High-purity ethylene production via indirect carbon dioxide electrochemical reduction

High-purity ethylene production from CO2 electroreduction (CO2RR) is a coveted, yet arduous feat because the product stream comprises a blend of unreacted CO2, H2, and other off-target CO2 reduction products. Here we present an indirect reduction strategy for CO2-to-ethylene conversion, one that employs 2-bromoethanol (Br-EO) as a mediator. Br-EO is initially generated from CO2RR and subsequently undergoes reduction to ethylene without the need for energy-intensive separation steps. The optimized AC-Ag/C catalyst with Cl incorporation reduces the energy barrier of the debromination step during Br-EO reduction, and accelerates the mass-transfer process, delivering a 4-fold decrease of the relaxation time constant. Resultantly, AC-Ag/C achieved a FEethylene of over 95.0 ± 0.36% at a low potential of −0.08 V versus reversible hydrogen electrode (RHE) in an H-type cell with 0.5 M KCl electrolyte, alongside a near 100% selectivity within the range of −0.38 to −0.58 V versus RHE. Through this indirect strategy, the average ethylene purity within 6-hour electrolysis was 98.00 ± 1.45 wt%, at −0.48 V (vs RHE) from the neutralized electrolyte after CO2 reduction over the Cu/Cu2O catalyst in a flow-cell.


Supplementary Note 1
Herein, Ag was chosen as the catalyst for the reduction of Br-EO because its remarkable catalytic ability in cleaving the C-halogen bond.On the other hand, Ag-based catalysts, such as Ag nanoplates, nanowires, or single-atom, have also shown significant efficacy in CO2-to-CO conversion.But in our present investigation, no discernible formation of CO was observed via online GC analysis during the reduction of Br-EO.Several factors may account for this result.
Firstly, the potential window examined for the reduction of Br-EO (-0.08 ~ -0.58 V, vs RHE) does not encompass the reduction bias necessary for the generation of detectable concentrations of CO.We measured the reduction product distribution over Ag and AC-Ag electrode for pure CO2 reduction.As shown in Fig. S12a, the Ag foil electrode gave a detectable CO peak at the potentials of -0.78 and -0.88 V (vs RHE, the current density was insufficient to yield adequate products for GC analysis at more positive potentials).While for AC-Ag, only H2 was observed at -0.58 and -0.68 V (vs RHE, Fig. S12b).These results align with previously reported initial reduction potentials of Ag foil electrodes for CO2 reduction.To evaluate the activity of the AC-Ag/C catalyst for CO production, the LSV curve in CO2-saturated 0.5 M KCl was collected.It is evident that the initial potential, defined by the potential at which the current reaches 1 mA cm -2 , stands at -0.61 V (vs RHE, Fig. S12c).On-line GC analysis was conducted from -0.48 to -0.78 V (vs RHE), encompassing the applied potential for Br-EO reduction (-0.48 and -0.58 V, vs RHE).However, the appearance of the CO peak was not observed until the potential reached -0.68 V (vs RHE), suggesting that CO2 reduction over the catalysts employed in our study occurs at a potential outside the range of focus for the Br-EO reduction in this work.
Secondly, the presence of Br-EO in electrolytes will lead to the competitive adsorption between Br-EO and CO2, thereby suppressing the CO2 electroreduction.At the potential of -0.78 V (vs RHE), wherein CO can be discerned in CO2-saturated 0.5 M KCl, the CO peak disappeared upon the introduction of 50 mM Br-EO into the electrolyte (Fig. S12d).Instead, a pronounced peak denoting the presence of C2H4 emerges (Fig. S12e).Detailed analysis through in situ ATR-SEIRAS revealed that all traces of CO2 reduction intermediates disappeared, leaving only signals indicative of Br-EO reduction, in 0.5 M KCl with 50 mM Br-EO.These results collectively demonstrated that Br-EO in electrolytes exhibits a marked predilection for reduction, thereby implying a paucity of active sites available for the adsorption and subsequent reduction of CO2.

Supplementary Note 2
We assessed the electrochemical performance of the Br-EO reduction over pristine carbon black and Ag/C without electrochemical activation (denoted as Ag/C), in 0.5 M KCl while CO2 bubbled through.Within the potential range like AC-Ag/C (-0.18 ~ -0.58 V vs RHE), the pristine carbon black showed an almost negligible current response, exemplified by J registering at only -1.8 mA cm -2 at -0.58 V (vs RHE), which is significantly inferior to that of AC-Ag/C (Fig. S13a).Noteworthy is that ethylene production was only discerned at -0.48 and -0.58 V (vs RHE), with selectivity of 50.99% and 74.4%.Regarding the Ag/C sample, it demonstrated a significantly enhanced reductive current, boasting a J of -45.0 mA cm -2 at -0.58 V (vs RHE).A selectivity analysis revealed the near-exclusive production of ethylene, with only a marginal generation of H2, manifesting faradaic efficiency lower than 4.0% at all potentials (Fig. S13b).However, despite the notable high selectivity for ethylene observed in the case of Ag/C, its current densities are still lower than that of AC-Ag/C.These results can conclude that the excellent activity of AC-Ag/C is attributable to the presence of Ag particles, and the electrochemical activation can further improve the activity for Br-EO reduction.

Supplementary Note 3
Oleamine served as the capping agent to imbue stability upon the Ag nanoparticles.
Nevertheless, these residual organic molecules may influence the electrochemical behavior of the as-obtained Ag nanoparticles.According to the FTIR analysis, our synthesized Ag particles, in their as-yielded state, exhibit vestiges of oleamine adorning their surfaces, evidenced by the characteristic absorption peaks corresponding to pure oleamine (Fig. S14a).The TGA curve of pristine Ag particles, conducted under an Ar atmosphere, commences a mass loss at approximately 200 o C, ultimately registering a weight reduction of 22.72% (Fig. S14b).
To scrutinize the influence of lingering oleamine molecules on the reduction of Br-EO catalyzed by Ag/C, we subjected the Ag/C catalyst to treatment at 250 o C, hereinafter referred to as Ag/C-after.The removal of oleamine was substantiated by the disappearance of characteristic oleamine peaks in the FTIR spectrum, accompanied by an absence of mass loss in the TGA curve.To exclude the potential influence of electrochemical activation procedures on the residual surface oleamine molecule, we conducted a comparative analysis of the electrochemical performance between the pristine Ag/C and the Ag/C-after without electrochemical activation.In terms of current density, it was observed that Ag/C-after could deliver a slightly superior current when compared to the Ag/C counterpart (Fig. S14c), likely attributable to the enhanced exposure of active sites.However, regarding the selectivity for ethylene and H2, both Ag/C and Ag/C-after exhibited comparable Faradaic efficiencies (Fig. S14d).These findings collectively substantiate that the pronounced selectivity for the conversion of Br-EO to ethylene primarily emanates from the Ag particles themselves, rather than the remnants of oleamine on their surfaces.

