Tae Won Park
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Tae Won Park has a total of 25 co-authors. Here are the top co-authors by number of shared publications:
- Cheol Seong Hwang
Seoul National University
4 shared publications - Joon-Kyu Han
Seoul National University
3 shared publications - Dong Hoon Shin
Department of Materials Science and Engineering and Inter‐University Semiconductor Research Center Seoul National University Seoul 08826 Republic of Korea
2 shared publications - Kyung Seok Woo
Ulsan National Institute of Science and Technology
2 shared publications - Soo Hyung Lee
Department of Materials Science and Engineering and Inter‐University Semiconductor Research Center College of Engineering Seoul National University Seoul 08826 Republic of Korea
2 shared publications Dong Yun Kim
Department of Materials Science and Engineering and Inter‐University Semiconductor Research Center College of Engineering Seoul National University Seoul 08826 Republic of Korea
1 shared publication
Latest Publications
- 12 January 2025
- Published by Wiley in Advanced Materials
- Vol. 37 (8) , e2412549
An Analysis of Components and Enhancement Strategies for Advancing Memristive Neural Networks
Advancements in artificial intelligence (AI) and big data have highlighted the limitations of traditional von Neumann architectures, such as excessive power consumption and limited performance improvement with increasing parameter numbers. These challenges are significant for edge devices requiring higher energy and area efficiency. Recently, many reports on memristor‐based neural networks (Mem‐NN) using resistive switching memory have shown efficient computing performance with a low power requirement. Even further performance optimization can be made using engineering resistive switching mechanisms. Nevertheless, systematic reviews that address the circuit‐to‐material aspects of Mem‐NNs, including their dedicated algorithms, remain limited. This review first categorizes the memristor‐based neural networks into three components: pre‐processing units, processing units, and learning algorithms. Then, the optimization methods to improve integration and operational reliability are discussed across materials, devices, circuits, and algorithms for each component. Furthermore, the review compares recent advancements in chip‐level neuromorphic hardware with conventional systems, including graphic processing units. The ongoing challenges and future directions in the field are discussed, highlighting the research to enhance the functionality and reliability of Mem‐NNs.
- 12 January 2025
- Published by Wiley in Advanced Materials
- Vol. 37 (8) , e2412549
An Analysis of Components and Enhancement Strategies for Advancing Memristive Neural Networks
Advancements in artificial intelligence (AI) and big data have highlighted the limitations of traditional von Neumann architectures, such as excessive power consumption and limited performance improvement with increasing parameter numbers. These challenges are significant for edge devices requiring higher energy and area efficiency. Recently, many reports on memristor‐based neural networks (Mem‐NN) using resistive switching memory have shown efficient computing performance with a low power requirement. Even further performance optimization can be made using engineering resistive switching mechanisms. Nevertheless, systematic reviews that address the circuit‐to‐material aspects of Mem‐NNs, including their dedicated algorithms, remain limited. This review first categorizes the memristor‐based neural networks into three components: pre‐processing units, processing units, and learning algorithms. Then, the optimization methods to improve integration and operational reliability are discussed across materials, devices, circuits, and algorithms for each component. Furthermore, the review compares recent advancements in chip‐level neuromorphic hardware with conventional systems, including graphic processing units. The ongoing challenges and future directions in the field are discussed, highlighting the research to enhance the functionality and reliability of Mem‐NNs.
- 28 August 2024
- Published by Wiley in Advanced Materials
- Vol. 36 (40) , e2410191
In‐Materia Annealing and Combinatorial Optimization Based on Vertical Memristive Array
Due to its area and energy efficiency, a memristive crossbar array (CBA) has been extensively studied for various combinatorial optimization applications, from network problems to circuit design. However, conventional approaches include heavily burdening software fine‐tuning for the annealing process. Instead, this study introduces the “in‐materia annealing” method, where the inter‐layer interference of vertically stacked memristive CBA is utilized as an annealing method. When mapping combinatorial optimization problems into the configuration layer of the CBA, exponentially decaying annealing profiles are generated in nearby noise layers. Moreover, in‐materia annealing profiles can be controlled by changing compliance current, read voltage, and read pulse width. Therefore, the annealing profiles can be arbitrarily controlled and generated individually for each cell, providing rich noise sources to solve the problem efficiently. Consequently, the experimental and simulation of Max‐Cut and weighted Max‐Cut problems achieve notable results with the minimum software burden.
- 28 August 2024
- Published by Wiley in Advanced Materials
- Vol. 36 (40) , e2410191
In‐Materia Annealing and Combinatorial Optimization Based on Vertical Memristive Array
Due to its area and energy efficiency, a memristive crossbar array (CBA) has been extensively studied for various combinatorial optimization applications, from network problems to circuit design. However, conventional approaches include heavily burdening software fine‐tuning for the annealing process. Instead, this study introduces the “in‐materia annealing” method, where the inter‐layer interference of vertically stacked memristive CBA is utilized as an annealing method. When mapping combinatorial optimization problems into the configuration layer of the CBA, exponentially decaying annealing profiles are generated in nearby noise layers. Moreover, in‐materia annealing profiles can be controlled by changing compliance current, read voltage, and read pulse width. Therefore, the annealing profiles can be arbitrarily controlled and generated individually for each cell, providing rich noise sources to solve the problem efficiently. Consequently, the experimental and simulation of Max‐Cut and weighted Max‐Cut problems achieve notable results with the minimum software burden.
