A5066896226- Advanced Memory and Neural Computing
- Ferroelectric and Negative Capacitance Devices
- Semiconductor materials and devices
- Neuroscience and Neural Engineering
- CCD and CMOS Imaging Sensors
- GaN-based semiconductor devices and materials
- Neural dynamics and brain function
- Transition Metal Oxide Nanomaterials
- Antenna Design and Analysis
- ZnO doping and properties
- Physical Unclonable Functions (PUFs) and Hardware Security
- Nanowire Synthesis and Applications
- Ga2O3 and related materials
- Electronic and Structural Properties of Oxides
- Industrial Vision Systems and Defect Detection
- Medical Image Segmentation Techniques
- Surface Roughness and Optical Measurements
- Energy Harvesting in Wireless Networks
- Advanced Vision and Imaging
- Optical measurement and interference techniques
- Image Processing Techniques and Applications
- Drug Solubulity and Delivery Systems
- Cancer, Hypoxia, and Metabolism
- Advanced Antenna and Metasurface Technologies
- Antenna Design and Optimization
Gyeongsang National University
2022-2025
Seoul National University of Science and Technology
2024-2025
Georgia Institute of Technology
2023-2024
Dong-Eui University
2024
Hanyang University
1999-2024
Seoul National University
2016-2023
Korea Polytechnic University
2021-2022
Inha University
2020-2021
Hongik University
2021
Korea Hydro and Nuclear Power (Korea)
2020
We propose a hardware-friendly architecture of convolutional neural network using 32 × memristor crossbar array having an overshoot suppression layer. The gradual switching characteristics in both set and reset operations enable the implementation 3-bit multilevel operation whole that can be utilized as 16 kernels. Moreover, binary activation function mapped to read voltage ground is introduced evaluate result training with boundary 0.5 its estimated gradient. Additionally, we adopt fixed...
To apply resistive random‐access memory (RRAM) to the neuromorphic system and improve performance, each cell in array should be able operate independently by reducing device variation. In addition, it is necessary lower operating current of RRAM enable gradual switching characteristics mimic low‐energy operations biological. most filamentary RRAMs, however, overshoot occurs forming stage, shows large variation, high current, abrupt set reset characteristics. Herein, shortcomings occurring...
Neuromorphic computing offers parallel data processing and low energy consumption can be useful to replace conventional von Neumann computing. Memristors are two-terminal devices with varying conductance that used as synaptic arrays in hardware-based neuromorphic devices. In this research, we extensively investigate the analog symmetric multi-level switching characteristics of zinc tin oxide (ZTO)-based memristor for systems. A ZTO semiconductor layer is introduced between a complementary...
Using "capacitive" crossbar arrays for compute-in-memory (CIM) offers higher energy efficiency compared to "resistive" arrays. The non-volatile capacitive (nvCap) synapse has been recently reported in MFM and MFS stacks with asymmetric electrode interfaces. In this work, a foundry FeFET is leveraged as the first time, where tunable gate-to-drain gate-to-source capacitances ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math...
Compute-in-memory (CIM) has emerged as a compelling approach to address the ever-increasing demand for energy-efficient computing edge artificial intelligence (AI) applications. Nonvolatile capacitive synapses have been recently proposed further boost energy and area efficiency of CIM hardware with following attractive features when compared resistive synapses: negligible static power consumption, selectorless access, low interconnect voltage drop exemption from sneak-path leakage, limited...
This article surveys the recent development of semiconductor memory technologies spanning from mainstream static random-access memory, dynamic and flash toward emerging candidates such as resistive, ferroelectric, magnetic memories. Pathways for future technological innovations are presented.
In this Letter, we present reset-voltage-dependent precise tuning operation of TiOx/Al2O3-based memristive devices. For the high resistance state (HRS) with reset voltage, abrupt set operations are observed a large variation, while HRS obtained by low voltage provides gradual and uniform switching behaviors. The improvement programming accuracy analyzed regarding cycle-to-cycle as well device-to-device variations. We believe that these results can be applied to operate memristors in areas...
Although many studies have been continuously conducted to reduce the power consumption of a resistive random access memory (RRAM) cross-point array with current-compliance effect, it has difficult yet realize intrinsic self-compliance effects in an RRAM device itself. In this study, simple oxygen-rich TiOy layer is inserted into Al2O3/TiOx-based stack as current suppression layer, and XPS analysis provided compare stoichiometry TiOx layers. A region formed between different breakdown...
