CO2-Free Power Generation on an Iron Group Nanoalloy Catalyst via Selective Oxidation of Ethylene Glycol to Oxalic Acid in Alkaline Media

An Fe group ternary nanoalloy (NA) catalyst enabled selective electrocatalysis towards CO2-free power generation from highly deliverable ethylene glycol (EG). A solid-solution-type FeCoNi NA catalyst supported on carbon was prepared by a two-step reduction method. High-resolution electron microscopy techniques identified atomic-level mixing of constituent elements in the nanoalloy. We examined the distribution of oxidised species, including CO2, produced on the FeCoNi nanoalloy catalyst in the EG electrooxidation under alkaline conditions. The FeCoNi nanoalloy catalyst exhibited the highest selectivities toward the formation of C2 products and to oxalic acid, i.e., 99 and 60%, respectively, at 0.4 V vs. the reversible hydrogen electrode (RHE), without CO2 generation. We successfully generated power by a direct EG alkaline fuel cell employing the FeCoNi nanoalloy catalyst and a solid-oxide electrolyte with oxygen reduction ability, i.e., a completely precious-metal-free system.

9. Figure S6. CVs using FeCoNi/C and Pt/C modified working electrodes in KOH aqueous solution containing EG.
10. Figure S7. Onset potential estimations in voltammograms using FeCoNi/C and Pt/C modified working electrodes.
12. Figure S9. Calibration curve between GC peak areas, detected by FI detector, vs. amounts S3 of CO.
13. Figure S10. Calibration curve between GC peak areas, detected by FI detector, vs. amounts of NaCO 3 .
16. Figure S13. Amounts of electrochemically evolved gaseous products on FeCoNi/C and Pt/C in EG oxidation at 1.0 V. Figure S14 Number of electrons, coulomb numbers, current efficiencies, and selectivities, related with oxidized product formation from EG on Fe/C, FeCoNi/C, and Pt/C. Figure S15. Powder XRD patterns of FeCoNi/C before and after the CA experiment.

