieve a zero-net CO2emission, the CO2 emission factor should be reduced by 5-20 times to as calculated

[00302] εୡ୰୧^୧ୡୟ୪ ൌ ^44.0 / 12.0^ / ^^ుిା^^ి ^ ൌ 3.67 kgେ^ଶ kgିେ^^^ / 50)

[0303] εᇱୡ୰୧^୧ୡୟ୪ ൌ ^44.0 / 12.0^ / ^ ిି^ ି^୫ిొూ ^ ൌ 3.67 kgେ^ଶ kgେ^^ / 48.48 kWh kgେ^^ൌ 0.08 kgେ^ଶkWhି^(Supplementary Equation 51)

[0304] Supplementary Notes

[0305] Supplementary Note 1: Catalytic behaviors at metal-CeO2 interfaces

[0306] The high-angle annular dark field (HAADF) imaging and energy-dispersive X-ray spectroscopy (EDS) mapping results (Figures 1h-j) indicated that the formation of CNF primarily follows a tip-growth mechanism, which pushes the metal nanoparticles away from the support and places them within the tip of CNF. Thus, the metal-support interface interaction should mainly play a role in the initial reduction of Fe and Co oxide species to form the above-mentioned metallic structures and / or the initial activation of CO and H2molecules. Temperature-programmed reduction (TPR) measurements indicated that Fe and Co species supported on CeO2can be fully reduced below 500°C, while those supported on γ-Al2O3 and SiO2 were only partially reduced after the 500°C-reduction. This decreased reducibility is likely responsible for their compromised performance in solid carbon formation.

[0307] Additional DFT calculations to explore the effect of the metal oxide / metal interface and validate the use of the metal slab models to describe the CeO2-supported metalCU24182 – 101879.000335 nanoparticles synthesized experimentally. The adoption of metal slab models in the DFT calculations is based on the consideration of the large metal particle size (20-30 nm) observed experimentally (FIG.22), which limits the area of interfacial regions and likely their catalytic contributions. Previously, similar slab models were found to describe well the experimentally observed trend in activity and selectivity for a wide range of monometallic and bimetallic nanoparticles supported by ceria according to our previous studies of CO2 reduction.6–8For the activation of CO in the current case, our DFT results show that the contribution from metal-CeO2 interfaces to the overall catalytic activity can be more limited due to the lower activity in C-O bond scission than that on the corresponding metal surfaces. Specifically, DFT calculations were performed using Fe4 / CeO2(111), Fe2Co2 / CeO2(111), and Co4 / CeO2(111) as simplified interface models to capture the metal-CeO2 interfaces (FIG. 20A). Compared to the corresponding metal surfaces, CO still prefers the atop site of Co or Fe at the interfaces (FIG.20B), while both Fe and Co at the interface are more oxidized and with lower coordination numbers, resulting in the binding to CO being slightly stronger (Supplementary Table 1). However, the adsorption of dissociated *C and *O fragments at these oxidized Fe and Co sites is less stable than that on the corresponding metal surfaces, thus the dissociation of CO is less favorable. This suggests that the low-coordinated sites of metal nanoparticles at the interface bind CO too strongly to enable subsequent activations, which likely leads to CO poisoning. In consideration of the fact that CO dissociation is a key step for initiating the CNF growth, the contribution from the metal-CeO2interfaces is likely limited.

[0308] Supplementary Note 2: Role of Fe3C

[0309] According to the DFT calculations, the CO dissociation on Fe3C(001) is highly unlikely due to the high activation energy (Ea= 2.49 eV, Supplementary Table 8). Thus, the transition from metallic Fe to Fe3C under the reaction environment likely leads to the deactivation of Fe catalyst for CO dissociation, which is a key step to initiate the overall CNF growth. Without the dissociation of CO to produce *C and *O, the subsequent C-C bond coupling, as well as the removal of *O fragment from the Fe3C surface are not viable. However, the DFT results (FIG.52) indicate that in the presence of *C and *O, the Fe3C surface can enable the C-C bond formation via a thermodynamically favorable process (^E = -0.72 eV). In addition, the removal of *O fragment via the formation of CO2(^E = 0.81 eV) or hydrogenation to *OH (^E = 0.43 eV) and sequentially *H2O (^E = 0.81 eV) is also notCU24182 – 101879.000335 difficult. Overall, the effect of Fe3C formation under reaction conditions on the CNF growth is negative by hindering the initial CO bond dissociation.

[0310] Supplementary Note 3: Role of additional metallic Co

[0311] DFT calculations demonstrate that the FeCo alloy surface can dissociate CO with an activation energy (Ea) of 1.64 eV and reaction energy (^E) of -0.19 eV and initiate the CNF growth by enabling the C-C bond formation under reaction conditions (^E = 0.26 eV, Ea= 1.13 eV). In comparison, Co alone is less active to dissociate CO (^E = 0.82 eV, Ea= 2.28 eV), but more active for the C-C bond formation (^E = -0.62 eV, Ea= 0.96 eV). Thus, the coexistence of Co with FeCo likely takes advantage of CO dissociation on FeCo alloy and C-C coupling of the produced *C on Co. Such synergy is ensured by the favorable diffusion of *C species from FeCo to Co. According to the DFT calculations (Supplementary Table 9 and FIG.30), the produced *C species is mobile on the FeCo alloy surface due to the more preferred diffusion (^E = 0.11 eV, Ea = 1.09 eV) than the C-C coupling. Note that the transition state (TS) for *C diffusion on FeCo(110) requires breakage of two carbon-metal bonds by shifting from 4-fold hollow to two-fold bridge site, thus resulting in a higher barrier than the diffusion energy (FIG.41). In contrast, the TS for *C diffusion on Co(0001) can maintain a hollow-like motif similar to the initial state, thus rendering a lower barrier that is close to the corresponding diffusion energy (Supplementary Figure 37). Consequently, the spillover of *C from the FeCo alloy to the Co-rich surface is likely feasible due to the small difference in binding of *C between FeCo and Co (0.66 eV, Supplementary Table 7) or low diffusion energy, and thus a similarly low diffusion barrier. Finally, the facile diffusion of *C on the Co surface (^E = 0.30 eV, Ea= 0.32 eV) should help facilitate C-C bond formation and thus CNF growth. Overall, CNF growth is feasible on the FeCo alloy surface, which can be further promoted by the coexistence of additional Co via feasible *C diffusion from FeCo to Co and thus an accelerated C-C bond formation.

