The semiconductor process begins with the photo process that transfers circuits onto a wafer as if "taking a picture," and after about 150 steps, it is completed into a chip in a total of three to four months. Here, you can experience that entire process firsthand.
4th, upon entering the Daegu Gyeongbuk Institute of Science and Technology (DGIST) Next-Generation Semiconductor Convergence Research Institute in Dalseong County, Daegu, a semiconductor manufacturing facility (FAB) clean room filled with yellow light came into view. HEPA filters to maintain cleanliness were installed on the ceiling, and directly below, a sub fab (sub FAB) was laid out with gas supply lines and pump equipment.
Lee Bong-ho, Head of Team of the Next-Generation Semiconductor Convergence Research Institute's Advanced Convergence Analysis Team, said, "We purposely tuned the lighting to yellow because of semiconductor materials that react to light," and explained, "The FAB's precision is at the 350 nm (nanometer; 1 nm is one-billionth of a meter) level, and the clean room's cleanliness is class 1000 (1,000 or fewer particles larger than 0.5 μm per cubic foot). While that differs from Samsung's 3 nm and class 1, it is sufficient for education and research."
Near the FAB, there was a space where students learn chip design using a design program connected to a supercomputer. A few steps farther led to an analysis room equipped with a scanning electron microscope (SEM), an aberration-corrected transmission electron microscope (TEM) capable of atomic-scale observation, and a focused ion beam (FIB) system for microfabricating samples.
Lee said, "There are not many universities that have a system where you can experience everything from semiconductor design to processing and analysis," and added, "Korea tends to focus on fragmentary technologies, but in line with DGIST's motto of cultivating convergent talent, we are striving to produce integrated talent equipped across semiconductor design, processing, and analysis capabilities."
DGIST said it will focus on three fields—physical AI, human digital twin, and quantum sensors—based on world-class research infrastructure that includes the Next-Generation Semiconductor Convergence Research Institute with a FAB, the laboratory animal center, and the supercomputing artificial intelligence (AI) education and research center.
DGIST President Lee Geon-woo has long said, "The major weakness of domestic universities is that they lack flagship fields." To find a solution, he formed a Future Strategy Field Discovery Committee by bringing together Yoon Ui-joon, president of the National Academy of Engineering of Korea, and 10 young DGIST researchers, and after a year of discussions starting in Jul. last year, finalized three major future strategic research fields.
Lee said, "From now on, wherever you are in the world, when you think of DGIST, I will make sure physical AI (artificial intelligence), human digital twin, and quantum sensors come to mind," adding, "Our goal is to become a top-tier research hub—one of the best in the world—in all three fields."
Physical AI refers to AI with a tangible form. Autonomous vehicles or robots equipped with AI are representative examples. Park Kyung-jun, DGIST planning director and head of the physical AI initiative, said, "In 2016, AI AlphaGo beat a human at Go, but a person still had to place the stones," and explained, "Now we have entered the era of 'physical AI,' where AI has a body, perceives and judges on its own, and carries through to action."
He cited the three elements of physical AI as perception (environment recognition), intelligence (analysis and decision), and action (physical action), and said, "Although Korea fell behind in the AI race, it can instead leverage its hardware strengths to take the lead during the shift to physical AI."
DGIST has faculty and staff ranked No. 2 in Korea in AI by citation count, as well as a robotics-specialized department and infrastructure such as the national robot test field and the robot industry cluster. The institute also plans to establish a dedicated AI college jointly across the institute.
The second strategy is the "human digital twin." If a digital twin is a technology that implements objects or systems from the real world in virtual space for simulation, a human digital twin is a technology that replicates the human body in virtual space to predict diseases and present treatments.
Lee Kyung-tae, a professor in DGIST's Department of Electrical Engineering, Electronics, and Computer Science, said, "If existing medicine relied on patients' experiential data, a digital twin enables customized prediction and experimentation," adding, "It will bring tremendous change to new drug development and personalized treatment."
DGIST's roadmap unfolds in three stages over 10 years. By 2028, it will secure diverse biometric data and turn them into a database. By 2031, it will complete AI models for the human digital twin. Finally, by 2035, through global standardization and ecosystem expansion, it aims to establish the technology as one that brings innovation to the medical and pharmaceutical industries. It has already signed agreements with Yeungnam University Medical Center, Kyungpook National University Hospital, Seoul National University Hospital, and the Korea Brain Research Institute.
The third strategy is quantum sensors. This technology measures physical quantities with far higher precision and sensitivity than existing sensors by using principles of quantum mechanics that apply to the microscopic world. Quantum sensors are called one of the three major quantum technologies along with quantum computing and quantum communication. Quantum sensors capable of precisely capturing extremely minute biological responses, such as brain signals, are expected to be used across various industries in the future, including medicine, space, and defense.
Yoo Cheon-yeol, head of DGIST's quantum sensing promotion strategy team, said, "Quantum sensors are the core foundational technology that make physical AI and digital twins possible," adding, "DGIST has already established the Department of Quantum Information Science to foster talent, and based on world-class semiconductor fabrication infrastructure, we are ramping up research on quantum devices and sensors."