-
대학/대학원
-
연구/산학
-
대학생활
-
입학/취업
-
커뮤니티
-
대학소개
-
개교 80주년
연구/산학
PKNU Research 1000
New Beginning, New Inspiration
Pukyong National University, the first university in Busan,always paves a new path through specialized and converged research to lead the era of the fourth indutrial revolution.
| Yeo Byung-Chul | Identifying Next-Generation Semiconductor Memory Material Candidates Using Generative AI | |||
| 작성자 | 대외홍보센터 | 작성일 | 2026-04-09 |
| 조회수 | 59 | ||
| Yeo Byung-Chul | Identifying Next-Generation Semiconductor Memory Material Candidates Using Generative AI | |||||
 |
대외홍보센터 |  |
2026-04-09 |  |
59 |
Identifying Next-Generation Semiconductor Memory Material Candidates Using Generative AI
- Joint research by Pukyong National University and KIST proposes new ferroelectric materials applicable to next-generation semiconductor devices; published in Advanced Science
Professor Yeo Byung-Chul of the Department of Energy Resources Engineering at Pukyong National University (President Sang-Hoon Bae) announced that, through joint research with Dr. Lee Hyun-Jae, Dr. Kang Sung-Woo, and Dr. Lee Jung-Hoon of the Korea Institute of Science and Technology (KIST), the team has identified candidate materials for next-generation semiconductor memory using generative artificial intelligence (AI). The research findings were published in the international journal Advanced Science (Impact Factor: 14.1).
To overcome the limitations of conventional semiconductor memory material discovery-previously restricted to a small number of materials-the research team proposed a new materials design framework that combines a diffusion model-based generative AI for atomic structure generation with computational science techniques.
Using the generative AI model, the team generated a large number of candidate crystal structures and applied a multi-step validation process integrating various machine learning methods with density functional theory (DFT) calculations.
As a result, the research team identified two promising materials-Ca₃P₂ (calcium phosphide) and LiCdP (lithium cadmium phosphide). Both materials exhibit electrical insulation properties while also demonstrating characteristics of ferroelectric materials, which can switch their electric polarization states in response to external stimuli, confirming their potential as next-generation memory devices.
In particular, LiCdP showed material properties comparable to-or even surpassing-those of existing high-performance polarization materials. Meanwhile, Ca₃P₂ is significant in that it presents a previously unreported low-temperature crystal structure candidate. Furthermore, detailed electronic structure calculations revealed that both materials hold strong potential for energy applications, such as photocurrent devices based on solar energy.
This study demonstrates that generative AI can be used to discover previously unknown functional materials and is expected to serve as a key enabling technology for the development of next-generation semiconductor memory, energy devices, and advanced electronic components.
The research team stated, “This study shows that new ferroelectric materials can be efficiently explored by combining generative AI with computational science,” adding, “We will continue to pursue research on the development of next-generation semiconductor and energy materials.”
Meanwhile, this research was supported by the Nano and Materials Technology Development Program (Materials Global Young Connect / National Strategic Technology Materials Development_HUB) of the National Research Foundation of Korea, funded by the Ministry of Science and ICT.