[0312] Supplementary Note 4: Ensemble, electronic, and strain effects of FeCo Alloy

[0313] The formation of FeCo alloy displayed different binding behaviors to the reaction intermediates involved in the syngas conversion from each parent metal. Specifically, the C-anchored species tended to bind at Co site and the O-anchored species tended to bind at Fe or Fe-rich site (FIGs.43, 44, and 45; Supplementary Tables S7 and S8). For the binding of CO, the top site of Fe(110) provides the strongest binding (Eb = -2.17 eV),CU24182 – 101879.000335 followed by the top site of Co on FeCo(110) surface (Eb= -2.09 eV) and the hcp site of Co on Co(0001) (Eb = -1.89 eV) in a decreasing sequence. By comparison the top site of Fe on FeCo(110) (Eb= -1.89 eV) is less active than Co.

[0314] Such selective bindings can be attributed to the ensemble and electronic effects introduced by forming FeCo alloy while the strain effect is trivial (FIGs.46, 47, and 48, Supplementary Tables S9 and S10). Upon the alloying of Fe with Co, the overall surface Fe 3d states shifted away from the Fermi level with d band center (εd) decreased from -1.79 eV to -2.31 eV. That is, the Fe site is deactivated by forming FeCo alloy. This is due to the contribution from electronic effect, which involved the alloying-introduced spin transfer between the spin-up and spin-down d-band of Fe, as observed previously.9The down-shifted εd of Fe resulted in the weakened Fe-CO binding on FeCo(110) as compared to that on Fe(110) (Supplementary Table 9, FIG.46). Differently, the Co site is activated, where the shift in Co 3d states by alloying with Fe was toward the Fermi level (εd: from -2.00 eV to - 1.94 eV) (Supplementary Table 9, FIG.42). Although the magnitude of εdshift for Co was not as significant as that of Fe, more drastic variation in profile of Co 3d states was observed going from delocalized states as seen for Co(0001) to localized states. This is attributed to the ensemble effect, where the nearest neighbor coordination number of Co varied from 9 to 6 and the nearest neighbor environments of Co varied from Co to Fe upon alloying with Fe to a bcc FeCo alloy structure (Supplementary Table 9, FIG.43). As a result, the Co site on FeCo(110) can bind CO more strongly than Co(0001) to facilitate the activation of CO. Yet, it is still not as strong as that of Fe(110). Note that the contributions from strain are rather small (~ 1%) for either Fe or Co sites upon alloying.

[0315] Similar adsorption trend, Fe > FeCo > Co in the sequence of decreasing binding, was observed for *CO2, *O, *C, *OH and *H (Supplementary Tables S7 and S8), where the Co sites on FeCo(110) surface participated in the binding either alone as together with the neighboring Fe sites. Note that, the stabilization of the O-anchor of the adsorbates (e.g. *CO2, *O, *OH) was found to always involve the Fe sites. This is likely associated with the higher electronegativity of oxygen than carbon, which can reactivate the Fe site on FeCo(110) by pulling the atom outward to strengthen the binding on interaction. Such effect was well exemplified for the binding of *H2O, where only the Fe site was preferred. In this case, the trend in binding, FeCo (Eb= -0.66 eV) > Fe (Eb= -0.56 eV) > Co (Eb= -0.52 eV), was different from that seen for the other adsorbates. On FeCo(110), the molecule preferredCU24182 – 101879.000335 to be anchored at the Fe top site rather than the Co top site (Eb= -0.50 eV), which was also demonstrated by the higher flexibility of surface Fe sites than Co sites (Supplementary Table 10 and FIG.48) in response to H2O adsorption. Nevertheless, the difference in binding energy is rather small.

[0316] In addition, the alloying-induced modification in electronic structures, including the appropriate combination of deactivated Fe as compared to Fe(110) and activated Co as compared to Co(0001), also ensured the enhanced stability of bimetallic FeCo alloy with respect to monometallic Fe, being able to closely mimic the high stability observed for monometallic Co (Figure 3a).

[0317] Overall, the interplay between ensemble and electronic effects plays exceptional roles in enabling reasonable stability, selective bindings different species and thus selective CO conversion toward carbon nanofiber growth on FeCo alloy, which cannot be observed for Fe and Co alone.

[0318] Supplementary Note 5: Effects of Pd loading on the electrochemical reaction

[0319] 10%Pd / C demonstrated increasing H2 / CO ratios from 0.1 to 0.2, 0.6, and 1.6 with stable cell voltages over time (Supplementary Figure 49). Compared with -150 mA cm-2, however, a higher current density of -200 mA cm-2only promoted H2formation. The total CO2 conversion (Xtotal) was significantly improved at -150 mA cm-2(89.1%) and -200 mA cm-2(93.0%), which benefited the downstream thermochemical process by mitigating the reverse Boudouard reaction that resulted from unreacted CO2. In comparison, 20% and 40%Pd / C catalysts (FIGs.54 and 55) exhibited similar performance as the current density increased, but with lower H2 / CO ratios (e.g., 1.3 and 1.1 v.s.1.6 at -200 mA cm-2) as well as comparable formation rate of CO with 10%Pd / C.

