AI semiconductor plasma-based integrated process KIMM develops intelligent system for 2D semiconductor manufacturing

A technology has been developed that enables artificial intelligence (AI) to autonomously analyze and control ultra-thin next-generation 2D semiconductors at the atomic-layer level based on process diagnostic results. Previously, differences in process conditions often caused quality variations, making uniform production difficult. This research establishes a foundation for simultaneously improving process reproducibility and productivity by combining plasma processing technology with AI-based data analysis. In particular, the integration of synthesis, etching, monitoring, and process prediction into a single system provides a technological foundation for the future automation and intelligentization of semiconductor manufacturing processes.

A research team led by Senior Researcher Hyeong-U Kim of the Semiconductor Manufacturing Research Center at the Korea Institute of Machinery and Materials (President Seok-Hyeon Ryu, hereinafter referred to as KIMM) developed synthesis and etching processes for 6-inch next-generation 2D semiconductors (MoS₂ and WS₂) using low-temperature plasma-based PECVD and RIE equipment and implemented them into an AI-based intelligent system.

   *PECVD: Plasma Enhanced Chemical Vapor Deposition

   **RIE: Reactive Ion Etcher

The research team developed a technology capable of predicting process state by measuring light emission and gas mass variations generated during the process in real time and analyzing them using machine learning techniques. In particular, the team acquired time-series multimodal data using various real-time diagnostic tools such as OES, ToF-MS, and QMS, and successfully applied them to machine learning models to precisely predict semiconductor thickness at the atomic-layer level.

   *OES: Optical Emission Spectroscopy

   **ToF-MS: Time of Flight Mass Spectrometer

   ***QMS: Quadrupole Mass Spectrometer

Conventional next-generation 2D semiconductor processes have primarily relied on high-temperature methods, resulting in low compatibility with existing semiconductor production lines and limitations in achieving large-area uniform processing. In addition, atomic-layer etching technologies often required long process times and suffered from low productivity. In contrast, this research distinguishes itself by adopting a low-temperature plasma-based process compatible with existing mass-production equipment while implementing atomic-layer etching through a single-process approach, thereby improving both process efficiency and productivity.

This technology is expected to be applicable across various semiconductor industries, including AI semiconductors, next-generation electronic devices, and displays. In addition, because the technology utilizes the existing OES viewports of commercial production equipment, it can be implemented without structural modifications to the equipment, significantly enhancing its industrial applicability. As process data continue to accumulate, the technology is also expected to evolve into a key platform for autonomous and intelligent semiconductor manufacturing.

Senior Researcher Hyeong-U Kim stated, “This research is meaningful in that it successfully implemented 2D semiconductor processes on 6-inch wafers at the atomic-layer level under low-temperature conditions.” He added, “By applying multimodal data-based machine learning technologies, we simultaneously achieved process prediction and optimization, significantly improving both process reproducibility and productivity.” He further noted, “We plan to expand this technology into a core platform for the AI intelligentization of next-generation semiconductor manufacturing framework.”

This research was supported by KIMM’s Creative Challenge Research Program and Young Frontier Program, KIMM-SKKU academic collaboration, KIMM’s institutional research program, and the Ministry of Trade, Industry and Energy’s semiconductor workforce development project. Through a total of five research projects, the team has continuously advanced related technologies and produced more than 30 SCI journal publications.

###

The Korea Institute of Machinery and Materials (KIMM) is a non-profit government-funded research institute under the Ministry of Science and ICT. Since its foundation in 1976, KIMM is contributing to economic growth of the nation by performing R&D on key technologies in machinery and materials, conducting reliability test evaluation, and commercializing the developed products and technologies.