Introduction
Computational materials design is a pivotal force in accelerating the transition towards a sustainable energy future. This in silico approach, now powerfully augmented by the integration of artificial intelligence (AI), significantly reduces the time and cost associated with traditional experimentation. This, in turn, accelerates the identification and optimization of high-performance materials essential for developing next-generation energy technologies. These innovations span a wide spectrum, from advanced biofuels, environment-friendly components in batteries, and efficient perovskite solar cells, to catalysts for green hydrogen production/storage and novel materials for carbon capture. Critically, this design paradigm inherently allows for the proactive consideration of resource availability and environmental impact, ensuring that materials innovation aligns squarely with our global energy challenges and the building of a circular materials economy.
Research Focus
-
Mesoscale and Multiphysics Modeling
Connection of atomic scale averages to the mesoscale using coarse-grained, discrete element, fluid dynamics, and finite element models; multiphysics models of materials processing and manufacturing.
-
First-Principles and Molecular Dynamics Simulations
Density functional theory (DFT) for solids, alloys, and compounds; classical, reactive, and coarse-grain molecular dynamics of materials.
-
Data and Machine Learning
High-throughput DFT databases, automation of simulations and learning from molecular dynamics, machine learning for the acceleration of first-principles thermodynamic calculations, machine learning for materials property predictions, reinforcement learning for steering of experiments and manufacturing processes.
-
High-performance Computing
Integration of edge-to-exascale computing and machine learning acceleration hardware in advanced materials design, scale-up, and manufacturing.
M3L
Data-enabled Discovery of Materials for New Energy Economy
Develop an ability to steer searches compounds, alloys, and metastable states with unique properties using genetic algorithms for generations of crystal structure guesses with progressively better structural motif or stability.
- A search of alloys and compounds in the phase diagram
- Materials for a circular economy
- Design of corrosion-resistant microstructure
- Nanostructured coatings
- Design of molecular and hybrid crystals
- Additive-manufacturing
- 3D printing of electronics and sensors
- Bio-based and bio-sourced materials such as bioplastics
Creating high performance materials
As articulated in a recent DOE report “The ability to predict and control mesoscale phenomena and architectures is essential if atomic and molecular knowledge is to blossom into the next generation of technology opportunities, societal benefits, and scientific advances”. The challenge of modeling phenomena in the mesoscale is scaling the system size and the physics to the boundaries of nano-and-microscale where continuum behaviors emerge. Due to recent advances in precision in 3D printing, mesoscale assembly of materials through proper control of transport, reactions, and phase segregation processes are possible. However, there remain large gaps in understanding mesoscale. The outcome of the research at M3L is an ability to predict Structure-Property-Performance-and-Processing (SPPP) correlations in complex mesoscale architectures under dynamic thermo-chemo-mechanical conditions and bridge fundamental science constructs to the engineering scale.
Selected Publications
Hyun Park, Xiaoli Yan, Ruijie Zhu, Eliu A Huerta, Santanu Chaudhuri, Donny Cooper, Ian Foster, Emad Tajkhorshid: A generative artificial intelligence framework based on a molecular diffusion model for the design of metal-organic frameworks for carbon capture.
https://www.nature.com/articles/s42004-023-01090-2