Seminar & Colloquium
[세미나: 4월 13일(수), 오전 9시 30분] Fritz-Haber-Institut der Max-Planck-Gesellschaft, 장우선 박사
[Title]
An integrated digital materials design ecosystem: A sans pareil solution via data-centric flexi-architectonics for green energy harvesting
[Speaker]
장우선 박사, Postdoctoral Research Fellow, Fritz-Haber-Institut der Max-Planck-Gesellschaft, Berlin, Germany
[Education]
- Feb 2019 Ph. D. in Materials Science and Engineering, Yonsei University, Seoul, Republic of Korea
Implementing Nanoarchitectonics in Low-Dimensional Materials: Computational Materials Exploration and Design, advised by Prof. Aloysius Soon
- Feb 2016 M. S. in Materials Science and Engineering, Yonsei University, Seoul, Republic of Korea
An Ab Initio Study of Structurally Modified Low-Dimensional Phosphorus, advised by Prof. Aloysius Soon
- Feb 2014 B. S. in Materials Science and Engineering, Yonsei University, Seoul, Republic of Korea
Platinum Atomic Layer on Titanium Nitride Support
[Professional experiences]
- Mar 2022 NRF Sejong Research Fellow,
- Feb 2020 Alexander von Humboldt Research Fellow, Fritz-Haber-Institut der Max-Planck-Gesellschaft, Berlin, Germany
- Mar 2019 Affiliated Research Fellow, Center for Artificial Synesthesia Materials Discovery, Yonsei University, Seoul, South Korea
- Nov 2018 Postdoctoral Research Fellow, Fritz-Haber-Institut der Max-Planck-Gesellschaft, Berlin, Germany
| Date | Wednesday, April 13th, 2022
| Time | 9:30 ~
| Venue | 온라인 강의 (https://snu-ac-kr.zoom.us/j/98245303500)
회의 ID: 982 4530 3500
[Abstract]
Scientific data are a significant raw material of the 21st century. Engineering materials by design is proven to be much effective than the conventional approaches - mainly the brute-force type approach, well known as a trial-and-error approach. A much more promising approach ? the data-driven approach, "big data" from experiments and simulations is collected and exploited via a open-source database, from which new knowledge for materials design is mined and refined by means of a high-throughput process.
Yet, these data-driven processes are not so-called "intelligent" by design. Thus, a combination of high-quality, curated data (from experiments and simulations), well-managed digital infrastructure, and artificial intelligence (AI) tools expands the basic science and engineering of materials. This seamlessly-integrated digital materials design ecosystem aims to provide the next-generation strategy to discover novel functional materials, for example, from first-principles. In this talk, I will apply the de novo flexi-architectonics for green energy harvesting using low-dimensional heterosystems.
Coupling first-principles electronic structure calculations with data-driven artificial (and natural collective) intelligent algorithms, I will explain how this approach can be efficiently used to search for multimodal global optimization solutions to active metastable low-dimensional heterosystems.
Extending both the time-and-length scale in describing the amorphous nature of key technological materials, I will demonstrate how kernel-based machine-learning potentials are developed from extensive ab initio molecular dynamics calculations for surface oxide of liquid metals and ultrathin films of carbon.
To exploit the value of heterogeneous data, a FAIR - Findable, Accessible, Interoperable, and Re-purposable - data infrastructure is essential. To provide perspective, I will provide as an outlook how a data-informed AI model for in-situ X-ray spectroscopy enables one to characterize and analyze low-dimensional transition metal (hydr)oxides for green energy harvesting.
By these examples, this talk will aim to provide how such an integrated digital materials design ecosystem can provide a sans pareil solution to targeted low-dimensional heterosystems by flexi-architectonics for green energy harvesting.
| Host | Prof. Seungwu Han (883-1520)