Atomically dispersed asymmetric cobalt electrocatalyst for efficient hydrogen peroxide production in neutral media

Electrochemical hydrogen peroxide (H2O2) production (EHPP) via a two-electron oxygen reduction reaction (2e- ORR) provides a promising alternative to replace the energy-intensive anthraquinone process. M-N-C electrocatalysts, which consist of atomically dispersed transition metals and nitrogen-doped carbon, have demonstrated considerable EHPP efficiency. However, their full potential, particularly regarding the correlation between structural configurations and performances in neutral media, remains underexplored. Herein, a series of ultralow metal-loading M-N-C electrocatalysts are synthesized and investigated for the EHPP process in the neutral electrolyte. CoNCB material with the asymmetric Co-C/N/O configuration exhibits the highest EHPP activity and selectivity among various as-prepared M-N-C electrocatalyst, with an outstanding mass activity (6.1 × 105 A gCo−1 at 0.5 V vs. RHE), and a high practical H2O2 production rate (4.72 mol gcatalyst−1 h−1 cm−2). Compared with the popularly recognized square-planar symmetric Co-N4 configuration, the superiority of asymmetric Co-C/N/O configurations is elucidated by X-ray absorption fine structure spectroscopy analysis and computational studies.


Non-aqueous
Others (2)    Supplementary Note 1.The identification of N K-edge features was struggling due to the contamination from the beamline setup.The contamination was not supposed to be present on the clean bare Au substrate.Silicon window was used to measure this sample at B07-B beamline at Diamond Light Source, to avoid the N K-edge interference from silicon nitride window.The contamination was probably due to the glue used for the silicon window assembly after discussion with beamline scientists.We apologize for this as we cannot find a good way to obtain the real N K-edge at the moment.This supplementary note is to prevent causing misleading understanding.Fitting range: 3.1 ≤ k (Å -1 ) ≤ 12.3; 1 ≤ R (Å) ≤ 2.6.
Due to the limit of independent points, only the combination of two paths was considered here.All the attempts showed well-matched fitting results, demonstrating the infeasibility of EXAFS fitting in identifying the accurate coordination information of C/N/O.Moreover, the ± 20% error bounds in the fitting parameters make it more challenging to identify the accurate coordination number and distance of cobalt species in CoNCB based on the EXAFS analysis.To the best of our knowledge, the same issues exhibit in all the metal-C/N/O materials prepared by the annealing method and without known electronic structure, and there is not a feasible way to solve this issue currently for the metal-C/N/O materials with very low loading of metal.However, this issue has been consciously ignored by previous research in which many ambiguous or even misleading conclusions were reported.

Figure S2 .
Figure S2.STEM images of CoNCB.(a) HAADF-STEM image and (b) ABF-STEM image of CoNCB.The magnification is 60k and scale bar is 50 nm.

Figure S4 .
Figure S4.STEM images of NCB.(a) ABF-and (b) HAADF-STEM images of metal-free NCB material.Scale bar of both: 2 nm.

Figure S6 .
Figure S6.Schematic drawings of CoPr structure.(a) front view and top view.(b) Simplified illustration of D4h symmetry.(c) Schematic illustration of 3d orbital splitting of low spin CoPr with D4h symmetry.

Figure S11 .
Figure S11.EXAFS fitting of Co foil.(a) k 2 -weighted extracted EXAFS signal of experimental result of Co foil and related fitting result using Co-Co path.(b) FT magnitudes of k 2 -weighted EXAFS spectra without phase correction.(c) Real part of FT magnitudes.

Figure S12 .
Figure S12.EXAFS fitting of CoNCB.(a) k 2 -weighted extracted EXAFS signal of experimental result of CoNCB and related fitting result using Co-N path and Co-O path.(b) FT magnitudes of k 2 -weighted EXAFS spectra without phase correction.(c) Real part of FT magnitudes.(d) Fitted structure model..

Figure S14 .
Figure S14.EXAFS fitting of CoNCB.(a) k 2 -weighted extracted EXAFS signal of experimental result of CoNCB and related fitting result using Co-C path and Co-N path.(b) FT magnitudes of k 2 -weighted EXAFS spectra without phase correction.(c) Real part of FT magnitudes.(d) Fitted structure model.

Figure S15 .
Figure S15.Calculated Tafel plots from disk LSV current density; labels show the corresponding Tafel slopes.Koutecky-Levich equation was used with 5.8 mA cm -2 as the theoretical limiting current of 4e -ORR process.

Figure S16 .
Figure S16.Effect of Co loading.(a) ORR polarization curves of disk current density (solid line) and ring current density (dash line) of CoNCB with different Co loading content in 0.1 M PBS (pH = 7).(b) Calculated H2O2 selectivity (H2O2 %).The optimal loading content is determined to be 0.05 wt % with both high activity and selectivity.

Figure S17 .
Figure S17.Effect of catalyst loading.(a) ORR polarization curves of disk current density (solid line) and ring current density (dash line) of CoNCB with different catalyst loading in 0.1 M PBS (pH = 7).(b) Calculated H2O2 selectivity (H2O2 %).They exhibited similar ring current density in the potential range of 0.55-0.75eV, but the H2O2 selectivity significantly decreased for catalyst loading of 40 ug/cm 2 , indicatingthat the generated peroxide was trapped within catalyst layer and further reduced to H2O.Therefore, we opted for a very thin layer of catalyst (4 ug/cm 2 ).

Figure S19 .
Figure S19.Flow cell set-up.(a) The digital picture of the flow cell set-up.(b) The digital picture of initial CoNCB@Gas diffusion electrode fixed with silicon membrane.(c) The digital picture of spent CoNCB@Gas diffusion electrode fixed with silicon membrane.The leaching electrolyte droplet would happen after long-time running, which would inhibit the oxygen transfer and decrease the active catalyst areas.

Figure S20 .
Figure S20.UV-vis titration test.(a) UV-vis absorbance spectra of standard Ce(SO4)2 solutions (up to 0.5 mM) in 0.5 M H2SO4.(b) The linear calibration curve (shown as an inset) at the peak wavelength (319 nm).(c) UV-vis absorbance spectra of Ce 4+ solutions after reacting with electrolytes at different times.(d) Measured working potential with time at the fixed current density of 20 mA cm -2 .

Figure S23 .
Figure S23.Experimental and simulated XANES spectra of Co foil and CoPr.The simulated (sim) XANES spectra of Co foil and CoPr reference aligned well with the experimental (exp) spectra.

Table S3 . Fitting parameters of Co-K edge EXAFS of Co foil.
S0 2 is the amplitude reduction factor; R is the interatomic distance (the distance between absorber and backscatter atoms) in the fitting results; C.N. is the coordination number; σ 2 is Debye-Waller factor (a measure of thermal and static disorder in adsoberscatterer distances); ΔE0 is the edge-energy shift; R factor is used to evaluate the goodness of the fitting results.

Table S4 . Fitting parameters of Co-K edge EXAFS of CoNCB.*
* The σ 2 value is set to be the same for each path in the fitting process.The ΔE0 value is set to be the same for each path in the fitting process.S0 2 was fixed to be 0.74 determined by Co foil fitting.