Data-Driven Inorganic Materials Design Tool

Data-Driven Inorganic Materials Design Tool

AtomWork Creator is an inverse-design platform developed for solid-state battery materials. By combining machine learning models with composition-space search algorithms, it automatically proposes new material compositions exhibiting high ionic conductivity. Using training data built from AtomWork Battery, it trains an NGBoost model and predicts ionic conductivity using 34-dimensional periodic-table descriptors as features. The search engine employs Bayesian optimization based on the Markov chain Monte Carlo (MCMC) method, enabling efficient exploration of a vast composition space while accounting for chemical constraints such as allowed elements, forbidden elements, and charge-neutrality conditions.

Other Protocols

• Electrode particle design based on reaction transport models
• Machine learning for evaluation of electrode structure and battery characteristics
• 3D Structure Analysis
• Operando Electrode Reaction Analysis
• Operando Nano X-ray CT Imaging
• Operando Soft X-ray Absorption Spectroscopy
• Operando Mechanical Characterization of Battery Cross‑Sections
• Explosion-Proof Operando X-ray CT Imaging
• Rapid safety screening with small batteries
• Autonomous Experimentation System for Electrolyte Materials Discovery
• High-Throughput Pouch Cell Production and Evaluation Platform
• Cathode & Solid Electrolyte Database