A Geometric–Delocalization Model for Material Classification in the M–D Parameter Space
Abstract
This study introduces a two-parameter geometric–electronic framework for classifying materials based on structural regularity and electronic delocalization. The model uses two normalized indices: (1) M_geom, capturing coordination environment, crystallographic symmetry, and density; and (2) D_delok, measuring electronic delocalization through hybridization character, π-orbital participation, dimensionality, and band-gap behavior. Using Materials Project structures for seven representative materials —including metallic alloys, carbon allotropes, silicates, and actinide compounds—we demonstrate that the proposed model consistently separates metals, semimetals, and insulators in the M–D phase space. The approach provides a compact, data-consistent alternative to traditional organic–inorganic classification schemes and suggests a more universal structure–property mapping framework. Language editing assistance was provided using an AI-based tool. Originality and AI-use statement: This work is an original research output by Begüm Yıldırım. AI tools, if used, were limited to language refinement, grammar correction, formatting, translation assistance, and clarity improvement. The conceptual framework, research direction, interpretation, models, and conclusions belong to the author. External sources, datasets, or prior works are cited where applicable.
Fields of Study & Subjects
Machine Learning in Materials Science
Advanced Chemical Physics Studies
Thermodynamic and Structural Properties of Metals and Alloys
Publication details
- Journal / source
- Zenodo (CERN European Organization for Nuclear Research)
- Publication date
- 13 January 2026
- Document type
- Preprint