Computational Discovery and Design of Materials : Session IV
Sponsored by: TMS Materials Processing and Manufacturing Division, TMS: Computational Materials Science and Engineering Committee, TMS: Integrated Computational Materials Engineering Committee
Program Organizers: Houlong Zhuang, Arizona State University; Duyu Chen, University of California, Santa Barbara; Ismaila Dabo, Pennsylvania State University; Yang Jiao, Arizona State University; Sara Kadkhodaei, University Of Illinois Chicago; Mahesh Neupane, Army Research Laboratory; Xiaofeng Qian, Texas A&M University; Arunima Singh, Arizona State University; Natasha Vermaak, Lehigh University
Tuesday 2:30 PM
March 21, 2023
Room: Cobalt 502A
Location: Hilton
Session Chair: Natasha Vermaak, Lehigh University; Xiaofeng Qian, Texas A&M University
2:30 PM Invited
Applying Data-driven Models in Materials Science: Unraveling Hidden Relationships between Structures and Properties: Mingjie Liu1; 1University of Florida
Designing materials at atomic scale is promising to develop new materials for energy applications. However, we are still facing the challenges from huge materials space to explore, mismatch between simulations and experiments, and multiobjective design requirements. Data science and machine learnings are demonstrated as useful tools to tackle those challenges. In this talk, I will introduce the work of applying data-driven models in carbon materials design with examples of single-atom catalysts and ion-selective membrane. From those examples, I will show how the data-driven models can be used to accelerate the exploration of the materials space from atomic scale.
3:00 PM Invited
Computational Design for Metallic Meso-architected Materials for Dynamics: H Alicia Kim1; Brianna McNider1; Ryan Fancher1; Po-Shun Chiu1; Jaeyub Hyun1; Nicholas Boechler1; 1University of California, San Diego
We present our latest research in formulating a computational design strategy for mitigating damage resulting from high-strain rate loading events. It is well-known that unprecedented properties can be achieved by tailoring nonlinearities derived from designed mesostructure. Several examples of materials with designed nonlinearity have shown outstanding promise for shock and impact mitigation. Yet, the examples of mesostructure-derived effective material nonlinearities are few and far between. Two primary challenges exist: (1) nonlinearity dramatically expands this unintuitive design space compared to linear material problems, introducing a plague of local minima and non-uniqueness; (2) even if we had an answer to (1), it is not guaranteed that the best nonlinear material can be manufactured. The researchers of the Center for DREAMS (Dynamically Responsive Emergent Architected Material Systems) in the UC system develop a computational approach via mathematical optimization to readily discover the mesostructures of architected materials with desired effective nonlinearity.
3:30 PM Cancelled
Crystal Structure Generation using Wasserstein Generative Adversarial Network: Zahra Gholami Shiri1; Michael Alverson1; Taylor Sparks1; Hasan Sayeed1; 1University of Utah
The commonplace approach of discovering new materials is experimentation while materials informatics can accelerate innovation. Herein we have implemented deep learning techniques to generate inorganic solid materials using Fourier-Transformed Crystal Properties (FTCP) framework. We have developed two Wasserstein Generative Adversarial Networks (WGAN) models using Artificial (ANN) and Convolution Neural Networks (CNN) to predict new crystal structures. Each model has been trained on more than forty thousand crystallographic information files collected from the Materials Project computational database. It has been demonstrated that the convolution network outperforms the artificial network. Generated structures by ANN suffer from overlapping atoms in 3-dimensional space and unrealistic lattice parameters. Employing convolution networks has addressed this issue. The WGAN model using convolution architecture and FTCP representation has the potential to pave the way for inverse design of stable inorganic solid materials.
3:50 PM Break
4:10 PM Invited
Atomistic Modeling of Electronic Transport and Electrochemistry: Yuanyue Liu1; 1University of Texas at Austin
I will present our first-principles studies of (1) carrier transport in 2D semiconductors. I will discuss what limits the carrier mobility in current 2D semiconductors, and present new materials with very high mobility discovered through high-throughput screening; [1] (2) electrocatalytic mechanism of single metal atom in N doped graphene. The conventional models often oversimply the complexity of realistic electrochemical interface, and do not offer give kinetic information especial for electrochemical steps. I will present an advanced model to simulate the reaction kinetics at electrochemical interface, and its success in explaining the mechanisms of Ni-N-C for electrochemical CO2 reduction and Co-N-C for oxygen reduction reaction. [2][1] PRL ’20, JACS ’19, JACS ‘18 [2] Chem. Rev. ‘22, JACS ’21, JACS ’20, JACS ’18
4:40 PM
Design and Development of High Strength High Conductivity Alloys using ICMDŽ Approach: Qiaofu Zhang1; Tom Kozmel1; Peter Jacobson1; Jiadong Gong1; Greg Olson1; 1QuesTek Innovations LLC
Electrification is the dominant strategy for decarbonization of industry in order to meet the greenhouse gas reduction goal. This electrification trend requires novel materials developments of materials/alloys with high conductivity to reduce energy losses, and high strength to provide structural support and adoption to various conditions. Using QuesTek’s Materials By DesignŽ approach, several HSHC alloys have been designed and developed for different applications. At the computational design stage, ICME models were developed to connect the correlation link of process-microstructure-properties, and alloys were designed using the developed models and QuesTek’s in-house Integrated Computational Materials Design (ICMDŽ) platform. Experimental prototyping were performed to achieve the desired microstructure and target properties to validate the design. Three types of alloys, Cu-alloys, Al-alloys, and carbon-nanotube (CNT) additive materials will be discussed in this presentation.