References.
References section includes the previous reports for EG electrooxidation which could not be cited in manuscript, because of reference number limitation.
1-1. Preparation of Fe/C monometallic catalyst. The Fe/C nanoalloy catalyst was obtained by the similar procedure for the preparation of FeCoNi/C, except for the scales of starting material, reagents, and solvents. We found that Fe nanoparticles supported on carbon with metal loading higher than 40 wt. % tend to have diameters more than ca. 90 nm. For elimination of the need for considering size-effects on catalytic activities, therefore, Fe nanoparticles supported on carbon having an averaged diameter similar to that of the ternary nanoalloy were prepared by adjusting the metal loading. Fe II (OAc) 2 (0.1396 g, 0.8 mmol) was dissolved into a mixed solvent of PEG (0.3525 g) and TEG (200 mL). After vigorous stirring with bubbling Ar for 30 minutes at room temperature, the reaction mixture was heated up to 80 °C and kept stirred for 30 minutes. The MeOH (10 mL) suspension of Vulcan (0.083 g) was added to this pale brown colored solution, and stirred for 30 minutes at room temperature. After stirring at 80 °C for 3 minutes, an aqueous solution (10 mL) of NaBH 4 (0.3029 g) was added. The metal-oxide composite and carbonsupported Fe nanoparticle catalyst were prepared in similar way as in the case of FeCoNi/C. The metal content was determined to be 28.5 wt.% by ICP-MS analysis.
1-2. Powder X-ray diffraction (XRD) measurement. Powder X-ray diffraction measurements were carried out with Cu Kα radiation (λ = 1.54059 Å) using a Rigaku SmartLab at room temperature. Synchrotron powder XRD measurements were performed at the RIKEN materials sience beamline BL44B2 of SPring-8. 1 The data were acquired using the Debye-Scherrer camera equipped with an imaging plate as an X-ray detector. The incident wavelengths were 0.579057 Å for FeCoNi/C and Fe/C samples, which were obtained by calibration using CeO 2 as a standard powder sample. The X-ray beam was collimated by a double slit 0.5 mm by 3.0 mm. Powder samples of FeCoNi/C or Fe/C was sealed in borosilicate glass capillaries under vacuo. The samples were irradiated by X-ray at 300 K. Figures S1 and S2 show the fitting results of FeCoNi/C and Fe/C, respectively. The optimized parameters were listed in Table S2.
1-3. TEM measurement of Fe/C monometallic catalyst. TEM image of Fe/C was taken with a JEM-2010HCKM operated at 200 kV, and obtained image is shown in Figure S4. For these measurements, Fe/C mounted copper grid was prepared by the similar procedures with that of FeCoNi/C for BF-STEM image measurement, as above described. washing with acetone for 3 times and drying in vacuo for over night (ca. 12 h). The felt was heattreated under N 2 gas at 400 °C for 30 minutes followed by H 2 gas at 300 °C for 10 minutes, respectively. After cooling the felt to room temperature, it was fixed to the handmade stainless clip, and which was used as working electrode. As the counter electrode, coiled Pt wire was used. A Hg/HgO reference electrode (RE-6A, BAS Co. Ltd.) with filling 1 M KOH aqueous solution was used. All potentials were measured against this Hg/HgO reference, which has a potential of 0.098 V vs. the normal hydrogen electrode (NHE), and finally converted to that vs. referenced hydrogen electrode (RHE). For CV using a prepared catalyst, VersaSTAT4 potentiostat (Princeton Applied Research, AMETEC Inc.) was used with applying abovementioned three electrodes. A sample vial (100 mL in volume, ALS Co. Ltd.) equipped with a gas-tight Teflon cap was used. The electrolyte aqueous solution (80 mL, 20 wt.% KOH, 30 wt.% EG) was introduced, and working, reference, and counter electrodes were placed inside the vial. The electrolyte solution was prepared by using ultrapure water (DIRECT-Q®3UV, Millipore Corp., Merck Ltd.). After the Teflon cap was tightly closed, N 2 gas was bubbled in cell for 30 minutes in order to purge the air from the inside of the cell. After the deaeration, the current value was recorded against the applied potential with 10 mV/s scan rate and 10 scan cycles. The CV measurement of the blank, the same procedures abovementioned, except for using electrolyte solution (80 mL, 20 wt.% KOH), was carried out. On the other hand, for the CV measurement using the Pt/C catalyst, commercially available 20 wt% Pt/C (2.5 mg, Alfa Aesar) mounted carbon felt (1 cm 2 ) was used as working electrode.
The EG (0.7 mg) suspension of prepared catalysts (50 mg) was applied on the carbon felts (4 cm 2 × 4, 16 cm 2 total). The felts were heat-treated under N 2 gas at 400 °C for 30 minutes followed by H 2 gas at 300 °C for 10 minutes. After cooling the felts to room temperature, they were fixed to the handmade stainless clip, and which was used as working electrode. Counter and reference electrodes were the same with those at CV measurement. All potentials were measured against vs.
Hg/HgO and converted to them vs. RHE. For chronoamperometry measuremtns, VersaSTAT4 potentiostat was used with applying abovementioned three electrodes. The electrochemical experiments were fully carried out at inside of the glove box filled with N 2 gas. A home-build double compartment cell, where each compartment is separated by proton conducting membrane (Nafion®, NRE-212, Sigma-Aldrich) and equipped with gas-tight Teflon caps for each was used.  From the headspace of this vessel, gas sample (2 mL) was collected and analyzed by GC to determine the amount of dissolved CO 2 gas. 1-6. Definitions of number of electrons, current efficiency, and selectivity.

Number of Electrons
The number of electrons is defined as the number of electrons which are related to the oxidized product formation from EG. The number of electrons was calculated from the amount of an oxidized product, which was quantified by HPLC, from the reaction solution in anodic cell. For example, 8 electrons are required for oxalic acid formation from EG. Therefore, if x mol of oxalic acid was detected during the reaction, the "number of electrons" for oxalic formation can be calculated using equation as follows (Eq. 1).
The number of electrons was finally divided by the metal weight in the catalyst, and shown as per metal weight (g) in the Figures 3a and S14, respectively.

Current Efficiency
The current efficiency is defined as the percentage of the electrons, which is relevant to the product formation, out of the total number of electrons which pass through the circuit and counted by potentiostat during the experiment. This value can be calculated from "number of electrons" and "coulomb number" which is counted by potentiostat, using equation as described below (Eq. 2). If this value was close to 100 (%), the counted electrons can be considered to be based on the EG oxidation, i.e. oxidized product formation from EG. Contrastively, when it was close to 0 (%), the counted electrons are based on the other factor, e.g. catalyst self-oxidation. 2

Selectivity
The selectivity is defined as the percentage of the electrons related to the specific product formation out of total electrons for all the product formations. This value can be calculated from the "number of electrons" for specific product and sum of them for all the products. For example, "selectivity" of oxalic acid can be calculated using equation as follows (Eq. 3).