Title:

Can Unknown Materials Properties Be Accurately Predicted without Parameters?

Speaker:

Prof. Zi-Kui Liu

Affiliation:

Pennsylvania State University

When:

Friday, March 29, 2024 at 11:00:00 AM   Add to Calendar

Where:

MRDC Building, Room 4211

Host:

Ting Zhu
ting.zhu@me.gatech.edu

Abstract

In today's science, the only predictive theory is quantum mechanics. With several well-defined approximations, the density functional theory (DFT) provides the numerical solution for the many-body interactions in quantum mechanics. DFT represents the outcome of those interactions and articulates that for a given system, there exists a ground-state configuration at 0 K that its energy is at its minimum value with a universal functional of the interacting electron gas density. However, up to date, quantitative agreement between DFT-based predictions and experiments is lacking in the literature due to the focus of DFT on the ground-state configuration and a few non-ground-state configurations. These bottom-up approach have not been able to capture all conceivable configurations that the system embraces at finite temperature as stipulated by the top-down Gibbs statistical mechanics. Furthermore, Gibbs statistical mechanics does not include the entropy contributions from individual configurations. Since Gibbs thermodynamics is applicable to equilibrium systems only and thus misses the irreversible internal processes in nature, resulted in the separate development of phenomenological irreversible thermodynamics. Hillert integrated them together by first emphasizing their differences and then examining their connections. Over the last two decades, the author's group developed a multiscale entropy approach (recently termed as zentropy theory) that integrates DFT-based quantum mechanics and Gibbs statistical mechanics plus replacing the total energy of individual configurations by their free energies. Zentropy theory has demonstrated its capability to accurately predict entropy and free energy of magnetic materials and is being applied to ferroelectric materials and superconductors. Furthermore, in combination with the combined law for nonequilibrium systems developed by Hillert, the author developed the theory of cross phenomena beyond the phenomenological Onsager Theorem for irreversible thermodynamics. The zentropy theory and theory of cross phenomena jointly provide quantitative predictive theories for experimental observables and will be discussed in the presentation.

Biography

Dr. Zi-Kui Liu is the Dorothy Pate Enright Professor at the Department of Materials Science and Engineering, College of Earth and Mineral Science, The Pennsylvania State University. He obtained his BS from Central South University (China), MS from University of Science and Technology Beijing (China), PhD from Royal Institute of Technology (KTH, Sweden). He was a research associate at University of Wisconsin-Madison and a senior research scientist at Questek Innovation, LLC. He has been at the Pennsylvania State University since 1999, the Editor-in-Chief of CALPHAD journal since 2001, and the President of CALPHAD, Inc. since 2013. He co-founded the NSF Center for Computational Materials Design and served its director from 2005 to 2014. Dr. Liu coined the name 'Materials Genome' in 2002 and led the incorporation of the nonprofit Materials Genome Foundation in 2018, and his company, Materials Genome, Inc., owns its trademark. Dr. Liu is a Fellow of TMS and ASM International. He served as the President of ASM International and a member of ASM International Board of Trustees and the TMS Board of Directors. He received the ASM J. Willard Gibbs Phase Equilibria Award, the TMS William Hume-Rothery Award, the ACers Spriggs Phase Equilibria Award, the Wilson Award for Excellence in Research from the Pennsylvania State University, and the Lee Hsun Award from Institute of Metals Research, Chinese Academy of Science. Dr. Liu's current research activities are centered on (1) DFT-based first-principles calculations and deep neural network machine learning for prediction and modeling of materials properties through integration of quantum, statistical, and irreversible thermodynamics, and (2) their applications for designing and tailoring materials chemistry, processing, and performances.