[0320] Supplementary Note 6: Morphology of solid products after separation

[0321] The morphology of the post-separation solid products was evaluated using scanning electron microscopy (SEM), transmission electron microscopy (TEM) and selected area electron diffraction (SAED). SEM images (FIG.59) across different size scales (100 nm to 1 μm) highlighted the prevalence of CNF (tens of nm in diameter, hundreds-to-thousands of nm in length) in the solid products. Moreover, the open structure of CNF tips, where metal nanoparticles were placed, suggested effective metal removal by treatments such asCU24182 – 101879.000335 ultrasonication and nitric acid digestion as indicated by the inductively coupled plasma optical emission spectroscopy (ICP-OES) analysis.

[0322] To address potential X-ray diffraction interferences between CNF and CeO2, SAED was employed to assess CNF crystallinity by examining the solid products across different regions and scales. The TEM images and SAED patterns in FIG.60 indicated the dominance of high-crystallinity CNF (characterized by (002) and (100) diffraction patterns and d spacing), with minor CeO2residuals observed, emphasizing the importance of optimizing separation methods and / or utilizing the resulting CNF as support.

[0323] Supplementary Note 7: Effects of water vapor and CO2 on solid carbon formation

[0324] The presence of water vapor at the outlet of the electrolyzer was observed (FIG.4B and FIGs.53-55). The effect of water vapor on solid carbon formation was investigated by co-feeding CO and H2 with steam to the thermal reactor. As shown in FIG. 62, the addition of steam helped stabilize the syngas conversion process and enhanced solid carbon formation when there was a deficiency of H2, but its importance decreased when H2 became abundant. For example, the catalyst weight gain (CWG) increased from 120 mg to 220 mg after co-feeding water vapor with the mixture of CO / H2 / N2 (15 / 3 / 32 ml / min), while the CWG values were comparable (320 mg vs.305 mg) for the case of CO / H2 / N2(15 / 30 / 5 ml / min). This finding agreed with the observation of a similar CWG (250 mg) at a current density of -200 mA cm-2that yielded a higher H2 / CO ratio (FIG.56) than the case with -150 mA cm-2.

[0325] The impact of CO2on the syngas conversion process was also explored by co-feeding the CO / H2 stream with different amounts of CO2. In general, the presence of CO2 can increase the CWG values, likely due to the occurrence of the reverse water-gas shift (RWGS) reaction, which provided more CO for carbon fixation. Moreover, CO2 is a byproduct resulting from the Boudouard reaction (2CO → C(s) + CO2). Based on the above findings, it is reasonable to infer that recycling CO2 should have a positive or at least a non- detrimental effect on subsequent CNF formation.

[0326] Supplementary Tables

[0327] Supplementary Table 1 Comparison for CO adsorption and dissociation at the metal-CeO2interfaces and on metal surfaces.CU24182 – 101879.000335

[0328] DFT-calculated CO adsorption energy (Eads) and dissociation energy (^E) at Fe4-CeO2(111), Fe2Co2-CeO2(111), and Co4-CeO2(111) interfaces and on Fe(110), FeCo(110), and Co(0001) surfaces. Interface Eads(CO) ^E(*CO^*C+*O) Slab Eads(CO) ^E(*CO^*C+*O) Model eV eV Model eV eV Fe4 / CeO2(111) -2.37 0.48 Fe(110) -2.17 -0.89 Fe2Co2 / CeO2(111) -2.32 0.56 FeCo(110) -2.09 -0.19 Co4 / CeO2(111) -2.25 0.96 Co(0001) -1.86 0.82(BET); MCM-41 (hexagonal), ~1000 m2g-1; Fe(NO3)3∙9H2O, ≥99.999% trace metals basis; 2) Alfa-Aesar: SiO2, 168 m2g-1; γ-Al2O3, 80-120 m2g-1; Co(NO3)2∙6H2O, 99.999% metal basis

[0331] Supplementary Table 3 EDS-quantified compositions (Fe and Co) within the tip of CNF over the spent Fe3Co6 / CeO2 catalyst. Elements Series Mass fraction (wt%) Atom fraction (at%) Error (3σ) (wt%) Fe K 33.59 34.80 4.62 Co K 66.41 65.20 7.85

[0332] Note: The quantified region is the tip of CNF in Figure 1j in the main text. The atomic ratio of Co to Fe was 1.9, very close to the nominal synthesis ratio (2:1), suggesting that almost all the Fe and Co elements should reside within the tip of CNF.

[0333] Supplementary Table 4 Results of Raman peak deconvolution and integration for the spent Fe3Co6 / CeO2 samples and the commercial CNF.CU24182 – 101879.000335 Samples (reaction Param D D’’ G D’ ID / IA / Ratio Ratio conditions) eters band band band band IGIG-1a-2bI (single reactor: Positio 1338 1576. 16 00 mg of catalyst, n .0 15 02. 1 13.9 7 6 CO / H2 / N2=3 / 3 / 4 ml min-1, 450 °C, 1 Width 142. 353.8 42.1 atm, 4.5 h) Area 206. 2 4.6 0.6 1 8.6 44.0 29.8 8 5 0.77 0.92 II (single reactor: Positio 1339 1576. 1602. mg of catalyst, n .3 1510100 .7 5 8 CO / H2 / N2=15 / 15 / 2 0 ml min-1, 450 °C, Width 143. 7 98.1 56.0 44.6 1 atm, 4.5 h) Area 206. 4.7 0.6 8 29.0 43.3 33.0 7 7 0.76 0.89 III (tandem reactor: Positio 1338EC (-150 mA cm2)- n .8 7 4 TC (200 mg of FeCo catalyst, Width 127. 1 92.9 51.0 38.7 CO2 / N2=10 / 10 ml min-1, 450 °C, 1 Area 181. 2 4.4 0.5 8 3.5 40.8 25.0 0.81 1.04 atm, 11 h) 6 8 IV ((tandem reactor: Positio 1337 200 mA cm )- n .4 1 1577. 1603. EC (--2517.5 5 3 TC (200 mg of FeCo catalyst, Width 141. 1 87.1 50.3 39.9 CO2 / N2=10 / 10 ml min-1, 450 °C, 1 Area 205. 2 4.8 0.5 3 4.1 42.1 27.6 0.74 1.05 atm, 11 h) 7 7 Commercial CNF Positio 1342 1571. 16 .1 15 06. n 05.0 5 9 Width 160.50.0 Area 222. 3.6 0.6 4 36.8 61.6 36.5 1 0 1.00 1.00

