Abstract
We report an artificial optoelectronic synapse based on a copper-phthalocyanine (CuPc) and para-sexiphenyl (p-6P) heterojunction structure. This device features stable conductance states and their linear distribution in long-term potentiation (LTP) characteristic curve formed by continuous input light pulses. These superior synaptic characteristics originate from the fact that the number of photo-holes moving into the CuPc channel and photo-electrons being trapped at the p-6P/dielectric interface is constant at every light pulse. A single-layer neural network is theoretically formed with these optoelectronic synaptic devices and its feasibility is studied in terms of training/recognition tasks of the Modified National Institute of Standards and Technology digit image patterns. Owing to the excellent LTP characteristic and through the use of a unidirectional update method, its maximum recognition rate is as high as 78% despite the use of a single-layer network. This study is expected to provide a foundation for future studies on optoelectronic synaptic devices toward the implementation of complex artificial neural networks.
| Original language | English |
|---|---|
| Article number | 104095 |
| Journal | Nano Energy |
| Volume | 66 |
| DOIs | |
| Publication status | Published - 2019 Dec |
Fingerprint
All Science Journal Classification (ASJC) codes
- Renewable Energy, Sustainability and the Environment
- Materials Science(all)
- Electrical and Electronic Engineering
Cite this
}
Solar-stimulated optoelectronic synapse based on organic heterojunction with linearly potentiated synaptic weight for neuromorphic computing. / Qian, Chuan; Oh, Seyong; Choi, Yongsuk; Kim, Jeong Hoon; Sun, Jia; Huang, Han; Yang, Junliang; Gao, Yongli; Park, Jin Hong; Cho, Jeong Ho.
In: Nano Energy, Vol. 66, 104095, 12.2019.Research output: Contribution to journal › Article
TY - JOUR
T1 - Solar-stimulated optoelectronic synapse based on organic heterojunction with linearly potentiated synaptic weight for neuromorphic computing
AU - Qian, Chuan
AU - Oh, Seyong
AU - Choi, Yongsuk
AU - Kim, Jeong Hoon
AU - Sun, Jia
AU - Huang, Han
AU - Yang, Junliang
AU - Gao, Yongli
AU - Park, Jin Hong
AU - Cho, Jeong Ho
PY - 2019/12
Y1 - 2019/12
N2 - We report an artificial optoelectronic synapse based on a copper-phthalocyanine (CuPc) and para-sexiphenyl (p-6P) heterojunction structure. This device features stable conductance states and their linear distribution in long-term potentiation (LTP) characteristic curve formed by continuous input light pulses. These superior synaptic characteristics originate from the fact that the number of photo-holes moving into the CuPc channel and photo-electrons being trapped at the p-6P/dielectric interface is constant at every light pulse. A single-layer neural network is theoretically formed with these optoelectronic synaptic devices and its feasibility is studied in terms of training/recognition tasks of the Modified National Institute of Standards and Technology digit image patterns. Owing to the excellent LTP characteristic and through the use of a unidirectional update method, its maximum recognition rate is as high as 78% despite the use of a single-layer network. This study is expected to provide a foundation for future studies on optoelectronic synaptic devices toward the implementation of complex artificial neural networks.
AB - We report an artificial optoelectronic synapse based on a copper-phthalocyanine (CuPc) and para-sexiphenyl (p-6P) heterojunction structure. This device features stable conductance states and their linear distribution in long-term potentiation (LTP) characteristic curve formed by continuous input light pulses. These superior synaptic characteristics originate from the fact that the number of photo-holes moving into the CuPc channel and photo-electrons being trapped at the p-6P/dielectric interface is constant at every light pulse. A single-layer neural network is theoretically formed with these optoelectronic synaptic devices and its feasibility is studied in terms of training/recognition tasks of the Modified National Institute of Standards and Technology digit image patterns. Owing to the excellent LTP characteristic and through the use of a unidirectional update method, its maximum recognition rate is as high as 78% despite the use of a single-layer network. This study is expected to provide a foundation for future studies on optoelectronic synaptic devices toward the implementation of complex artificial neural networks.
UR - http://www.scopus.com/inward/record.url?scp=85072066051&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85072066051&partnerID=8YFLogxK
U2 - 10.1016/j.nanoen.2019.104095
DO - 10.1016/j.nanoen.2019.104095
M3 - Article
AN - SCOPUS:85072066051
VL - 66
JO - Nano Energy
JF - Nano Energy
SN - 2211-2855
M1 - 104095
ER -