Machine Learning Prediction of Two-Dimensional Polymerization of Nitrogen in FeN_x.

Area of Science:

• Materials Science
• Computational Chemistry
• Solid State Physics

Background:

• Nitrogen-rich iron nitrides (FeNx) are of interest for their unique polymeric nitrogen structures and properties.
• Accurate modeling of transition metal nitrides using DFT+U requires careful selection of U values.
• Previous FeNx research had not reported polymerized nitrogen structures beyond one-dimensional.

Purpose of the Study:

• To develop a machine learning-integrated DFT+U (DFT+U_ML) approach for modeling FeNx systems.
• To systematically explore the FeNx (x = 1, 2, 4, 6, 8, or 10) system.
• To discover novel FeNx phases and resolve existing controversies in their ground states.

Main Methods:

• Developed a machine learning-integrated DFT+U (DFT+U_ML) approach.
• Applied DFT+U_ML to systematically explore the FeNx system.
• Investigated FeNx phases for stability, mechanical, and energetic properties.

Main Results:

• Reported a novel two-dimensional nitrogen-polymerized phase, P21/c-FeN4.
• Resolved the controversy regarding the ambient-pressure ground state of FeN.
• The P21/c-FeN4 phase is stable at 0 GPa and shows enhanced energetic and mechanical properties compared to 1D FeNx.

Conclusions:

• The DFT+U_ML approach provides a robust method for modeling transition metal nitrides.
• The novel P21/c-FeN4 phase presents potential for energetic material applications.
• This study offers new strategies for designing nitrogen-rich energetic materials and guiding synthesis.