Professor Jong-Ho Lee's Research Team at SNU Achieves Remarkable Success with Three Papers Accepted to IEEE IEDM 2025
Developing Breakthrough Technologies in Ultra-Low-Power Memory, Hardware Security, and Intelligent Sensors to Advance Next-Generation Semiconductors



▲ (Clockwise from top left) Prof. Jong-Ho Lee (Department of Electrical and Computer Engineering, former Minister of Science and ICT), Prof. Jae-Joon Kim (ECE), PhD candidate Kyung Min Lee (ECE), PhD candidate Sangwoo Ryu (ECE), Prof. Gyuweon Jung (School of Transdisciplinary Innovations), Dr. Jinwoo Park (ECE), Postdoctoral Researcher Ryunhan Gu (ECE), and PhD candidate Hoonhee Shin (ECE)

Seoul National University College of Engineering announced that a research team led by Prof. Jong-Ho Lee of the Department of Electrical and Computer Engineering—who formerly served as Korea’s Minister of Science and ICT—has achieved the remarkable achievements of having three research papers accepted to the IEEE International Electron Devices Meeting (IEDM) 2025, the world's most prestigious conference in semiconductor device technology.

The accepted research spans three core areas of next-generation semiconductor innovation:
▲ Ultra-low-power memory devices
▲ Hardware-based security chips
▲ Intelligent gas-sensing systems
These achievements reflect the strength of integrated research across materials, devices, and systems.

As advanced technologies such as artificial intelligence (AI), the Internet of Things (IoT), and autonomous driving continue to expand, the need for low-power and high-reliability semiconductor technologies is rising rapidly. However, conventional semiconductor devices face significant limitations, including high power consumption, weak security, and reduced sensing accuracy under real-world environmental conditions.

To address these challenges, Prof. Lee’s research team has focused on developing:
▲ Highly efficient non-volatile memory,
▲ Unclonable hardware security chips, and
▲ Selective gas-sensing systems capable of operating in complex environments.
Their achievements have now been recognized with the acceptance of three papers at IEEE IEDM 2025.

The three papers present:
▲ A key device technology for low-power neuromorphic computing,
▲ A security technology leveraging intrinsic semiconductor characteristics, and
▲ A low-power, high-performance sensor system
—all of which are expected to become foundational technologies for future integrated AI systems.

In the first study, the team introduced a major advancement in ferroelectric memory, a type of non-volatile memory. By incorporating a pseudomorphic structure into the thin-film layers surrounding hafnium–zirconium oxide (HfZrO₂, HZO), a well-known ferroelectric material, the researchers achieved world-leading polarization characteristics, and extremely low operating voltage. This breakthrough demonstrates the potential for implementing ultra-low-power memory and hardware-based neuromorphic AI architectures.

The second study utilizes another next-generation non-volatile memory technology, the magnetic tunnel junction (MTJ), to develop an innovative physical unclonable function (PUF). The team demonstrated that by exploiting the breakdown probability characteristics of MTJ devices, it is possible to reliably generate and store unclonable hardware security keys, maintain strong reliability even at high temperatures. This technology offers a promising path toward high-security semiconductor solutions for future electronic systems.

In collaboration with Prof. Gyuweon Jung of SNU’s School of Transdisciplinary Innovations, the team achieved a global first by developing a mixed-gas discrimination system that integrates both gas sensors and circuits on a single chip. The core innovation is a Self-Cancellation (SC) circuit architecture applied to low-power sensors, enabling ultra-miniaturized and ultra-low-power sensor-system implementation. The system can selectively detect target gases within complex mixtures without external signal processing, and in demonstrations, it successfully monitored egg freshness in real time.

All three technologies are compatible with standard CMOS semiconductor process technologies, which gives them strong potential for commercialization. The ultra-low-power memory developed by the team can support next-generation low-power AI and neuromorphic chips, while the hardware-based security technology offers a reliable solution for authentication and encryption in devices such as smartphones, automobiles, and IoT systems. In addition, the intelligent sensor system can be incorporated into a wide range of applications, including portable electronics, food-safety monitoring, environmental sensing, healthcare, and industrial-safety platforms. By establishing foundational technologies for low-power, high-security, and high-intelligence semiconductors, these achievements are expected to contribute significantly to strengthening Korea’s national semiconductor competitiveness.

Prof. Jong-Ho Lee stated, “These studies highlight the value of a convergent approach combining materials, devices, circuits, and systems. They demonstrate that Korean semiconductor technology can achieve world-class competitiveness across AI, security, and sensing applications.” He added, “We will continue research efforts to ensure that these breakthroughs—from basic materials research to fabrication-compatible demonstrations—lead to practical next-generation semiconductor technologies.”

Sangwoo Ryu, first author of the first study, is affiliated with the Department of Electrical and Computer Engineering at Seoul National University and is pursuing his PhD while employed at SK Hynix.

Kyung Min Lee, first author of the second study, is also affiliated with the Department of Electrical and Computer Engineering at Seoul National University and is pursuing his PhD while enrolled in Samsung Electronics’ academic training program.

Ryunhan Gu, first author of both the first and second studies, earned his PhD in the Department of Electrical and Computer Engineering at Seoul National University and is now a postdoctoral researcher at SNU’s Inter-University Semiconductor Research Center (ISRC).

Hoonhee Shin, first author of the third study, is enrolled in the integrated master–PhD program in the Department of Electrical and Computer Engineering at Seoul National University.

Jinwoo Park, co–first author of the third study, received his PhD from the Department of Electrical and Computer Engineering at Seoul National University and is currently working at Samsung Electronics.


▲ Fig. 1. Concept and advantages of proposed psuedomorphic HTH-HZO stack. HTH interlayer-guided HZO growth enhances ferroelectric properties and enables a large memory window in FeCap operation.


▲ Fig. 2. (a) Illustration of the proposed MBD-PUF concept and its key features. (b) Comparison of the proposed method with non-volatile memory-based PUF schemes. (c) Overview of the chip architecture, and the formation process of the PUF response and the concealment method.


▲ Fig. 3. (a) Schematic illustration of proposed sensor-circuit integrated computing (SCIC) system. (b) Top SEM image of the fabricated SCIC system. (c) Response of individual sensors and output voltage of the system during egg freshness monitoring. While the response of each sensor fluctuates daily without a distinct pattern, the output voltage shows a gradual increase over time.


[Contact Information]
Joon Hwang, Semiconductor Materials & Devices Laboratory, Department of Electrical and Computer Engineering, Seoul National University / joonh1007@snu.ac.kr