Operando Studies of Electrochemical Denitrogenation and Its Mitigation of N-Doped Carbon Catalysts in Alkaline Media

Abstract

N-doped carbons (NCs) have excellent electrocatalytic performance in oxygen reduction reaction, particularly in alkaline conditions, showing great promise of replacing commercial Pt/C catalysts in fuel cells and metal-air batteries. However, NCs are vulnerable when biased at high potentials, which suffer from denitrogenation and carbon corrosion. Such material degradation drastically undermines the activity, yet its dynamic evolution in response to the applied potentials is challenging to examine experimentally. In this work, we used differential electrochemical mass spectroscopy coupled with an optimized cell and observed the dynamic behaviors of NCs under operando conditions in KOH electrolyte. The corrosion of carbon occurred at ca. 1.2 V vs RHE, which was >0.3 V below the measured onset potential of water oxidation. Denitrogenation proceeded in parallel with carbon corrosion, releasing both NO and NO₂. Combined with the ex situ characterizations and density-functional theory calculations, we identified that the pyridinic nitrogen moieties were particularly in peril. Three denitrogenation pathways were also proposed. Finally, we demonstrated that transferring the oxidation reaction sites to the well-deposited metal hydroxide with optimized loading was effective in suppressing the N leaching. This work showed the dynamic evolution of NC under potential bias and might cast light on understanding and mitigating NC deactivation for practical applications.

Figure 1

(a) LSV curve of NC with Faradaic currents from both OER and NC oxidation; (b) CV curves of NC after 0, 10, 50, and 100 cycles from 0.87 to 2.03 V vs RHE. (c) LSV curves of NC after various CV cycles at 1600 rpm; (d) comparison of E_1/2 and electron transfer number of NC in ORR. All measurements are performed in oxygen saturated 0.1 M KOH electrolyte at 25 °C using a rotating disc electrode.

Figure 2

Comparison of NC, NC-O₁₀, NC-O₅₀, and NC-O₁₀₀ properties containing (a, b) TEM micrographs of NC and NC-O₁₀₀, the insets demonstrate the mesopores; (c, d) deconvoluted core-level XPS spectra of N 1s and C 1s; (e) FTIR spectra; and (f) Raman spectra.

Figure 3

(a, b) Schematic configurational comparison of two DEMS cells; the CV curves and the corresponding ionic currents obtained from (c) probe cell and (d) Au film cell; measurements are performed in oxygen saturated 0.1 M KOH electrolyte at 25 °C with a scan rate of 10 mV s^–1.

Figure 4

(a) Multiple CV cycles and the corresponding ionic currents obtained in the Au-film cell with (solid line) and without (dotted line) the deposition of NC; (b) enlarged graphs of the square area in (a) showing the onset potentials of each gas evolution reaction; (c) potentiostatic analysis with stepwise potential increase of 0.25 V per step and the corresponding ionic current. All measurements are performed in oxygen saturated 0.1 M KOH electrolyte at 25 °C, the voltage window is between 0.87 and 2.12 V vs RHE, and the scan rate is 10 mV s^–1.

Figure 5

(a) Configurations of selected oxidation sites of NC used in the DFT calculation; N atoms are orange in color; (b) plot of the calculated oxidation free energy of various moieties; and (c) hypothetic routes of NC denitrogenation.

Figure 6

(a) Schematic illustration of CFNC with various loadings of CoFe(OH)_x; and (b–e) corresponding TEM micrographs.

Figure 7

(a–c) Selected ionic currents from DEMS measurement of NC and CFNC with different loadings of metal hydroxides; the subscript in the sample name is the mass loading in percentage; (d) LSV curves of pristine and oxidized catalysts at 1600 rpm; the oxidation was performed via applying a constant potential at 2 V vs RHE for different periods of time, and the subscript in the sample name is time in minute; (e) comparison of E_1/2 and electron transfer number; (f) N 1s core-level XPS spectra of CFNC and CFNC-C₁₂₀. All measurements are performed in oxygen saturated 0.1 M KOH electrolyte at 25 °C.