Supplementary Note 6: Discussions about the possible acidic environment of anolyte
We first monitored the pH fluctuations of the anolyte post-electrolysis.A solution with saffron yellow was observed in the anolyte after 2 hours of electrolysis at -200 mA cm -2 , registering a pH of 5.80, indicative of a mild acidic environment (Fig. S40a).After standing for 19 hours, the pH of the anolyte increased to 6.10, subsequently ascending to 6.25 and 6.54 at 38 and 57 hours, respectively.The observed mild acidity, rather than a robust acidic environment, was partially ascribed to the cation exchange membrane (CEM) used in the electrolytic cell.The protons, that are generated in the anolyte, can partially traverse the CEM and interact with hydroxide ions (OH -) present in the catholyte.To substantiate this assertion, we established an electrolysis system featuring a cathodic reaction of CO2 reduction and anodic oxidation of ferrocyanide ions (Fe(CN)6 4-) (Fig. S40b).Notably, the oxidation of Fe(CN)6 4-does not entail the generation or depletion of protons, hence its corresponding influence on the pH variations of the anolyte can be discounted.The aqueous solution of 0.5 M KCl containing 25 mM H2SO4 was used as the anolyte (pH = 1.52, as presented in the upside section of Figure S40c).On the cathodic side, CO2 can react with KOH in the catholyte and generate KHCO3.The resultant HCO3 -may permeate the CEM and further interact with the protons present within the anolyte, potentially modulating its pH.Consequently, we scrutinized the pH of the anolyte following a 10-hour incubation period with continuous CO2 flow through the cathode chamber.The pH of the anolyte was detected to be 1.54 (downside section of Fig. S40c), almost mirroring that of the pristine electrolyte.This outcome attests to the marginal impact of HCO3 -crossover on the pH dynamics of the anolyte within our experimental framework.Conversely, after the electrolysis of the CO2 -Fe(CN)6 4-system under -50 mA cm -2 for 2 hours, the anolyte was collected, revealing a corresponding pH of 2.22 as indicated by the pH meter (Fig. S40d).This means the electric field-driven proton crossover can significantly buffer the pH reduction of anolyte, thereby preventing the formation of a strong acid environment.
Next, we endeavored to elucidate the impact of an acidic environment on the selectivity of Br-EO electroreduction.As delineated in Fig. S40e, the pH of a 0.5 M KCl aqueous solution was modulated by adding H2SO4 across varying concentrations.
In a dilute acidic KCl aqueous solution (CH2SO4 = 0.001 mM, pH = 5.7), the FEethylene still reached 98.6%.With the proton concentration increased to 50 mM (equal to the concentration of added Br-EO, pH = 1.59), we also observed a notable ethylene selectivity of 88.5%, and the FEethylene remaining above 50.0%even upon introducing 0.1 M H2SO4.In the electrolyte comprising 0.5 M KCl + 25 mM H2SO4 + 50 mM Br-EO, ethylene emerges as the principal product across the potential window ranging from -0.228 to -0.528 (Fig. S40f).These findings underscore the feasibility of achieving high-purity ethylene production within a mild acidic solution.
Finally, the alkaline catholyte is gradually added into the anolyte, to remove the residue Br2 and neutralize the formed acid, transforming the solution from saffron yellow to colorless, as depicted in the inset of Fig. S40g.Notably, the gradually added alkaline catholyte did not induce any changes in the Br-EO in the anolyte, as confirmed by the 1 H NMR (Fig. S40g).Subsequently, the electrolysis of this mixed solution by AC-Ag electrode gave the exclusive ethylene peak in the online GC, with the FEethylene exceeding 95.0%, with no identifiable peak for H2 (retention time of 2.25 min) observed (Fig. S40h-S40i).These results further support that the neutralization of anolyte by the catholyte of the first electrolytic cell can effectively remove the potential influence of a weak acid environment.The specific energy for electrochemical C2H4 production can be calculated via the following equation. [1]  W  Concerning potential remedies, firstly, insights can be gleaned from established practices in the bromine industry and chlor-alkali industry, including their storage technologies and secure production protocols.For example, the utilization of a sunblock strategy for the anode chamber is conceivable.Given our operations at room temperature and atmospheric pressure, glass-lined vessels can endure the rigors of Br2involved testing, while vessels with ceramic linings also prove viable.The oxidative corrosion potency of Cl2 surpasses that of Br2; hence, the design of electrolysis cells (comprising structure, separator, and texture) and electrode design can provide valuable insights for our Br2-involved indirect pathway.Notably, stable Pt and IrO2/TiO2 electrodes employed in the chlor-alkali industry can seamlessly integrate into our system.Furthermore, safety specifications established for the chlor-alkali industry offer guidance in mitigating safety concerns within our system.

Fig. S4
Fig. S4 The digital photos of Ag-foil electrode after oxidation and subsequent reduction step.

Fig. S5
Fig. S5 The XRD pattern of Ag foil electrode after oxidation in 0.5 M KCl electrolyte.

Fig. S9
Fig. S9 Area in TEM image for SEAD measurement of AC-Ag.

Fig. S10 5 M
Fig. S10 LSV curves for Br-EO reduction over different Ag-based catalysts in CO2saturated 0.5 M KCl aqueous solution.Scan rate is 10 mV s -1 .The size of Ag and AC-Ag are both 1 cm 2 .The mass loading for AC-Ag/C is 1 mg cm -2 .

Fig. S12
Fig. S12 Product characterizations.GC spectra of (a) pristine Ag electrode at -0.78 and -0.88 V (vs RHE) in 0.5 M KCl.(b) AC-Ag electrode at -0.58 and -0.68 V (vs RHE) in 0.5 M KCl.(c) LSV curve of AC-Ag/C catalyst in 0.5 M KCl saturated with CO2 (30 sccm).GC spectra of (d) FID for detecting CO, (e) FID for detecting ethylene of AC-Ag/C catalyst in 0.5 M KCl at different potentials and in 0.5 M KCl containing 50 mM

Fig. S13
Fig. S13 Electrochemical data for Ag/C electrode.(a) The J of carbon black, Ag/C, and AC-Ag/C for Br-EO reduction (50 mM) in 0.5 M KCl aqueous solution.(b) The Faradaic efficiency of H2 and ethylene for Ag/C.Error bars correspond to the standard deviation of three measurements.

