Abstract
Carbon materials underpin a wide range of energy and sustainability technologies, including batteries, supercapacitors, electrocatalysts, adsorbents, and separation membranes. However, rationally designing these materials remains difficult. The carbonization process is rarely resolved at the atomic level, and the resulting structures are often too disordered to be described completely. This dual opacity has long forced researchers to rely on empirical optimization. Materials informatics (MI) offers a practical way to extract useful structure‐property relationships from heterogeneous data even when a full mechanistic understanding is lacking. In this Perspective, we discuss how AI‐assisted literature mapping, regression analysis, machine‐learning‐based structural modeling, image‐based characterization, and multimodal data integration are reshaping carbon materials research. Through representative case studies, we show that MI is valuable not only for prediction but also for organizing complex datasets, identifying governing factors, and guiding more rational design strategies. Further progress will require tighter links among data‐driven methods, physically meaningful descriptors, experimentally grounded workflows, and sustainability‐oriented materials development.
Funding Information
  • Japan Science and Technology Agency (CREST: JPMJCR24S6)
  • Japan Science and Technology Agency (GteX: JPMJGX24S0)
  • New Energy and Industrial Technology Development Organization (JPNP21005)
  • New Energy and Industrial Technology Development Organization (JPNP25002)
  • Japan Society for the Promotion of Science (24K17761)
  • Japan Society for the Promotion of Science (23H00227)
  • Japan Society for the Promotion of Science (26KJ0572)

This publication has 90 references indexed in Scilit: