Simulations & Digital Technology

Development and discovery of new materials using Materials Informatics (MI)

Development and discovery of new materials using MI

Materials Informatics (MI) is a promising field of study where vast amounts of materials data are processed by AI and deep learning. Although materials R&D activities have long been carried based on extensive experiments and researchers' intuition, the scope of MI applications is rapidly expanding.
ENEOS' goal is to discover and develop innovative materials for a sustainable low carbon future in the fields of renewable energies, catalysts, functional materials and lubricants. By integrating experiments, simulations and AI, ENEOS aim to realize materials development through "physicochemical laws" and "statistical analysis".

AI x Simulation Platform: Matlantis™

Recent advances in simulations and AI technologies drastically accelerate materials R&D. In-silico atomistic-scale simulations predicting materials properties are suitable for MI to generate and analyze huge amount of materials data.
ENEOS have developed the universal atomic-level simulator Matlantis™ in collaboration with Preferred Networks (PFN). In 2021, these two companies established a joint-venture called Preferred Computational Chemistry (PFCC) to offer Matlantis™ as a SaaS solution. Matlantis™ computes materials properties at extreme speed with high accuracy. It can predict the structures and properties of molecules, crystals, amorphous structures, interfaces and even macromolecules.
ENEOS are also focusing research efforts on low carbon energy carriers such as hydrogen, accelerating the pace of materials R&D with Matlantis™.

Case Study 1: Virtual Screening of Catalysts for Ammonia Synthesis

To enhance catalyst design strategy, reaction efficiencies of several alloy catalysts were calculated using Matlantis™ and the results analyzed by means of reaction rate simulations. Candidate materials exhibiting high reaction efficiencies were discovered within a week, while similar calculations would have taken years using conventional simulation tools.

Virtual screening of 2,000 alloy catalysts
Analysis of catalytic activity using Matlantis™

Case Study 2:Design of Lubricants and Greases

With Matlantis™, ENEOS design lubricants and greases to improve the efficiency and performance of mechanical systems and reduce their environmental impact. For instance, by identifying the molecular structures of additives that enhance friction and wear properties or by clarifying chemical reactions that are crucial for wear resistance through the elucidation of complex phenomena. Based on these simulation results, ENEOS are promoting the design of ideal lubricants and greases that can achieve the desired tribological properties.

Analysis of adsorption energy of lubricant additives
Simulation of tribochemical reactions of lubricant additives

Case Study 3: Multi-Scale Simulation×AI for Tire Rubber Material Development

ENEOS employ Matlantis™ and other simulation tools to develop rubber materials for tires with complex hierarchical structures. Finite Element Methods (FEM) and molecular dynamics (MD) are conventionally used to develop rubber materials, but Matlantis™ has further expanded the computational domain to the atomic scale. The combination of these simulation techniques at various scales with machine learning (AI) is driving the development of materials that contribute to improved tire performance.

Hierarchical structure of rubber materials for tires and computational domain of Multi-Scale simulation

Related Links

Original Paper

  1. 1.Takamoto, S., Shinagawa, C., Motoki, D. et al. Towards universal neural network potential for material discovery applicable to arbitrary combination of 45 elements. Nat Commun 13, 2991(2022).


  1. 1.PFN Research & Development Blog Development of Universal Neural Network for Materials Discovery
  3. 3.Matlantis™ - YouTube

Hommoku Insights (R&D News, Japanese)

  1. 1.汎用原子レベルシミュレータ「Matlantis™」を活用した3つの研究成果についてトライボロジー会議2022秋で発表
  2. 2.研究成果が世界的に権威のある論文誌「Nature Communications」に掲載

Press Releases (Japanese)

  1. 1.Matlantis のコア技術「PFP」に関する論文が Nature Communications の Editor’s Highlights に選出
  2. 2.PFCC、新物質開発や材料探索を高速化する 汎用原子レベルシミュレータMatlantisをクラウドサービスとして提供開始
  3. 3.新物質開発や材料探索を加速する 汎用原子レベルシミュレータを提供する合弁会社の設立に合意
  4. 4.株式会社Preferred Networksとの協業について~AI技術を活用したJXTGグループの事業強化・創造に向けて~