[0334] Notes:to the protocols established in literature14,15, the Raman spectroscopy between 900 and 2000 cm-1(FIG.19A) indicated five primary peaks: D’ band (~1600 cm-1), G band (1570-1580 cm-1, graphitic), D’’ band (1505-1520 cm-1), D band (~1340 cm-1), and a supplementary shoulder band (~1200 cm-1). The origin of the small shoulder peak has not been clearly identified and thus was not considered for peak deconvolution. The D and G bands are associated with the defect and graphitic carbonCU24182 – 101879.000335 structures, respectively. However, it should be noted that the D band originates from not only amorphous carbon agglomerates but also CNF-related factors, including defects in the curved graphene sheets, CNF ends and edges, and finite size of crystalline domains of the CNFs. In comparison, the D'' band has been identified to originate from only amorphous carbon. Thus, two intensity ratios were calculated: the ratio of the D and G bands (ID / IG) to indicate crystallinity, and the ratio of the D'' and G bands (ID'' / IG) to reflect the degree of amorphous characteristics in the resulting solid carbon products. In addition, the fraction or the purity of CNFs over the solid carbon was approximated based on the ratio of the measured ID’’ / IG ratio to the counterpart of the commercial CNF sample, i.e., (IA / IG)catal-1 / (IA / IG)CNF-1. A ratio of 1 was defined here to represent 100% purity, indicating a purity same as that of the commercial CNF. However, it is crucial to clarify that this ratio of 1 does not necessarily imply a genuine 100% pure CNF. Likewise, the crystallinity was approximated by (ID / IG)catal-1 / (ID / IG)CNF-1.

[0335] Ratio-1a: (ID / IG)catal-1 / (ID / IG)CNF-1.

[0336] 1.

[0337] compositions from EDS mappings overthe spent Fe3Co6 / CeO2 sample after the thermochemical experiment. Elements Line Type k Factor Atomic % Wt% Wt% Sigma 85235the FeCo / CeO2 catalyst under different conditions. E Samples Conditions Shells CNBond0len σ2(Å2) shift R gth (Å) (eV) factor fcc Co foil 25 °CaCo-Co 12 2.49±0.000.008 Fe3Co6 / N2@25 °CbCo-O 6.2±0.3 1.91±0.00 0.0030 2.68 0.008 CeO2 H2@500 °CcCo-M 3.1±0.1 2.45±0.00 0.0028 8.06 0.008 Co-M’ 1.1±0.2 2.98±0.01 0.0028 8.06CU24182 – 101879.000335 Rxn.@450 Co-C 0.8±0.1 1.80±0.02 0.0012 -1.79 0.003 °C_25 °CdCo-M 2.6±0.1 2.42±0.00 0.0022 3.44 Co-M’ 1.8±0.1 2.81±0.00 0.0022 3.44 bcc Fe 25 °CeFe-Fe 8 2.46±0.00 0.0049 4.18 0.008 foil Fe-Fe’ 6 2.84±0.01 0.0041 4.18 Fe3Co6 / N2@25 °CfFe-O 6.1±0.4 1.96±0.01 0.0070 1.74 0.008 CeO2Fe-Fe* 2.5±0.6 2.96±0.02 0.0070 1.74 H2@500 °CgFe-M 3.4±0.3 2.43±0.01 0.0051 5.44 0.008 Fe-M’ 1.4±0.5 2.89±0.02 0.0051 5.44

[0339] Notes: CN―average coordination number (normalized to all the absorbers) around absorbing center atom; σ2―mean square variation in path length; R factor―quality of fitting. Rxn.@450 °C_25 °C indicated that the catalyst was exposed to the CO and H2stream (CO / H2 / N2=15 / 15 / 20 ml min-1) at 450 °C for 5 h, followed by cooling down to 25 °C for the ex situ XAS spectrum collection; M and M’ could be Fe or Co due to the typical limitation of EXAFS in differentiating elements at the similar distance as the Z difference is less than 3; Fe-Fe* represented the Fe-Fe bond in the iron oxide structure. The fitting parameters are summarized as below:

[0340] a: ∆k=2.8-10.0 Å-1, ∆R=1.0-3.0 Å, fitted with the scattering path from the Co;

[0341] b: ∆k=2.8-10.0 Å-1, ∆R=1.0-2.0 Å, fitted with the scattering path from Co3O4;

[0342] c: ∆k=3.0-10.0 Å-1, ∆R=1.0-3.0 Å, fitted with the scattering paths from FeCo;

[0343] d: ∆k=3.0-12.0 Å-1, ∆R=1.0-3.0 Å, fitted with the scattering paths from Co2C and FeCo;

[0344] e: ∆k=3.0-10.0 Å-1, ∆R=1.0-3.0 Å, fitted with the scattering paths from Fe;

[0345] f: ∆k=2.7-10.0 Å-1, ∆R=1.0-3.0 Å, fitted with the scattering paths from Fe2O3;

[0346] f: ∆k=2.7-10.0 Å-1, ∆R=1.0-3.0 Å, fitted with the scattering paths from Fe2O3;