- Open Access
- 9 July 2024
- Published by AIP Publishing in Journal of Applied Physics
- Vol. 136 (2)
Comparative study on the stability of ferroelectric polarization of HfZrO2 and AlScN thin films over the depolarization effect
This study investigates the stability of the (Hf,Zr)O2 (HZO) and AlScN thin films against the depolarization effect. In order to adversely interfere with the polarization screening of ferroelectric (FE) film and induce the depolarization effect, an Al2O3 film was inserted between the TiN bottom electrode and the FE film. Subsequently, to minimize charge injection through Al2O3 films, the double remanent polarization–voltage (2Pr–V) curve was measured employing voltage pulses with 1 μs-length. As the thickness of the Al2O3 film increased, the charge injection through the Al2O3 film decreased, leading to a decrease in the slope of the 2Pr–V curve. The magnitude of the depolarization field applied to the FE film was determined by simulating the 2Pr–V curve using an equivalent circuit model. Subsequently, the proportion of reversed domains during the delay period due to the depolarization effect was measured. It was revealed that HZO films exhibited vulnerability to the depolarization effect, with 16% of domains being reversed. On the other hand, AlScN films demonstrated superior stability, with polarization loss of less than 1%. Ultimately, it was revealed that AlScN films demonstrated superior stability against the depolarization effect compared to HZO films due to their higher coercive field and more uniform polarization distribution across the film area.
- 9 July 2024
- Published by AIP Publishing in Journal of Applied Physics
- Vol. 136 (2)
Open AccessComparative study on the stability of ferroelectric polarization of HfZrO2 and AlScN thin films over the depolarization effect
This study investigates the stability of the (Hf,Zr)O2 (HZO) and AlScN thin films against the depolarization effect. In order to adversely interfere with the polarization screening of ferroelectric (FE) film and induce the depolarization effect, an Al2O3 film was inserted between the TiN bottom electrode and the FE film. Subsequently, to minimize charge injection through Al2O3 films, the double remanent polarization–voltage (2Pr–V) curve was measured employing voltage pulses with 1 μs-length. As the thickness of the Al2O3 film increased, the charge injection through the Al2O3 film decreased, leading to a decrease in the slope of the 2Pr–V curve. The magnitude of the depolarization field applied to the FE film was determined by simulating the 2Pr–V curve using an equivalent circuit model. Subsequently, the proportion of reversed domains during the delay period due to the depolarization effect was measured. It was revealed that HZO films exhibited vulnerability to the depolarization effect, with 16% of domains being reversed. On the other hand, AlScN films demonstrated superior stability, with polarization loss of less than 1%. Ultimately, it was revealed that AlScN films demonstrated superior stability against the depolarization effect compared to HZO films due to their higher coercive field and more uniform polarization distribution across the film area.
- 8 April 2024
- Published by AIP Publishing in Applied Physics Letters
- Vol. 124 (15)
A ternary gate-connected threshold switching thin-film transistor
Multi-valued logic has been a significant focus of research in various fields with the advancement of information technology. One approach to realizing ternary logic is integrating of a threshold switching (TS) device with a transistor, but this method often entails a complex fabrication process. This work suggests a ternary gate-connected threshold switching thin-film transistor (TS-TFT) by serially connecting the TS device with a bottom-gate thin-film transistor (TFT). The fabrication process is simplified with a structure that shares electrodes and insulators. Different threshold voltages from TS and TFT devices produce stable multiple states. The Pt/HfO2/TiN TS device has an electronic trapping/detrapping switching mechanism that exhibits low power consumption and high reliability. With the superior electrical performance of an amorphous indium gallium zinc oxide TFT, the TS-TFT has stable endurance. Furthermore, pulse switching and ternary inverter are demonstrated from the practical point of view.
- 8 April 2024
- Published by AIP Publishing in Applied Physics Letters
- Vol. 124 (15)
A ternary gate-connected threshold switching thin-film transistor
Multi-valued logic has been a significant focus of research in various fields with the advancement of information technology. One approach to realizing ternary logic is integrating of a threshold switching (TS) device with a transistor, but this method often entails a complex fabrication process. This work suggests a ternary gate-connected threshold switching thin-film transistor (TS-TFT) by serially connecting the TS device with a bottom-gate thin-film transistor (TFT). The fabrication process is simplified with a structure that shares electrodes and insulators. Different threshold voltages from TS and TFT devices produce stable multiple states. The Pt/HfO2/TiN TS device has an electronic trapping/detrapping switching mechanism that exhibits low power consumption and high reliability. With the superior electrical performance of an amorphous indium gallium zinc oxide TFT, the TS-TFT has stable endurance. Furthermore, pulse switching and ternary inverter are demonstrated from the practical point of view.