Nanostructured semiconducting metal oxides such as SnO2, ZnO, TiO2, and CuO have been widely used to fabricate high performance gas sensors. To improve the sensitivity stability of sensors, we developed NO2 sensors composed ZnO/TiO2 core–shell nanorods (NRs) decorated with Au nanoparticles (NPs) synthesized via a simple low-temperature aqueous solution process, operated under ultraviolet irradiation realize room temperature operation. The fabricated sensor 10 nm-thick TiO2 shell layer shows...
Recently, various memory devices have been actively studied as suitable candidates for synaptic devices, which are important and computing units in neuromorphic systems. One of the ways to manage these is off-chip training, where it essential transfer pretrained weights accurately. Previous studies, however, a few limitations, such lack consideration program errors that occur during process. Although smaller error, higher accuracy, corresponding increase time must be considered. To evaluate...
Memristor devices can be utilized for various computing applications, and stochastic is one of them. The intrinsic characteristics the memristor cause unpredictable current fluctuations by capture emission electrons in a trap site. Herein, true random number generator (TRNG) using telegraph noise (RTN) as an entropy source proposed. TiO x /Al 2 O 3 memristors are fabricated, time probability RTN modulated to 50% with varying read‐voltage device conductance state. In addition, TRNG operations...
To develop proliposome formulations to improve the oral bioavailability of l-glutathione (GSH), GSH-loaded proliposomes were prepared using granule method. Mannitol was selected as an effective excipient achieve desired particle size, entrapment efficiency (EE), and solubility for delivery final formulation. evaluate effect surface charge on GSH, negative (F1-F4) positive (F5-F9) prepared. Particle size F1 F5 167.8 ± 0.9 175.9 2.0 nm, zeta potential -8.1 0.7 21.1 mV, respectively....
Abstract As interest in artificial intelligence (AI) and relevant hardware technologies has been developed rapidly, algorithms network structures have become significantly complicated, causing serious power consumption issues because an enormous amount of computation is required. Neuromorphic computing, a AI technology with memory devices, emerged to solve this problem. For application, multilevel operations synaptic devices are important imitate floating point weight values software...
To analyze the effect of intrinsic variations memristor device on neuromorphic system, we fabricated 32 × Al2O3/TiOx-based crossbar array and implemented 3 bit multilevel conductance as weight quantization by utilizing switching characteristics to minimize performance degradation neural network. The tuning operation for 8 levels was confirmed with a tolerance ±4μA (±40μS). endurance retention were also verified, random telegraph noise (RTN) measured according range evaluate internal...
Memristor‐based ternary content‐addressable memory (TCAM) has emerged as an alternative to conventional static random‐access (SRAM)‐based TCAM because of its high‐density integration and zero‐static energy consumption. Herein, 0T2R operation on a 32 × passive memristor crossbar circuit is experimentally verified. The effective margin, which the difference between match case 1‐bit mismatch case, improved through precise tuning operations. Moreover, number bits cases can be accurately detected...
In this work, we present a fabrication strategy for high-yield memristor crossbar arrays. Our approach uses an Al2O3/TiOx-based bilayer with combination of dielectric and oxygen reservoir layer. The process is optimized by controlling the thickness Al2O3 layer to decrease forming voltage, thus reducing possibility device failure due excessive current during process. We also investigate yield trends concentration TiOx layer, achieving over 98% under optimal conditions. then fabricate array...
In this work, a two-terminal TiO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><i>x</i></sub> -based memristor device has been fabricated and the methods for controlling its conductance are demonstrated. The exhibits bipolar analog resistive-switching characteristics margin over 10 fold between highest lowest resistance states (RS). It is revealed that can be adjusted with high resolution by either continuous voltage sweep mode or pulse mode....
Although vertical configurations for high-density storage require challenging process steps, such as etching high aspect ratios and atomic layer deposition (ALD), they are more affordable with a relatively simple lithography have been employed in many studies. Herein, the potential of memristors CMOS-compatible 3D stacked structures Pt/Ti/HfOx/TiN-NCs/HfOx/TiN is examined use neuromorphic systems. The electrical characteristics (including I-V properties, retention, endurance) were...