[0347] k1k2k3-weighted EXAFS fitting

[0348] Supplementary Table 7 Summary of DFT-calculated binding energies, reaction energies, and activation energies for different processes on the surfaces of mono- and bimetallics.CU24182 – 101879.000335

[0349] DFT-calculated binding energies of CO, CO2, H2O, H and C species, reaction energies and activation energies for CO dissociation, CO2 formation, H2O formation, O-H bond formation and C-C bond formation processes on FeCo(110), Fe(110), and Co(0001) surfaces. The references for CO, CO2, H2O, H and C adsorption energy were set as gas phase CO, CO2, H2O, H2, and bulk of graphite, respectively. CO dissociation energy and binding energy (eV) Models Eb(CO) ^E(*CO^*C+*O) Ea(*CO^*C+*O) FeCo(110) -2.09 Fe(110) -2.17- Co(0001) -1.86 0.82 2.28 C-C bond formation energy and C binding energy (eV) Models Eb(C)^E(*C+*C^*C-C)Eads(*C+*C^*C-C)FeCo(110) 0.23Fe(110) -0.26 Co(0001) 0.89 -0.62 0.96 CO2formation and binding energy (eV) ModelsEb(CO2)^E(*CO+*O^*CO2) Ea(*CO+*O^*CO2)FeCo(110) -0.52Fe(110) -0.68 1.72 1.73 Co(0001) -0.22 1.06 1.27 O-H bond formation energy and H binding energy (eV) Models Eb(H) ^E(*O+*H^*OH) Ea(*O+*H^*OH) FeCo(110) -0.75Fe(110) -0.84 0.71 1.67 Co(0001) -0.67 0.18 1.11 H2O formation and binding energy (eV) Models Eb(H2O) ^E(*OH+*H^*H2O) Ea(*OH+*H^*H2O) FeCo(110) -0.66Fe(110) -0.56 1.30 2.02 Co(0001) -0.52 0.59 1.50

[0350] Supplementary Table 8 Summary of DFT-calculated binding energies, reaction energies, and activation energies related to CO on the surfaces of metal carbides.CU24182 – 101879.000335

[0351] DFT-calculated binding energies for CO, reaction energies and activation energies for CO dissociation and binding energies on Fe2CoC(001), Fe3C(001), and Co3C(0001) surfaces. CO dissociation and binding energy (eV) Eb(CO)^E(*CO^*C+*O)Ea(*CO^*C+*O)- Co3C(001) -2.23 0.71 2.34

[0352] Supplementary Table 9 Diffusion of *C species on surfaces.

[0353] DFT-calculated reaction energies and activation energies for surface diffusion of *C species on FeCo(110) and Co(0001) surfaces. Surface diffusion of *C species Models ΔE, eV Ea, eV FeCo(110) 0.11 1.09 Co(0001) 0.30 0.32

[0354] Supplementary Table 10 Summary of DFT-calculated reaction energies for H2-assisted CO dissociation on the surfaces of mono- and bimetallics.

[0355] DFT-calculated reaction energies for *CO+*H^*COH and *CO+*H^*HCO and their comparisons with CO dissociation, on FeCo(110), Fe(110), and Co(0001) surfaces. CO dissociation and association energy (eV) Models ^E(*CO^*C+*O) ^E(*CO+*H^*HCO) ^E(*CO+*H^*COH) FeCo(110)Fe(110) -0.89 0.64 0.77 Co(0001) 0.82 1.06 0.85

[0356] Supplementary Table 11 Summary of DFT-calculated binding energies of gas phase molecules at different catalytic sites on the surfaces of mono- and bimetallics.

[0357] DFT-calculated binding energies for CO (a), CO2 (b), and H2O (c) species at the different sites on FeCo(110), Fe(110), and Co(0001) surfaces. Only the sites that can stabilize the species during the DFT relaxation are shown while the sites that automaticallyCU24182 – 101879.000335 move to the neighbor stable sites during DFT relaxation are not listed. Each binding energy is referenced to each corresponding gas phase molecule.

[0358] a Binding energy of CO Models Stable binding site Eb(CO), eV Top-Co -2.09 Top-Fe -1.89 FeCo(110) Bridge-Co-Co -2.07 Bridge-Fe-Fe -1.81 Bridge-Fe-Co -1.90 Top-Fe -2.17 Fe(110) Long-Bridge-Fe-Fe -2.11 Hollow-Threefold -2.12 Top-Co -1.86 Co(0001) Hollow-FCC -1.88 Hollow-HCP -1.89

[0359] b Binding energy of CO2Models Stable binding site Eb(CO2), eV Top-Co-C, Bridge-Fe-Co-O, Bent-CO2 -0.52 FeCo(110) Top-Fe-C, Bridge-Fe-Co-O, Bent-CO2-0.39 No direct binding, Linear-CO2-0.23 Top-Fe-C, Short-Bridge-Fe-Fe-O, Bent-CO2-0.68 Top-Fe-O, Long-Bridge-Fe-Fe-C, Bent-CO2-0.64 Fe(110) Top-Fe-O, Short-Bridge-Fe-Fe-C, Bent-CO2 -0.54 No direct binding, Linear-CO2 -0.20 Top-Co-C, Bridge-Co-Co-O, Bent-CO2 -0.08 Co(0001) Top-Co-O, Bridge-Co-Co-C, Bent-CO2 -0.12 No direct binding, Linear-CO2 -0.22

[0360] cCU24182 – 101879.000335 Binding energy of H2O Models Stable binding site Eb(H2O), eV Top-Fe -0.66 FeCo(110) Top-Co -0.50 Fe(110) Top-Fe -0.56 Co(0001) Top-Co -0.52

[0361] Supplementary Table 12 Summary of DFT-calculated binding energies of important reaction intermediates at different catalytic sites on the surfaces of mono- and bimetallics.

