Article
Engineering, Electrical & Electronic
Chih-Ying Chen, Yu-Hsiu Feng, Hong-Lin Lu, Feng-En Chang, Jui-Yuan Chen
Summary: In this study, an integrated structure called one phase-change memory one resistive random access memory (1P1R) was proposed to suppress the sneak current during stacking. The 1P1R device remained in a high resistance state to suppress the sneak current, and switched to a working state to write/read its state. The feasibility of the 1P1R structure was confirmed through electrical measurement, and the property analysis provided insight into its speculated mechanism. The results demonstrated that the novel 1P1R structure could effectively suppress sneak current, showing potential for 3D IC manufacturing.
ACS APPLIED ELECTRONIC MATERIALS
(2023)
Article
Chemistry, Physical
Seung Woo Han, Moo Whan Shin
Summary: This study fabricates a high-performance flexible RRAM device using a precisely controlled UV laser annealing process, which changes the concentration of O Frenkel defect pairs in the ZnO layer and produces a ZnO/Al mixed interface layer with high quality oxygen reservoirs. The laser-annealed flexible RRAM shows stable resistive switching, performance enhancement, high on/off ratio, cycling endurance, and low power consumption, even at a bending radius of up to 5 mm.
JOURNAL OF ALLOYS AND COMPOUNDS
(2022)
Article
Chemistry, Physical
Seung Woo Han, Chul Jin Park, Moo Whan Shin
Summary: This study demonstrates that the diffusion of aluminum atoms and oxygen vacancies significantly affect the resistive switching behavior of zinc oxide-based random resistive access memory (RRAM). The diffusion of aluminum atoms into the zinc oxide layer acts as dopants, producing additional oxygen vacancies and contributing to the formation of conductive filaments. Additionally, the formation of an aluminum oxide layer by the redox reaction between aluminum atoms and oxygen leads to the instability of the reset process.
SURFACES AND INTERFACES
(2022)
Article
Engineering, Electrical & Electronic
Yu-Chung Lien, Tsung-Ta Wu, S. Simon Wong
Summary: Demonstrated is a resistive random-access memory (RRAM) utilizing nitrogen-doped aluminum oxide (AlOxNy) resistive dielectric deposited with atomic layer deposition (ALD), suitable for high-density 3-D vertical via array. The RRAM exhibits almost forming-free operation, sub-microamperes programming currents, and appropriate resistance ranges with ALD electrodes. As the via size scales down, the set voltage and reset current decrease, while the memory window widens due to the device's 3-D nature, showing programming endurance of 5 x 10(4) cycles and retention up to 10 years at 90°C.
IEEE TRANSACTIONS ON ELECTRON DEVICES
(2021)
Article
Materials Science, Multidisciplinary
Joong Hyeon Park, Sobia Ali Khan, Mehr Khalid Rahmani, Jihwan Cho, Moon Hee Kang
Summary: We fabricated organic RRAM devices using a low-cost solution-process method and enhanced their performance by simple UVO treatment. The devices exhibited bipolar resistive switching behavior with high ON/OFF ratio and endurance. The conduction mechanisms of the devices were investigated and Schottky emission and Ohmic conduction were identified as the main mechanisms.
MATERIALS RESEARCH EXPRESS
(2022)
Article
Computer Science, Information Systems
Usman Bature Isyaku, Mohd Haris Bin Md Khir, I. Md Nawi, M. A. Zakariya, Furqan Zahoor
Summary: Numerous works have shown the study and enhancement of switching properties of ZnO-based RRAM devices. Various native point defects affecting ZnO are discussed. Different methods like doping elements, multi-layered structures, suitable electrodes, and hybrid structures are explored for enhancing the switching dynamics.
Article
Engineering, Electrical & Electronic
Dongsheng Cui, Yawei Du, Zhenhua Lin, Mengyang Kang, Yifei Wang, Jie Su, Jincheng Zhang, Yue Hao, Jingjing Chang
Summary: A memory device with an Ag/Ga2O3/Pt structure has been successfully fabricated, exhibiting both bipolar resistive switching (BRS) and unipolar resistive switching (URS) behaviors. It was found that the bipolar and unipolar modes can be set by applying a positive voltage with the same compliance current (I-cc) of 1 mA. The reset process involves a polarity change of sweeping voltages without I-cc to switch between the bipolar and unipolar modes. The conduction mechanisms are identified as conducting filaments (CFs) for the low resistance state (LRS), and schottky emission for BRS, and space charge limited conduction mechanism for URS in the high resistance states (HRS), respectively.
IEEE ELECTRON DEVICE LETTERS
(2023)
Article
Nanoscience & Nanotechnology
Jeong Hyun Yoon, Young-Woong Song, Wooho Ham, Jeong-Min Park, Jang-Yeon Kwon
Summary: With the era of big data, the traditional von Neumann architecture is insufficient due to high latency and energy consumption. Neuromorphic computing, imitating biological neurons for parallel processing, is a promising solution. Resistive random access memory (RRAM) with fast-switching speed and scalability is a potential candidate. However, devices excelling in all aspects are rarely proposed.
Article
Chemistry, Multidisciplinary
Dhananjay Mishra, Krishnaiah Mokurala, Ajit Kumar, Seung Gi Seo, Hyeon Bin Jo, Sung Hun Jin
Summary: This study demonstrates the effectiveness of a new p-type copper iodide semiconductor in a flexible, low-voltage resistive random-access memory. The CuI RRAM devices, implemented through a room-temperature solid iodination process, show consistent On/Off ratio, excellent endurance, and long retention period. The study also showcases the use of blue light illumination for multi-level data storage and explores the thermal stability and key switching mechanism in CuI RRAM devices. Furthermore, the longevity of CuI devices is improved through PMMA encapsulation.
ADVANCED FUNCTIONAL MATERIALS
(2023)
Article
Chemistry, Multidisciplinary
Ke-Jing Lee, Wei-Shao Lin, Li-Wen Wang, Hsin-Ni Lin, Yeong-Her Wang
Summary: This paper presents a sol-gel process-prepared SrZrTiO3 (SZT) thin film for the insulator of resistive random-access memory (RRAM). Aluminum (Al) was embedded in the SZT thin film to enhance the switching characteristics. The RRAM with embedded Al in SZT thin film showed significant improvements in device parameters compared to pure SZT thin-film RRAM.
Article
Physics, Applied
Shih-Kai Lin, Min-Chen Chen, Ting-Chang Chang, Chen-Hsin Lien, Cheng-Hsien Wu, Yu-Shuo Lin, Pei-Yu Wu, Yung-Fang Tan, Wei-Chen Huang, Yong-Ci Zhang, Sheng-Yao Chou, Chung-Wei Wu, Simon M. Sze
Summary: This study investigates the influence of a supercritical fluid treatment on the characteristics of resistive random access memory. The treatment improves memory characteristics and induces oxygen ion doping and defect generation in the switching region.
APPLIED PHYSICS EXPRESS
(2022)
Article
Materials Science, Multidisciplinary
P. R. Sekhar Reddy, Venkata Raveendra Nallagatla, Yedluri Anil Kumar, G. Murali
Summary: This study investigates the resistive switching behavior of HfAlOx/ZrO2 thin films. The Au/HfAlOx/ZrO2/TiN stack shows superior performance compared to the Au/HfAlOx/TiN memory stack. Both stacks demonstrate good retention and endurance characteristics.
PROGRESS IN NATURAL SCIENCE-MATERIALS INTERNATIONAL
(2022)
Article
Engineering, Electrical & Electronic
Ruofei Hu, Jianshi Tang, Yue Xi, Zhixing Jiang, Yuyao Lu, Bin Gao, He Qian, Huaqiang Wu
Summary: A nitrogen-oxyanion-doped hafnium oxide RRAM with improved forming voltage, on/off ratio, and endurance is demonstrated. The critical electric field of N-doped RRAM for forming is 40% lower than that of undoped RRAM. The N-doped RRAM achieves 3x improvement in on/off ratio and 10x improvement in endurance at the forming voltage of 2 V, which is suitable for integration with advanced silicon technology nodes.
IEEE ELECTRON DEVICE LETTERS
(2023)
Article
Engineering, Electrical & Electronic
Xueyong Zhang, Byung-Kwon An, Tony Tae-Hyoung Kim
Summary: This paper proposes a time-based sensing scheme for read operations in SLC and MLC RRAM arrays, improving read reliability and reducing bit error rate (BER) compared to conventional schemes. The proposed scheme operates at 0.7-1.2 V supply and consumes 49 fJ/bit for read operation under a nominal 1.2 V supply. (48 words)
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I-REGULAR PAPERS
(2022)
Article
Chemistry, Physical
Hee Won Suh, Dong Su Kim, Ji Hoon Choi, Hak Hyeon Lee, Kun Woong Lee, Sung Hyeon Jung, Won Seok Yang, Jeong Jae Kim, Ji Sook Yang, Ho Seong Lee, Hyung Koun Cho
Summary: In this study, a unique crossbar array structure was proposed, which can be independently driven through the novel design of a multilayer RRAM. The ANPs-Cu2O active layers prepared by electrodeposition showed different set voltages and uniform reset voltages. The VCu-controlled DALs effectively prevented reverse current and achieved selector-less RRAM. The experimental results demonstrated that the array exhibited excellent memory performance and endurance without interfering with adjacent cells.