Fig. S14
Fig. S14 Analysis for the potential influence of residue capping agent.(a) FTIR, (b) TGA curves of pristine Ag particle and Ag particle after thermal treatment.(c) Total current density, and (d) FE of H2 and ethylene for Ag/C and Ag/C after thermal treatment in 0.5 M KCl containing 50 mM Br-EO with CO2 bubbling (30 sccm).Error bars correspond to the standard deviation of three measurements.

Fig. S15
Fig. S15 FEethylene for AC-Ag electrode in CO2-and N2-saturated 0.5 M KCl with the gas flow rate of 30 sccm.

Fig. S16
Fig. S16 Total current density of Br-EO reduction with different concentrations for AC-Ag/C electrode in 0.5 M KCl aqueous solution.

Fig. S17
Fig. S17 FE of H2 and ethylene, as well as the total current density, for the Br-EO reduction tested by a flow-through cell over the AC-Ag/C catalyst, using 0.5 M KCl electrolyte with 50 mM Br-EO.

Fig
Fig. S18 Electrochemical active specific area analysis.Pb underpotential deposition of (a) Ag foil, (b) AC-Ag, and (c) AC-Ag/C.(d) Electrochemical active surface areas of different Ag electrodes.

Fig
Fig. S20 Activation energy analysis.(a) FEethylene of Ag foil, AC-Ag, and AC-Ag/C at -0.38 V (vs RHE) under different temperatures, in 0.5 M KCl with 50 mM Br-EO.The size for Ag foil and AC-Ag is 1 cm 2 .The mass loading for AC-Ag/C is 1 mg cm -2 .(b) Linear fitting of the natural logarithm of the ethylene partial current densities versus the inverse temperatures.

Fig. S21
Fig. S21 Variations of FEethylene with the changes of H2O concentration in DMSO-based electrolyte over AC-Ag/C electrode, with the Br-EO concentration of 50 mM.

Fig. S22
Fig. S22 Dependence of Jethylene and the concentration of KCl.The Br-EO concentration is 50 mM and the working electrode is AC-Ag/C with mass loading of 1 mg cm -2 .The potentials are iR compensated.

Fig. S29
Fig. S29 In situ Raman spectra for AC-Ag/C catalysts in 0.5 M KCl electrolyte with and without 50 mM Br-EO, from open circuit potential (OCP) to -0.78 V (vs RHE).

Fig.
Fig. S30 Ag-O interaction analysis.(a) LSV curve for Ag and AC-Ag electrode collected in N2-saturated 0.1 M KOH aqueous electrolyte.The scan rate is 10 mV s -1 .(b) The charge distribution over Br-EO molecule.

Fig. S37
Fig. S37 LSV curves for the reduction of ethanol, acetate, aldehyde, and Br-EO over Ag/C electrode, in 0.5 M KCl electrolyte.

Fig. S40
Fig. S40 Analysis for the potential influence of the acidic anolyte.(a) The color and pH of anolyte in the first electrolytic cell after standing for different intervals.(b) Electrolytic system design for evaluating the influence of cation exchange membrane on the pH variation of anolyte.The indication of pH meter for (c) 0.5 M KCl + 25 mM H2SO4(upside), 0.5 M KCl + 25 mM H2SO4 with CO2 flow through the cathode chamber for 10 h (downside), and (d) 0.5 M KCl + 25 mM H2SO4 + 1 M KFe(CN)6 after electrolysis.(e) FE of H2 and ethylene for AC-Ag electrode collected in 0.5 M KCl + 50 mM Br-EO with different concentrations of H2SO4 at -0.328 V. (f) FE of H2 and ethylene for AC-Ag electrode collected in 0.5 M KCl + 25 mM H2SO4 + 50 mM Br-EO.(g) The 1 H NMR of Br-EO in the pristine anolyte and after neutralization with KOH catholyte.Inset is the photo of the anolyte after neutralization.Online GC spectra of (h) FID, (i) TCD detector of the online GC for the Br-EO electroreduction in the neutralized anolyte.

Fig. S41 Supplementary Note 7 :
Fig. S41 Electrochemical performance of Br-EO reduction.(a) The required FEBr-EO of the indirect route for achieving the comparable energy to that of the direct route at different total currents.(b) The cell potentials for acquiring different current densities that collected by galvanostatic analysis.

Table S1 .
The FE and the purity of ethylene for Br-EO reduction in electrolytes with