[0362] DFT-calculated binding energies for *C (a), *O (b), *H (c), and *OH (d) species at the different sites on FeCo(110), Fe(110), and Co(0001) surfaces. Only the sites that can stabilize the species during the DFT relaxation are shown while the sites that automatically move to the neighbor stable sites during DFT relaxation are not listed. The references for binding energy are shown in the table.

[0363] a Binding energy of *C species Models Reference Stable Binding Site Eb(C), eV Long-Bridge-Co- Co 0.23 FeCo(110) Long-Bridge-Fe-FeFe(110) Graphite C Long-Bridge-Fe-Fe -0.26 Hollow, HCP 0.89 Co(0001) Hollow, FCC 1.19

[0364] b Binding energy of *O species Models Reference Stable Binding Site Eb(O), eV Hollow-Threefold, Fe-Fe-Co 0.05 FeCo(110)*O in equilibrium with CO+*O^CO2Hollow-Threefold, Co-Co-Fe 0.29 Fe(110) Hollow, Threefold -0.31CU24182 – 101879.000335 Hollow, HCP 0.53 Co(0001) Hollow, FCC 0.65

[0365] c Binding energy of *H species Models Reference Stable Binding Site Eb(H), eV Hollow-Threefold, Fe-Fe-Co -0.75 FeCo(110) Hollow-Threefold,Co-Co-Fe H2(g) Fe(110) Hollow, ThreefoldHollow, HCP -0.67 Co(0001) Hollow, FCC -0.64

[0366] d Binding energy of *OH species Models Reference Stable Binding Site Eb(OH), eV Hollow-Threefold,Fe-Fe-Co FeCo(110) Hollow-Threefold,quilibrium with C - *OH in e o-Co-Fe Fe(110) *OH+1 / 2H2^H2O Hollow-ThreefoldHollow, HCP -0.54 Co(0001) Hollow, FCC -0.45

[0367] Supplementary Table 13 Summary of DFT-calculated d band center, nearest neighbor coordination number, and the strain for the surfaces of mono- and bimetallics.

[0368] DFT-calculated d band center (a), the nearest neighbor coordination number (b), and the strain (c) for top Fe and Co sites of the FeCo(110), Fe(110), and Co(0001) surfaces.

[0369] a d Band CenterCU24182 – 101879.000335 Models Elements εd ( eV ) Fe -2.31 FeCo(110) Co -1.94 Fe(110) Fe -1.79 Co(0001) Co -2.00

[0370] b Nearest Neighbor Coordination Number (NN-CN) Models Elements NN-CN Fe 6 FeCo(110) Co 6 Fe(110) Fe 6 Co(0001) Co 9

[0371] c Nearest Neighbor Strain (NNS) Models Elements NNS Fe +0.24% [reference to Fe(110)] FeCo(110) Co -1.16% [reference to Co(0001)]

[0372] Supplementary Table 14 Summary of DFT-calculated percentage change of Fe and Co sites before and after CO or H2O adsorption in z-direction on the surfaces of mono- and bimetallics.

[0373] DFT-calculated z-direction percentage change of Fe and Co sites for FeCo(110), Fe(110), and Co(0001) surfaces after CO or H2O adsorption with reference to the corresponding clean surfaces. z-direction percentage change of Fe and Co sites upon adsorption Models elements CO adsorption H2O adsorption Fe -0.20% +1.64% FeCo(110) Co +1.09% +1.16% Fe(110) Fe -0.36% +1.78%CU24182 – 101879.000335 Co(0001) Co +1.37% +1.10%

[0374] Supplementary Table 15 Element compositions from EDS mappings over the spent Fe3Co6 / CeO2sample after the electrochemical-thermochemical tandem experiment. Elements Line Type k Factor Atomic % Wt% Wt% Sigma CK series2.786 99.2 98.22 0.159451ICP results. Mass of sample (mg) Nominal mass Nominal mass of ICP measured mass Samples loading (%) metals (mg) metals (mg) Inital Spent Used for separation Co Fe Co Fe Co Fe 0.45 0.55 0.581.24

[0376] Notes: Tandem_NC_XX h denotes the solution sample obtained from the spent Fe3Co6 / CeO2sample after the tandem-reactor experiment, which was subsequently treated with concentrated nitric acid for XX hours. Single_NC_5 h refers to the solution obtained from the spent Fe3Co6 / CeO2sample after the single thermal-reactor experiment, which underwent a 5-h treatment with concentrated nitric acid.

[0377] References

[0378] 1. Reuter, K. & Scheffler, M. Composition, structure, and stability of RuO2(110) as a function of oxygen pressure. Phys. Rev. B 65, 035406 (2001).CU24182 – 101879.000335

[0379] 2. Reuter, K. & Scheffler, M. Composition and Structure of the RuO2(110) Surface in an O2 and CO Environment: Implications for the Catalytic Formation of CO2. Phys. Rev. B 68, 045407 (2003).

[0380] 3. Kang, J. et al. Surface characterization and methane activation on SnOx / Cu2O / Cu(111) inverse oxide / metal catalysts. Phys. Chem. Chem. Phys.23, 17186– 17196 (2021).

[0381] 4. Chase, M. W. & Organization (US), N. I. S. NIST-JANAF thermochemical tables. vol.9 (American Chemical Society Washington, DC, 1998).

[0382] 5. Douglas, A., Carter, R., Li, M. & Pint, C. L. Toward Small-Diameter Carbon Nanotubes Synthesized from Captured Carbon Dioxide: Critical Role of Catalyst Coarsening. ACS Appl. Mater. Interfaces 10, 19010–19018 (2018).

[0383] 6. Xie, Z., Wang, X., Chen, X., Liu, P. & Chen, J. G. General Descriptors for CO2-Assisted Selective C–H / C–C Bond Scission in Ethane. J. Am. Chem. Soc.144, 4186–4195 (2022).