APPLIED SURFACE SCIENCE
(2023)
Article
Engineering, Electrical & Electronic
Zhong Sun, Daniele Ielmini
Summary: Matrix computation is widely used in scientific and engineering fields, but it can be computationally complex in conventional digital computers. Analog matrix computing circuits based on resistive memory arrays provide a promising solution for fast and efficient matrix computations. This tutorial introduces the design principles, mapping strategies, stability requirements, and applications of analog matrix computing circuits.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II-EXPRESS BRIEFS
(2022)
Article
Neurosciences
Andrea Baroni, Artem Glukhov, Eduardo Perez, Christian Wenger, Enrico Calore, Sebastiano Fabio Schifano, Piero Olivo, Daniele Ielmini, Cristian Zambelli
Summary: This study proposes an IMC architecture based on RRAM crossbar arrays for accelerated neural network applications in survival analysis. The trade-off between performance and energy consumption is explored through synaptic weights mapping strategy and programming algorithms.
FRONTIERS IN NEUROSCIENCE
(2022)
Article
Chemistry, Multidisciplinary
Matteo Farronato, Piergiulio Mannocci, Margherita Melegari, Saverio Ricci, Christian Monzio Compagnoni, Daniele Ielmini
Summary: This study presents a charge-trap memory device with a MoS2 channel, which shows outstanding linearity of potentiation and is suitable for high-density neuromorphic computing.
ADVANCED MATERIALS
(2023)
Article
Engineering, Electrical & Electronic
Andrea Baroni, Artem Glukhov, Eduardo Perez, Christian Wenger, Daniele Ielmini, Piero Olivo, Cristian Zambelli
Summary: This paper analyzes the benefits of a new programming algorithm in RRAM arrays, which achieves better conductance control and lower variability through Set and Reset operations. The superior performance stability of this algorithm is demonstrated through data retention analysis and artificial neural network simulation.
IEEE TRANSACTIONS ON DEVICE AND MATERIALS RELIABILITY
(2022)
Article
Engineering, Electrical & Electronic
Marco Bertuletti, Irene Mu Noz-Martin, Stefano Bianchi, Andrea G. Bonfanti, Daniele Ielmini
Summary: In this work, a novel approach to the conceptual and technical design of integrated neural networks is proposed using novel in-memory computing circuits based on emerging nonvolatile memories. To reduce power consumption and complexity, a fully analog computing approach is introduced, replacing the analog-to-digital converter with a simple comparator. The impact of major nonidealities, such as PCM conductance variability, conductance drift, IR drop, and readout threshold, on accuracy is studied.
IEEE TRANSACTIONS ON ELECTRON DEVICES
(2023)
Article
Engineering, Electrical & Electronic
Piergiulio Mannocci, Enrico Melacarne, Daniele Ielmini
Summary: In-memory computing has become a promising candidate for distributed computing frameworks due to its energy efficiency and high throughput. This study proposes a closed-loop in-memory computing circuit for accelerating Ridge Regression and demonstrates its capabilities in realistic scenarios, comparing it with a commercial GPU. The results show significant improvements in energy efficiency and area efficiency, supporting the use of in-memory computing in future B5G and 6G networks.
IEEE JOURNAL ON EMERGING AND SELECTED TOPICS IN CIRCUITS AND SYSTEMS
(2022)
Article
Multidisciplinary Sciences
Rohit Abraham, Maksym V. Kovalenko, Daniele Ielmini, Alessandro Milozzi, Sergey Tsarev, Rolf Bronnimann, Simon C. Boehme, Erfu Wu, Ivan Shorubalko
Summary: This study demonstrates second-order dynamics in halide perovskite memristive diodes (memdiodes), capturing both timing-and rate-based plasticity. By exploiting ion migration, back diffusion, and modulable Schottky barriers, general design rules are established for higher-order memristors, enabling complex binocular orientation selectivity without the need for complicated circuitry.
Article
Multidisciplinary Sciences
S. Bianchi, I. Munoz-Martin, E. Covi, A. Bricalli, G. Piccolboni, A. Regev, G. Molas, J. F. Nodin, F. Andrieu, D. Ielmini
Summary: Authors propose a bio-inspired recurrent neural network based on resistive-switching synaptic arrays for autonomous exploration. The network utilizes homeostatic Hebbian learning for improved efficiency in reinforcement learning tasks. Experimental and theoretical discussions are presented to benchmark the accuracy and resilience of the proposed architecture.
NATURE COMMUNICATIONS
(2023)
Article
Nanoscience & Nanotechnology
Paolo Fantini, Nicola Polino, Andrea Ghetti, Daniele Ielmini
Summary: An ovonic threshold switch (OTS) based on chalcogenide glasses is used in storage class memory (SCM) arrays as a selecting device. The switching phenomenon of OTS, which involves electronic transport, joule heating, and phase transition, has attracted significant interest. This study reveals that the current-voltage characteristic near the switching point reflects carrier multiplication. The physical mechanism of threshold switching is explained by bipolar impact ionization leading to avalanche multiplication, resulting in the typical S-shaped characteristic. Numerical simulations based on this physics-based model accurately predict the switching properties of OTS at different chalcogenide thicknesses and compositions. These findings provide a theoretical framework for future design and optimization of OTS in memory and computing applications.
ADVANCED ELECTRONIC MATERIALS
(2023)
Article
Computer Science, Hardware & Architecture
Piergiulio Mannocci, Daniele Ielmini
Summary: Matrix-based computing is widely used in machine learning applications, but traditional von-Neumann architecture has limitations in energy and latency. In-memory computing (IMC) performs computation directly within the memory, eliminating the need for data transfer. This article presents a generalized closed-loop IMVM circuit that can perform any linear matrix operation by memory remapping, making it an ideal candidate for general-purpose machine learning accelerators.
IEEE JOURNAL ON EXPLORATORY SOLID-STATE COMPUTATIONAL DEVICES AND CIRCUITS
(2023)
Article
Engineering, Electrical & Electronic
Tommaso Rizzi, Andrea Baroni, Artem Glukhov, Davide Bertozzi, Christian Wenger, Daniele Ielmini, Cristian Zambelli
Summary: Resistive Random Access Memory (RRAM) technology has the potential to improve FPGA performance, reduce footprint, and lower energy requirements compared to CMOS-based products. However, high programming power consumption and non-ideal behaviors of the RRAM device hinder its integration in FPGAs. This work explores the impact of different programming procedures on run-time performance and highlights the importance of target resistive state selection and programming algorithm in reducing delay and energy metrics and improving robustness against variations.
IEEE TRANSACTIONS ON DEVICE AND MATERIALS RELIABILITY
(2023)
Article
Engineering, Electrical & Electronic
Manuel Escudero, Sabina Spiga, Mauro Di Marco, Mauro Forti, Giacomo Innocenti, Alberto Tesi, Fernando Corinto, Stefano Brivio
Summary: This paper presents a framework for the physical implementation of a tunable memristor Chua's circuit and successfully generates different oscillation patterns by programming a nonvolatile memristive device to different states. The characterization and modeling of the real device were used to extend the experimental work and draw further requirements for successful circuit implementation.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I-REGULAR PAPERS
(2023)
Review
Engineering, Electrical & Electronic
S. Brivio, S. Spiga, D. Ielmini
Summary: HfO2-based RRAM shows great potential in neuromorphic computing, with significant achievements in artificial neural networks and synapses. Its switching mechanisms, programming algorithms, and neuromorphic applications are thoroughly discussed and illustrated.
NEUROMORPHIC COMPUTING AND ENGINEERING
(2022)
Proceedings Paper
Engineering, Electrical & Electronic
Matteo Baldo, Lorenzo Turconi, Alessandro Motta, Elisa Petroni, Luca Laurin, Daniele Ielmini, Andrea Redaelli
Summary: This study presents a detailed investigation of the reset state of embedded phase change memories (ePCM) based on Ge-rich Ge2Sb2Te5 (Ge-GST) alloys. The resistance drift and conduction mechanisms are assessed under different bake times and temperatures. The evolution of the activation energy for conduction in the Ge-GST compound is described for the first time. The dependence of activation energy and resistance value on the forming state is also evaluated.
ESSDERC 2022 - IEEE 52ND EUROPEAN SOLID-STATE DEVICE RESEARCH CONFERENCE (ESSDERC)
(2022)
Article
Engineering, Electrical & Electronic
A. Redaelli, A. Gandolfo, G. Samanni, E. Gomiero, E. Petroni, L. Scotti, A. Lippiello, P. Mattavelli, J. Jasse, D. Codegoni, A. Serafini, R. Ranica, C. Boccaccio, J. Sandrini, R. Berthelon, J-C Grenier, O. Weber, D. Turgis, A. Valery, S. Del Medico, V Caubet, J-P Reynard, D. Dutartre, L. Favennec, A. Conte, F. Disegni, M. De Tomasi, A. Ventre, M. Baldo, D. Ielmini, A. Maurelli, P. Ferreira, F. Arnaud, F. Piazza, P. Cappelletti, R. Annunziata, R. Gonella
Summary: The effect of back-end of line (BEOL) process on the performance and reliability of Phase-Change Memory embedded in a 28nm FD-SOI platform (ePCM) is discussed in this paper. The study focuses on the microscopic evolution of the Ge-rich GST alloy during the process. It is found that Ge clustering occurs during fabrication, affecting the resistance and cell performance. An optimized process is identified to meet the requirements for demanding automotive applications.
IEEE JOURNAL OF THE ELECTRON DEVICES SOCIETY
(2022)
