Structure-Performance Descriptors and the Role of the Axial Oxygen Atom on M–N₄–C Single-Atom Catalysts for Electrochemical CO₂ Reduction

Abstract

Revealing and characterizing the catalytic sites, along with elucidating a convenient activity descriptor, can provide essential guidance in determining efficient electrocatalytic catalysts for the CO2 reduction reaction (CO2RR). In this work, the mechanism of CO2 reduction to methane (CH4) on 23 transition metal-coordinated nitrogen-doped carbon M–N4–C single-atom catalysts (SACs) was studied by density functional theory calculations, a step forward to revealing the effects of the axial O atom (M–N4O–C) on their catalytic activity. The electrocatalytic reduction activity of CO2 over M–N4–C SACs is strongly dependent on the outmost d-shell electron numbers and electronegativity of the selected metals. The introduction of the axial O atom changes the coordination structure of the central metal atoms, which not only improves the stability of M–N4O–C SACs (especially electrochemical stability) but also affects the adsorption strength of intermediate species and then improves or reduces the catalytic activity, which depends on the intrinsic properties of the metal atoms. More importantly, by considering the comprehensive effects of the number of outmost d-shell electrons, the electronegativity, coordinate numbers, and bonding length of the central metal atom and the nearest neighbor atom, a descriptor (φ) based on the intrinsic properties of materials was developed to correlate the catalytic activity. The volcano-shaped relationships between the φ and limiting potentials were well established. In particular, five SACs (Mn–N4–C, Cr–N4–C, Os–N4O–C, Ru–N4O–C, and Rh–N4O–C) close to the summit of the volcano were screened. Based on this descriptor, the catalyst activity can be predicted directly from the characteristics of the material instead of the expensive calculation of adsorption energies. This work is of great significance for understanding the mechanism of electrocatalytic CO2RR and the design of efficient and stable electrocatalysts.

A Machine Learning Model To Predict CO2 Reduction Reactivity and Products Transferred from Metal-Zeolites

“…Among these methods, CO 2 reduction reaction (CO 2 RR) by photocatalytic and electrocatalytic reactions has attracted considerable attention due to its high environmental compatibility and mild reaction conditions. Dispersing metal atoms on special sites of host supports has been applied in CO 2 RR catalysts due to highly active atomic utilization and unsaturated coordination environments of the metal center. − Great achievements have been made in experiments and calculations with various catalysts, such as metal-phthalocyanines (M-Pcs), − porpyridinic-like M–Nx complexes, − metal–organic frameworks (MOFs), − porous organic polymers, , and zeolites. − However, the interactions between the metal center and supported materials may lead to different pathways and hence the generation of different products (CO, CH 2 O, HCOOH, CH 3 OH, CH 4 , C 2 H 5 OH, C 2 H 4 , etc. ), resulting in a big challenge for precise control of reduction products.…”

“…To further test the applicability of the prediction model, the free energy changes were predicted on 2D, ,, 3D materials, , and molecular complexes for CO 2 RR (Figure S35). As shown in Figure b, the overall free energy trend for an external test on the MOF system, Mo 3 (HAB) 2 , was consistent with the calculated results.…”

“…A good agreement was shown between the ML-predicted CT values and the DFT computational values (Figure a). We also employed the above six descriptors to predict the CT values of some reported catalysts for CO 2 RR, such as, molecular complexes, M-N/C (M-N 4 , M-N 4 O, and M-g-CN), , and MOF materials. − The transferable potential for CT prediction in many other systems was exhibited from the satisfactory external test performance with an MAE of 0.045 e and R 2 of 0.84 (Figure S38).…”

“…Subsequently, the ML model was applied to molecular complexes, M-N/C (M-N 4 , M-N 4 O, and M-g-CN), , and MOFs, − a total of 34 catalysts for the prediction of final products in CO 2 RR. We divide the 0–1 scale bar into three parts, where the P -values less than 0.45 and greater than 0.55 indicate that the final products were CH 3 OH and CH 4 , respectively, and the P -values between 0.45 and 0.55 indicate that the final products were both likely CH 3 OH and CH 4 .…”

Surface Co‐Modification of Halide Anions and Potassium Cations Promotes High‐Rate CO2‐to‐Ethanol Electrosynthesis

“…In addition, the calculated dissolution potentials ( U dis ) of the Cu in KCuF 3 (110) and Cu(111) subsurface were all higher than −0.1 V versus RHE (Table S10, Supporting Information) that were significantly positive than the experimental operating potentials, indicating their excellent electrochemical stability. [ 45,46 ] According to our experimental and theoretical results, most of the K cations from KCuF 3 were dissolved and removed by electrolytes during the electrochemical reconstruction, while the remaining portion was stabilized via strong chemical bonding of KF and FCu on the surface.…”