Official websites use .gov
A .gov website belongs to an official government organization in the United States.
Secure .gov websites use HTTPS
A lock (
) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.
Summary
Computational materials design requires a variety of tools to model processing-structure-property relationships across a range of time and length scales. This work focuses on the development of these types of tools to support specific materials design applications specifically:
- Tools to facilitate the integration of data and models across multiple length scales
- Tools to enable evaluation of new thermodynamic models (particularly new sub-lattice models and lattice stability functions)
- Tools to enable high-throughput atomistic and DFT simulations and model development
Description
Microstructure-level Structure-Property Tools
OOF: Finite Element Analysis of Microstructures enables materials scientists calculate macroscopic properties from images of real or simulated microstructures. It reads an image, assigns material properties to features in the image, and conducts virtual experiments to determine the macroscopic properties of the microstructure. More information is available here.
pyMKS A python-based framework of the Materials Knowledge System (MKS) is a data science approach for solving multiscale materials science problems using physics, machine learning, regression analysis, signal processing, and spatial statistics to create processing-structure-property relationship.
Phase-based Modeling Tools
OpenCalphad: Multicomponent thermodynamic open-source software for performing a variety multicomponent multiphase calculations. The software is constructed to enable researchers to develop new thermodynamic models and assess model parameters for thermodynamic databases to describe experimental data as well as theoretical results from DFT calculations to calculate phase equilibria and phase diagrams.
Atomistic Simulations and Potential Development Tools
Physically-informed Neural Network (PINN) potentials is a novel method combining physics-based knowledge and neural networks to determine descriptions of the bonding forces between atoms.
pyFit-FF: A python package for training artificial neural network interatomic potentials (including PINN potentials), which uses the PyTorch optimization library.
Atomman: The Atomistic Manipulation Toolkit is a Python library for creating, representing, manipulating, and analyzing large-scale atomic systems of atom.
IPRPy: High-Throughput Computational Framework is a collection of tools and resources supporting the design of scientific calculations for evaluating basic materials properties across multiple classical interatomic potentials.
Density Functional Theory (DFT) Tools
JARVIS-Tools: A package of scripts used in generating and analyzing the datasets from JARVIS-DFT and JARVIS-FF data repositories.
JARVIS-Heterostructure: A set of tools for 2D materials in the JARVIS-DFT database. Some of the properties available are: work function, band-alignment, and heterostructure classification.
JARVIS-WannierTB: This tool predicts the Wannier Tight-binding hamiltonian derived properties of 3D and 2D materials from the pre-computed JARVIS-DFT database.