Accelerated discovery of NO₃RR single-atom catalysts via high-throughput DFT and machine learning

Area of Science:

• Materials Science
• Computational Chemistry
• Catalysis

Background:

• Nitrate (NO3RR) reduction is crucial for environmental remediation and sustainable chemistry.
• Developing efficient electrocatalysts for NO3RR is an ongoing challenge.

Purpose of the Study:

• To identify highly active and selective electrocatalysts for nitrate reduction.
• To elucidate the fundamental mechanisms governing nitrate electroreduction.

Main Methods:

• High-throughput density functional theory (DFT) calculations were employed to screen potential catalysts.
• Interpretable machine learning (ML), specifically SHAP analysis, was used to understand catalytic activity.
• Calculations focused on metal-substituted graphitic carbon nitride (g-C2N and g-C2NC) materials.

Main Results:

• Zirconium/graphitic carbon nitride (Zr/g-C2N) and Titanium/graphitic carbon nitride (Ti/g-C2NC) were identified as promising NO3RR catalysts.
• The limiting free-energy changes for NO3RR on Zr/g-C2N and Ti/g-C2NC were calculated to be 0.30 eV and 0.34 eV, respectively.
• SHAP analysis highlighted the importance of metal electronic structure, charge transfer, and adsorbate bonding in nitrate activation and hydrogenation.

Conclusions:

• Zr/g-C2N and Ti/g-C2NC represent highly active electrocatalysts for nitrate reduction.
• The study provides fundamental insights into the electronic and structural factors governing NO3RR catalysis.
• This work paves the way for rational catalyst design for efficient nitrate conversion.