Computational modeling and simulation is an essential partner for scientific discovery and innovation in advanced materials design in the 21st century. ASL’s materials research effort is directed toward integrating recent advances in computing and underlying theory to experimental materials research. This combination of experimental and theoretical research provides a significant competitive advantage in materials design and discovery efforts. In combination with characterization, synthesis, and testing facilities, computational modeling is designed to aid materials design and to reduce the time in identifying potential candidates suitable for the development of new technologies.

Computational manipulation of material properties requires a fundamental understanding of a variety of processes which are multidisciplinary in nature. Following are some of ASL’s areas of concentration:

• Changes in the materials structure and properties at extreme temperatures and pressures
• Interaction between reactive gases and solids
• Interaction between light and matter
• Changes in properties (electronic, chemical, etc.) with particle size, morphology, and dopants
• Catalytic activity of nanoscale materials and reactive surfaces
• Diffusion of molecules, ions, and solvents in solids and surfaces
• Atomistic origin of thermodynamic and kinetic properties.

Techniques Our research goal is to create models that represent the experimental reality as closely as possible. Suitability of a particular theoretical model depends on the area of interest. No universal model is effective for all length and timescales in complex chemical and physical processes. As a result, ASL researchers utilize multiple theoretical models and approaches:

• Density functional theory (DFT)
• First principles molecular dynamics (FPMD)
• 2D/3D Lagrangian and Eulerian finite element techniques for multigrain modeling and simulation in ductile and heterogeneous materials
• Arbitrary Lagrangian and Eulerian methods for mesoscale simulation
• Smooth particle hydrodynamics and discrete element modeling
• Classical force field simulations
• QM/MM methods for multi-scale modeling
• Perturbation theory and linear response theory.

High-performance Computing

Research benefits obtained through combining computational modeling and simulations include:

• High efficiency: a rapid approach to scientific "experimentation"
• Cost effectiveness: computational performance is continuously improving
• Synergy: multidisciplinary problems require linking experiments with simulations

Other research activities that take advantage of ISP’s computational facility include: Continuum and mesoscale modeling Molecular dynamics (Lawrence Livermore National Laboratory collaboration) Quantum or ab-initio calculations Molecular modeling and simulations for energy and catalysis Mesoscale simulations of sand penetration

To achieve the research benefits of modeling and simulations, high-performance computers are required. ASL computers include:

IBM Cluster – e1350

• 64 compute nodes, with 2 dual core Xeon Processors on each node (3.0 GHz, 4 MB, L2 Cache)
• 16 GB RAM per compute node
• 73 GB SAS Hard Drive (Scratch Space per node)
• 1.75 TB GPFS Shared Data Drive
• Infiniband® non-blocking 20 gbps switch
• Separate data storage and management nodes

SGI Altix - 4700

• 128 nodes, dual core 1.6 GHz Itanium2 Processors, 533 MHz, 8 MB Shared Memory (SMP)
• 4 GB RAM per node
• 730 GB Hard Drive
  

SGI Origin - 300

• 32 Nodes, 600 MHz 14K RISC, SMP
• 1 GB RAM per node
• 350 GB Hard Drive

Materials modeling at ASL is dedicated to transforming fundamental research ideas to realistic materials of tomorrow.