The research team from Department of AI Semiconductor Engineering at Korea University Sejong Campus has been featured in the ‘Journal of Materials Chemistry C’ of the Royal Society of Chemistry for their cover paper on semiconductor design using artificial intelligence. This study is expected to secure foundational technologies for AI-driven semiconductor device design and make significant contributions to the transformation of the electronic design automation (EDA) industry.

△ From left: Professor Lee Ohjun (Catholic University of Korea, Department of Artificial Intelligence), Researcher Jo Minseon (Korea University, Institute of Industrial Technology, First Author), Professor Jung Sungyeop (Korea University, Department of AI Semiconductor Engineering and Institute of Industrial Technology, Corresponding Author), and Intern Researcher Marin Franot (Advanced Institute of Convergence Technology, original affiliation: Toulouse INP-ENSEEIHT, Co-Author)

 

The research team includes Professor Jung Sungyeop (Korea University, Department of AI Semiconductor Engineering and Institute of Industrial Technology, Corresponding Author), Researcher Jo Minseon (Korea University, Institute of Industrial Technology, First Author), Professor Lee Ohjun (Catholic University of Korea, Department of Artificial Intelligence, Co-Author), and Intern Researcher Marin Franot (Advanced Institute of Convergence Technology, original affiliation: Toulouse INP-ENSEEIHT, Co-Author).

 

This research utilized artificial neural networks and AI to model the electrical properties of organic semiconductor-based field-effect transistors. Traditionally, compact models for semiconductor device based on formulas rely on computer codes that explains the motions of the device and mathematical formulas embedded in circuit design software to describe device behavior and predict and evaluate device and circuit performance.

Developing compact models for new materials or structures that precisely, succinctly, and physically represent the electrical properties of semiconductor devices often requires significant time and resources. To overcome these limitations, the researchers used artificial neural networks and transfer learning to develop models that accurately reflect the physical mechanisms of semiconductor devices and predict electrical properties in a shorter time.

 

Researcher Jo, the first author of the paper stated, “Unlike conventional formula-based transistor modeling, we leveraged the function approximation capabilities of artificial neural networks and deep learning to construct more accurate and faster models. This method also efficiently and accurately predicted the nonlinear changes in electrical properties of amorphous semiconductors, such as charge mobility variations with temperature, electric field, and charge density.”

 

Professor Jung, the corresponding author, stated, “This study presents a new method in which artificial intelligence learns the physics of charge transport in semiconductors and, based on this, accurately predicts device characteristics. It is explained that this approach can also be applied to the modeling of next-generation DRAM access transistors or resistive memory devices for artificial synapses, which are based on oxide semiconductors with similar charge transport characteristics."

 

 

Professor Lee, a co-researcher, referenced the examples of this year's Nobel Prizes in Physics and Chemistry, which were awarded for foundational AI technology research and AI-applied protein structure prediction. He stated, "We anticipate that AI application research will become even more active across the entire semiconductor field in the future."

△ Cover of the Journal of Materials Chemistry C

This study was supported by the ‘Next-Generation Intelligent Semiconductor Technology Development (Device) Project’ (Project Number NRF-2022M3F3A2A01076569) and the ‘Basic Science Research Project’ (NRF-2021R1F1A1064384 and NRF-2022R1F1A1065516). The paper’s title is: A neural compact model based on transfer learning for organic FETs with Gaussian disorder.

 

KU Sejong Student PR Team, KUS-ON

Translator: Seo Yujeong