Wooyoung Shim

Department of Materials Science and Engineering, Yonsei University,
Seoul 03722, South Korea
Fax: +82-(2)-312-5357 E-mail address: wshim@yonsei.ac.kr

One of the intriguing questions in materials science is how elements can be combined to form a solid with desired properties. One part of the answer to this question can be derived from the fundamental relationship between the composition, structure, and basic bonding in materials. This lecture starts with a description of the fundamental atomic structure connected to the structure stability and correlated properties. These aspects will be covered in the context of thermodynamics, crystallography, and solid state physics that complete the correlations between the structures and properties and beyond. Pursuing the discovery and development of advanced material systems lies in this deep understanding of the materials fundamentals, which is also crucial to achieving global competitiveness of material research.

Sang Soo Han

Computational Science Research Center, Korea Institute of Science and Technology,
Seoul 02792, South Korea
Fax: +82-(2)-958-5451 E-mail address: sangsoo@kist.re.kr

Physical laws or principles that define the fundamental connections between a material’s composition and its structure and function have been usually used as the foundation for analytical or numerical models, and these principles-based approaches provide a route to design material candidates with respect to properties of interest.1 Recently, evolution of the keywords of big data and artificial intelligence in the material science and engineering, called data-driven paradigm, are opening new paths to the understanding, design, and engineering of next-generation materials systems. Compared with the principles-based approach, the data-driven approach can efficiently help to uncover a correlation between material structures and properties or the underlying mechanism, and then accelerate material design. In this talk, I will review and discuss the recent status of the data-driven approach including our examples regarding design of novel catalysts. A machine-learning technique is recently using for the inverse design – engineering a novel material with particular properties. Here, I will also cover the issue.

References
1. A. Ferguson and J. Hachmann, Mol. Syst. Des. Eng. 3, 429 (2018).

Yunseok Kim^1*

¹ School of Advanced Materials Science and Engineering, Sungkyunkwan University (SKKU). 2066 Seobu-ro, Jangan-gu, Suwon, Gyeonggi-do 16419, Republic of Korea
E-mail address: yunseokkim@skku.edu

The microstructual, physical, and chemical properties of materials have been explored by various microscopy techniques. As the research focus on the sizes or dimensions has shifted from bulk to nanosized materials, the microscopy techniques with better signal and spatial resolutions are continuously developing. Nonetheless, the interpretation of the microscopy images is still insufficient to fully understand materials properties due to technically limited signal or spatial resolution of the microscopies. Very recently, machine learning based analysis of the microscopy images have received considerable attention because it allows overcoming the current drawbacks of the microscopy techniques. The improved signal or spatial resolution could allow better interpretation of the microscopy images. In this tutorial, I will discuss about recent progress on the machine learning based analysis of the microscopy images measured by transmission electron microscopy (TEM), scanning tunneling microscopy (STM) and atomic force microscopy (AFM). In particular, I will also cover deep learning based analysis for undertstanding physical property and atomic structural imagings.

Ki Tae Nam

Department of Materials Science and Engineering, Seoul National University, Seoul 151-744, South Korea

For decades, as global population and energy demands increased, electrocatalysis has been regarded as an attractive approach for sustainable energy conversion, which converts earth-abundant molecules (H2O, CO2, N2, etc) to higher-value products (H2, hydrocarbons, ammonia, etc) with renewable energy. Electrocatalysts serves ad key materials to improve performance and reduce overall cost in practical application. Thus, substantial research efforts have been devoted for development of highly efficient electrocatalysts. To achieve superior catalytic performance, rational design rule for electrocatalyst synthesis should be established by mechanistic investigation and descriptors for catalytic behavior.

Water splitting is a promising energy conversion process towards environmentally sustainable energy schemes because electrolysis produces only hydrogen and oxygen, without any by-products. The oxygen evolution reaction (OER), an anodic half-cell reaction, requires extremely high overpotential due to its slow reaction kinetics. Interestingly, the water oxidizing cluster in PS II, cubical Mn4CaO5 cluster, efficiently catalyzes water oxidation. Inspired by Mn4CaO5 cluster, specific questions that we intensively focus for further applications include how to translate the underlying principles in Mn4CaO5 cluster into synthetic heterogeneous catalysts. Toward this vision, we have been developing a new catalytic platform based on sub-10 nm-sized MnO nanoparticles (MnO NPs) to bridge the gap between atomically defined biological catalysts. We captured key reaction intermediates and demonstrated their unique OER mechanism through combined electrokinetic and in-situ spectroscopic analysis.

For the next step for investigating the reaction mechanism and finding key descriptors for electrocatalysis, machine learning techniques can be a powerful tool for identifying the reaction intermediates and surface changes, combining with conventional in situ spectroscopic approaches. Currently, the machine learning algorithm has applied to DFT-based calculation for determination of catalyst descriptors, which boosts the speed and efficiency of calculations. In this tutorial, we summarize current advances of Mn-based OER catalysts and their mechanistic investigations. We also briefly introduce some examples of machine learning based catalyst researches, and suggest an insightful perspective of in situ spectroscopy combined with machine learning interpretation.

Sang-Joon Cho

Park Systems Corp, KANC 4F, Gwanggyo-ro 109 Suwon, South Korea
* Corresponding author: msjcho@parksystems.com

Abstract: Atomic Force Microscope (AFM) is a powerful instrument in characterizing nanoscale features, but it is known to lack accuracy and repeatability in measuring absolute dimensions. However, the importance of AFM analysis is growing due to the strong necessity to investigate and characterize innovative nanomaterials. Finding new materials with innovative characteristics in nanoscale have helped guide many industries to grow and the newly found materials have contributed breakthroughs in sectors such as energy, semiconductor industry, life science, etc. Quantitatively characterizing electrical, magnetic, mechanical, and morphological properties of these materials are major concerns for both research and industrial sectors. The quantitative analysis by AFM could open the door to new ways to control the development of the material's facets and help increase efficiency and decrease failure and cost.