DGIST Develops Highly Efficient Process Technology for Next-Generation AI Semiconductors!

DGIST Professors Jang Jae-eun and Kwon Hyuk-jun and the research team develop ferroelectric field-effect transistors suitable for heterojunction following a low-temperature annealing process using laser - The development of ferroelectric-based memory devices enables next-generation AI semiconductor systems - The results are published in Advanced Science, a top international journal in the field

□ DGIST’s (President Kunwoo Lee) Electrical Engineering and Computer Science Professor Jang Jae-eun (Ph.D. candidate Kim Dongsu) and Professor Kwon Hyuk-jun (Ph.D. candidate Jeong Heejae) and their research team have developed a high-efficiency process technology for next-generation AI memory transistors. The team developed a nanosecond pulsed laser-based “selective heat treatment method” and “thermal energy minimization control process technology” to overcome the shortcomings of the high-temperature process of ferroelectric field-effect transistors, which have non-volatile memory characteristics, high-speed operation, low power consumption, long lifetime, and durability. The new technology process enables the realization of heterojunction structures, which are the core technology of next-generation AI semiconductors.


□ As various electronic systems are developed in the Fourth Industrial Revolution, high-level information processing and storage technologies are in demand. Most of all, AI-type memory transistor technology, which performs computation and memory simultaneously, like the human brain, is an exceptional technology with advantages such as reduced manufacturing costs, power efficiency, high performance, lightweight, and improved integration, and in-depth research is being conducted in this field.


□ Recent research on AI-type memory transistors has been centered on ferroelectric field-effect transistors[1], which have non-volatile memory characteristics, high-speed operation, low power consumption, long lifetime, and durability. However, due to the characteristics of ferroelectrics[2] that require a “high-temperature phase formation process[3]” (over 600 degrees Celsius), there is the issue that the characteristics of the transistors are degraded or thermal damage is caused to the lower layers of logic circuits or input/output (I/O) circuits when the heterojunction of memory devices is performed.


□ To overcome the shortcomings of ferroelectric processes, a joint research team led by DGIST Professors Jang Jae-eun and Kwon Hyuk-jun conducted research to develop heat treatment methods for specific parts and control process technology to minimize thermal energy. The team introduced a “nanosecond pulsed laser annealing process” that enables selective heat treatment to solve the problem of high-temperature processes. The process has a short wavelength (355 nm), a penetration depth as thin as two-thousandths of a human hair, and fast cooling characteristics utilizing a pulsed laser in 30 millionths of a second, which is highly advantageous for controlling thermal energy and promoting ferroelectricity[4] in ferroelectrics.


□ Using this technology, the team developed memory transistors with ferroelectrics and semiconductor channels[5] that are active even at low temperatures and that have highly efficient AI semiconductor characteristics. The optimized device has an ultra-fast response time of a millionth of a second, performing over 100,000 stable write and erase operations in an accelerated measurement environment and maintaining the memory state for over 10 years. In addition, the transistor has a “memory operating range” of over 1.7V, an on-off current ratio over 100,000, showing excellent power consumption performance and high ferroelectric properties (2-Pr of 14.7 µC/cm2) and pattern recognition linearity, as well as demonstrating outstanding mechanical, electrical, and chemical performance.


□ “The thermal energy minimization laser annealing technology developed in this study is a new approach to enhance the commercialization potential of ferroelectric field-effect transistors for AI systems, and it is expected to drive innovation in three-dimensional integration technology by overcoming the limitations of conventional high-temperature processes,” said Professor Jang Jae-eun. “In this study, the devices optimized through laser annealing control have realized memory applications in heterojunction structures and flexible structures and demonstrated high compatibility with existing silicon processes. We expect to innovatively contribute to the realization of next-generation AI systems that include various devices in the future,” said Professor Kwon Hyuk-jun.


□ The research was supported by the Korea Research Foundation's Mid-Career Research Program and Samsung Electronics’ Industrial Project, and the results are published online in Advanced Science (IF=17.521), a top international journal in the field.

- Corresponding Author E-mail Address : [email protected].k

[1]  Ferroelectric Field-Effect Transistor (FeFET): A transistor that uses ferroelectric materials as gate insulators to perform data storage and logic operations.

[2] Ferroelectric: A material that is electrically polarized by an external electric field and remains polarized after the removal of the external electric field.

[3] High-Temperature Phase Formation Process: A method used primarily in the manufacturing and processing of semiconductor and ceramic materials that involves the formation or transformation of a specific phase of a material at high temperatures.

[4] Ferroelectricity: Refers to the property that certain materials have spontaneous electrical polarization in the absence of an external electric field, whose polarization direction can be reversed when an external electric field is applied.

[5] Semiconductor channel: A semiconductor material that forms the path through which current flows in a transistor, determining the mobility of electrons and thus the performance of the device.

Published: 17 Jun 2024

Contact details:


333, Techno jungang-daero, Hyeonpung-myeon, Dalseong-gun, Daegu, 42988

News topics: 
Academic discipline: