Van der Waals ferroelectric transistors: the all-round artificial synapses for high-precision neuromorphic computing
ABSTRACT: State number, operation power, dynamic range and conductance weight update linearity are key synaptic device performance metrics for high-accuracy and low-power-consumption neuromorphic computing in hardware. However, high linearity and low power consumption couldn't be simultaneously...
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2023-06-01
|
| Series: | Chip |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2709472323000072 |
| _version_ | 1797347070929010688 |
|---|---|
| author | Zhongwang Wang Xuefan Zhou Xiaochi Liu Aocheng Qiu Caifang Gao Yahua Yuan Yumei Jing Dou Zhang Wenwu Li Hang Luo Junhao Chu Jian Sun |
| author_facet | Zhongwang Wang Xuefan Zhou Xiaochi Liu Aocheng Qiu Caifang Gao Yahua Yuan Yumei Jing Dou Zhang Wenwu Li Hang Luo Junhao Chu Jian Sun |
| author_sort | Zhongwang Wang |
| collection | DOAJ |
| description | ABSTRACT: State number, operation power, dynamic range and conductance weight update linearity are key synaptic device performance metrics for high-accuracy and low-power-consumption neuromorphic computing in hardware. However, high linearity and low power consumption couldn't be simultaneously achieved by most of the reported synaptic devices, which limits the performance of the hardware. This work demonstrates van der Waals (vdW) stacked ferroelectric field-effect transistors (FeFET) with single-crystalline ferroelectric nanoflakes. Ferroelectrics are of fine vdW interface and partial polarization switching of multi-domains under electric field pulses, which makes the FeFETs exhibit multi-state memory characteristics and excellent synaptic plasticity. They also exhibit a desired linear conductance weight update with 128 conductance states, a sufficiently high dynamic range of Gmax/Gmin > 120, and a low power consumption of 10 fJ/spike using identical pulses. Based on such an all-round device, a two-layer artificial neural network was built to conduct Modified National Institute of Standards and Technology (MNIST) digital numbers and electrocardiogram (ECG) pattern-recognition simulations, with the high accuracies reaching 97.6% and 92.4%, respectively. The remarkable performance demonstrates that vdW-FeFET is of obvious advantages in high-precision neuromorphic computing applications. |
| first_indexed | 2024-03-08T11:42:24Z |
| format | Article |
| id | doaj.art-5cc46fca905c46fb9e276188b2d7f408 |
| institution | Directory Open Access Journal |
| issn | 2709-4723 |
| language | English |
| last_indexed | 2024-03-08T11:42:24Z |
| publishDate | 2023-06-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Chip |
| spelling | doaj.art-5cc46fca905c46fb9e276188b2d7f4082024-01-25T05:24:02ZengElsevierChip2709-47232023-06-0122100044Van der Waals ferroelectric transistors: the all-round artificial synapses for high-precision neuromorphic computingZhongwang Wang0Xuefan Zhou1Xiaochi Liu2Aocheng Qiu3Caifang Gao4Yahua Yuan5Yumei Jing6Dou Zhang7Wenwu Li8Hang Luo9Junhao Chu10Jian Sun11School of Physics and Electronics, Central South University, Changsha 410083, ChinaState Key Laboratory of Powder Metallurgy, Central South University, Changsha 410083, ChinaSchool of Physics and Electronics, Central South University, Changsha 410083, ChinaSchool of Physics and Electronics, Central South University, Changsha 410083, ChinaShanghai Frontiers Science Research Base of Intelligent Optoelectronics and Perception, Institute of Optoelectronics, Department of Materials Science, Fudan University, Shanghai 200433, China; National Key Laboratory of Integrated Circuit Chips and Systems, Zhang Jiang Fudan International Innovation Center, Fudan University, Shanghai 200433, ChinaSchool of Physics and Electronics, Central South University, Changsha 410083, ChinaSchool of Physics and Electronics, Central South University, Changsha 410083, ChinaState Key Laboratory of Powder Metallurgy, Central South University, Changsha 410083, ChinaShanghai Frontiers Science Research Base of Intelligent Optoelectronics and Perception, Institute of Optoelectronics, Department of Materials Science, Fudan University, Shanghai 200433, China; National Key Laboratory of Integrated Circuit Chips and Systems, Zhang Jiang Fudan International Innovation Center, Fudan University, Shanghai 200433, ChinaState Key Laboratory of Powder Metallurgy, Central South University, Changsha 410083, ChinaShanghai Frontiers Science Research Base of Intelligent Optoelectronics and Perception, Institute of Optoelectronics, Department of Materials Science, Fudan University, Shanghai 200433, China; National Key Laboratory of Integrated Circuit Chips and Systems, Zhang Jiang Fudan International Innovation Center, Fudan University, Shanghai 200433, ChinaSchool of Physics and Electronics, Central South University, Changsha 410083, China; Corresponding author.ABSTRACT: State number, operation power, dynamic range and conductance weight update linearity are key synaptic device performance metrics for high-accuracy and low-power-consumption neuromorphic computing in hardware. However, high linearity and low power consumption couldn't be simultaneously achieved by most of the reported synaptic devices, which limits the performance of the hardware. This work demonstrates van der Waals (vdW) stacked ferroelectric field-effect transistors (FeFET) with single-crystalline ferroelectric nanoflakes. Ferroelectrics are of fine vdW interface and partial polarization switching of multi-domains under electric field pulses, which makes the FeFETs exhibit multi-state memory characteristics and excellent synaptic plasticity. They also exhibit a desired linear conductance weight update with 128 conductance states, a sufficiently high dynamic range of Gmax/Gmin > 120, and a low power consumption of 10 fJ/spike using identical pulses. Based on such an all-round device, a two-layer artificial neural network was built to conduct Modified National Institute of Standards and Technology (MNIST) digital numbers and electrocardiogram (ECG) pattern-recognition simulations, with the high accuracies reaching 97.6% and 92.4%, respectively. The remarkable performance demonstrates that vdW-FeFET is of obvious advantages in high-precision neuromorphic computing applications.http://www.sciencedirect.com/science/article/pii/S2709472323000072Ferroelectric transistorsFerroelectricvan der Waals heterostructuresArtificial synapsesNeuromorphic computing |
| spellingShingle | Zhongwang Wang Xuefan Zhou Xiaochi Liu Aocheng Qiu Caifang Gao Yahua Yuan Yumei Jing Dou Zhang Wenwu Li Hang Luo Junhao Chu Jian Sun Van der Waals ferroelectric transistors: the all-round artificial synapses for high-precision neuromorphic computing Chip Ferroelectric transistors Ferroelectric van der Waals heterostructures Artificial synapses Neuromorphic computing |
| title | Van der Waals ferroelectric transistors: the all-round artificial synapses for high-precision neuromorphic computing |
| title_full | Van der Waals ferroelectric transistors: the all-round artificial synapses for high-precision neuromorphic computing |
| title_fullStr | Van der Waals ferroelectric transistors: the all-round artificial synapses for high-precision neuromorphic computing |
| title_full_unstemmed | Van der Waals ferroelectric transistors: the all-round artificial synapses for high-precision neuromorphic computing |
| title_short | Van der Waals ferroelectric transistors: the all-round artificial synapses for high-precision neuromorphic computing |
| title_sort | van der waals ferroelectric transistors the all round artificial synapses for high precision neuromorphic computing |
| topic | Ferroelectric transistors Ferroelectric van der Waals heterostructures Artificial synapses Neuromorphic computing |
| url | http://www.sciencedirect.com/science/article/pii/S2709472323000072 |
| work_keys_str_mv | AT zhongwangwang vanderwaalsferroelectrictransistorstheallroundartificialsynapsesforhighprecisionneuromorphiccomputing AT xuefanzhou vanderwaalsferroelectrictransistorstheallroundartificialsynapsesforhighprecisionneuromorphiccomputing AT xiaochiliu vanderwaalsferroelectrictransistorstheallroundartificialsynapsesforhighprecisionneuromorphiccomputing AT aochengqiu vanderwaalsferroelectrictransistorstheallroundartificialsynapsesforhighprecisionneuromorphiccomputing AT caifanggao vanderwaalsferroelectrictransistorstheallroundartificialsynapsesforhighprecisionneuromorphiccomputing AT yahuayuan vanderwaalsferroelectrictransistorstheallroundartificialsynapsesforhighprecisionneuromorphiccomputing AT yumeijing vanderwaalsferroelectrictransistorstheallroundartificialsynapsesforhighprecisionneuromorphiccomputing AT douzhang vanderwaalsferroelectrictransistorstheallroundartificialsynapsesforhighprecisionneuromorphiccomputing AT wenwuli vanderwaalsferroelectrictransistorstheallroundartificialsynapsesforhighprecisionneuromorphiccomputing AT hangluo vanderwaalsferroelectrictransistorstheallroundartificialsynapsesforhighprecisionneuromorphiccomputing AT junhaochu vanderwaalsferroelectrictransistorstheallroundartificialsynapsesforhighprecisionneuromorphiccomputing AT jiansun vanderwaalsferroelectrictransistorstheallroundartificialsynapsesforhighprecisionneuromorphiccomputing |