[0384] 7. Xie, Z. et al. Dry reforming of methane over CeO2-supported Pt-Co catalysts with enhanced activity. Appl. Catal. B: Environ.236, 280–293 (2018).

[0385] 8. Yan, B. et al. Dry reforming of ethane and butane with CO2 over PtNi / CeO2bimetallic catalysts. ACS Catal.6, 7283–7292 (2016).

[0386] 9. Richter, R. & Eschrig, H. Spin-polarised electronic structure of disordered BCC FeCo alloys from LCAO-CPA. J. Phys., F Met. Phys.18, 1813 (1988).

[0387] 10. Seyring, M., Simon, A., Voigt, I., Ritter, U. & Rettenmayr, M. Quantitative crystallographic analysis of individual carbon nanofibers using high resolution transmission electron microscopy and electron diffraction. Carbon 116, 347–355 (2017).

[0388] 11. Zhou, J.-H. et al. Structural characterization of carbon nanofibers formed from different carbon-containing gases. Carbon 44, 3255–3262 (2006).

[0389] 12. Carta, D., Mountjoy, G., Apps, R. & Corrias, A. Effect of the Support on the Formation of FeCo Alloy Nanoparticles in an SBA-16 Mesoporous Silica Matrix: An X-ray Absorption Spectroscopy Study. J. Phys. Chem. C 116, 12353–12365 (2012).

[0390] 13. Liu, Y. et al. In situ XAFS study on the formation process of cobalt carbide by Fischer–Tropsch reaction. Phys. Chem. Chem. Phys.21, 10791–10797 (2019).CU24182 – 101879.000335

[0391] 14. Choi, Y. C., Min, K.-I. & Jeong, M. S. Novel Method of Evaluating the Purity of Multiwall Carbon Nanotubes Using Raman Spectroscopy. J. Nanomater.2013, 615915 (2013).

[0392] 15. Nguyen Van, T. D., Sufian, S., Mansor, N. & Yahya, N. Characterization of Carbon Nanofibers Treated with Thermal Nitrogen as a Catalyst Support Using Point-of-Zero Charge Analysis. J. Nanomater.2014, 631069 (2014).

[0393] Aspects

[0394] Aspect 1. A method, comprising: electrolyzing a CO2 input and water so as to form a first product comprising CO and H2, the electrolyzing optionally being performed over a Pd / C catalyst or a catalyst that comprises any one or more of gold, silver, iron, cobalt, nickel, copper, or zinc; and thermochemically processing the first product so as to give rise to a second product that comprises carbon nanofibers or nanotubes.

[0395] The CO2 input can be present, for example, as a pure gas. The CO2 input can also be present, for example as part of an input. Such an input can include CO2as well as one or more other species. The input can comprise N2.

[0396] Aspect 2. The method of Aspect 1, wherein electrolyzing the CO2input and water comprises at least one of a CO2 reduction reaction and the hydrogen evolution reaction. The CO2reduction reaction can give rise, for example, to a syngas. As explained herein, the disclosed technology can include co-electrolysis of CO2 and water into syngas (CO and H2) with a subsequent thermochemical process.

[0397] The hydrogen evolution reaction can be performed over a catalyst. Such catalysts can comprise, as but some examples, noble and non-noble metal-based materials. Such catalysts can comprise platinum, palladium, ruthenium, iridium, transition metals, transition metal oxides, transition metal sulfides, transition metal phosphides, transition metal carbides, transition metal nitrides, and the like.

[0398] Aspect 3. The method of any one of Aspects 1-2, wherein the thermochemical processing includes at least one of the Boudouard reaction (termed R1 in some instances) and CO + H2 ^ C(s) + H2O (termed R2 in some instances). In some embodiments, the thermochemical processing includes both of the Boudouard reaction and CO + H2 ^ C(s) + H2O. In some embodiments, the thermochemical processing includes reaction. In some embodiments, the thermochemical processing includes the CO + H2 ^ C(s) + H2O reaction. Hydrogen (H2) produced during performance of theCU24182 – 101879.000335 disclosed technology can be recycled, as such hydrogen can be considered a renewable byproduct.

[0399] The Boudouard reaction and the CO + H2^ C(s) + H2O reaction can be performed over a catalyst that comprises any one or more of nickel, iron, cobalt, chromium, manganese, copper, or molybdenum.

[0400] Aspect 4. The method of any one of Aspects 1-3, wherein the thermochemical processing occurs at less than about 600 ºC. The thermochemical processing can occur at, for example, from about 50 to about 600 ºC, or from about 100 to about 550 ºC, or from about 150 to about 500 ºC, or from about 200 to about 450 ºC, or from about 250 to about 400 ºC, or even from about 300 to about 350 ºC.

[0401] Aspect 5. The method of any one of Aspects 1-4, wherein the thermochemical processing occurs at from about 400 to about 450 ºC.

[0402] Aspect 6. The method of any one of Aspects 1-5, wherein the thermochemically processing is performed over a bimetallic catalyst. The bimetallic catalyst can comprise, for example, FexCoy, such as Fe3Co6. Iron-cobalt catalysts are considered particularly suitable. As described herein, the disclosed technology is especially well-suited to performance with FeCo alloy and extra metallic Co.

[0403] Aspect 7. The method of Aspect 6, wherein the bimetallic catalyst comprises FeCo.

[0404] Aspect 8. The method of Aspect 6, wherein the thermochemically processing is performed over a catalyst that is any one or more of (1) supported by CeO2 or (2) comprises Co metal. FeCo catalyst is considered particularly suitable.

[0405] Aspect 9. The method of any one of Aspects 1-8, wherein the second product comprises CO2, and wherein at least some of the CO2of the second product is recycled and electrolyzed with water.

[0406] Aspect 10. The method of any one of Aspects 1-9, wherein the second product comprises H2O, and wherein at least some of the H2O of the second product is recycled and electrolyzed with CO2.

[0407] Aspect 11. The method of any one of Aspects 1-10, wherein the second product comprises H2.

[0408] Aspect 12. The method of any one of Aspects 1-11, wherein input CO2is an output of a process.CU24182 – 101879.000335

[0409] Aspect 13. The method of any one of Aspects 1-11, wherein input CO2is environmental CO2.

[0410] Aspect 14. The method of any one of Aspects 1-13, further comprising sequestering the carbon nanofibers or nanotubes.

[0411] Aspect 15. A system, the system comprising: a first reaction zone, the first reaction zone configured to receive a CO2 input and water, and the first reaction zone configured to support electrolysis of the CO2input and water to evolve a product that comprises CO; a second reaction zone, the second reaction zone configured to receive a product from the first reaction zone, the second reaction zone configured to support at least one of the Boudouard reaction (R1) and CO + H2^ C(s) + H2O (R2). The at least one of the Boudouard reaction (R1) and CO + H2 ^ C(s) + H2O (R2) can give rise to a product that comprises at least one of carbon or nanotubes.

[0412] The first reaction zone can comprise therein a catalyst that promotes the electrolysis reaction; such a catalyst can be a Pd / C catalyst. The first reaction zone can include an electrolyzer.

[0413] Aspect 16. The system of claim 15, wherein the first reaction zone comprises a Pd / C catalyst and / or a catalyst that comprises any one or more of gold, silver, iron, cobalt, nickel, copper, or zinc.

[0414] Aspect 17. The system of any one of claims 15-16, wherein the second reaction zone comprises a FeCo catalyst.

[0415] The system can be configured for performance of the thermochemical processing of the product of the electrolyzing. The second reaction zone can comprise therein a catalyst that promotes the electrolysis reaction; such a catalyst can include Fe and Co, in some embodiments.

[0416] Aspect 18. The system of any one of claims 15-17, further comprising a separation train. The separation train can operate to separate carbon (e.g., carbon nanofibers or nanotubes) evolved during the technology from other products of the thermochemical processing. This can be effected by separating the solid carbon phase from the gaseous effluents. The separation train can also be operated to separate H2from other gaseous effluents. This can be accomplished by, for example, pressure swing adsorption, cryogenic distillation, and membrane separation. The choice of the most suitable separation method for practical applications can depend on various factors, such as the gas volume, required H2CU24182 – 101879.000335 purity, energy efficiency, and cost considerations. Following the separation process, remaining CO and CO2, along with a minor amount of H2O if any, can be directly recycled to the electrolyzer for subsequent reactions.

[0417] Aspect 19. The system of claim 18, wherein the separation train is operable to effect at least one of (1) separating carbon from other products of the second reaction zone, or (2) separate H2 from other products of the second reaction zone.

[0418] Aspect 20. A system, the system configured to perform the method of any one of Aspects 1-14.

Claims

CU24182 – 101879.000335 What is Claimed:

1. A method, comprising: electrolyzing a CO2 input and water so as to form a first product comprising CO and H2, the electrolyzing optionally being performed over a Pd / C catalyst or a catalyst that comprises any one or more of gold, silver, iron, cobalt, nickel, copper, or zinc; and thermochemically processing the first product so as to give rise to a second product that comprises carbon nanofibers or nanotubes.

2. The method of claim 1, wherein electrolyzing the CO2 input and water comprises at least one of a CO2reduction reaction and the hydrogen evolution reaction.

3. The method of any one of claims 1-2, wherein the thermochemical processing includes at least one of the Boudouard reaction and CO + H2 ^ C(s) + H2O.

4. The method of any one of claims 1-2, wherein the thermochemical processing occurs at less than about 600 ºC.

5. The method of any one of claims 1-2, wherein the thermochemical processing occurs at from about 400 to about 450 ºC.

6. The method of any one of claims 1-2, wherein the thermochemically processing is performed over a bimetallic catalyst.

7. The method of claim 6, wherein the bimetallic catalyst comprises FeCo.

8. The method of claim 6, wherein the thermochemically processing is performed over a catalyst that is any one or more of (1) supported by CeO2and (2) comprises Co metal.

9. The method of any one of claims 1-2, wherein the second product comprises CO2, and wherein at least some of the CO2 of the second product is recycled and electrolyzed with water.CU24182 – 101879.000335 10. The method of any one of claims 1-2, wherein the second product comprises H2O, and wherein at least some of the H2O of the second product is recycled and electrolyzed with CO2.

11. The method of any one of claims 1-2, wherein the second product comprises H2.

12. The method of any one of claims 1-2, wherein input CO2 is an output of a process.

13. The method of any one of claims 1-2, wherein input CO2 is environmental CO2.

14. The method of any one of claims 1-2, further comprising sequestering the carbon nanofibers or nanotubes.

15. A system, the system comprising: a first reaction zone, the first reaction zone configured to receive a CO2 input and water, and the first reaction zone configured to support electrolysis of the CO2 input and water to evolve a product that comprises CO; a second reaction zone, the second reaction zone configured to receive a product from the first reaction zone, the second reaction zone configured to support at least one of the Boudouard reaction (R1) and CO + H2^ C(s) + H2O (R2).

16. The system of claim 15, wherein the first reaction zone comprises a Pd / C catalyst or a catalyst that comprises any one or more of gold, silver, iron, cobalt, nickel, copper, or zinc.

17. The system of any one of claims 15-16, wherein the second reaction zone comprises a FeCo catalyst.

18. The system of any one of claims 15-16, further comprising a separation train.CU24182 – 101879.000335 19. The system of claim 18, wherein the separation train is operable to effect at least one of (1) separating carbon from other products of the second reaction zone, or (2) separate H2from other products of the second reaction zone.

20. The system of claim 15, wherein the system is configured to perform the method of claim 1.

Images