Elsevier

Nano Energy

Volume 65, November 2019, 104035
Nano Energy

Full paper
Versatile neuromorphic electronics by modulating synaptic decay of single organic synaptic transistor: From artificial neural networks to neuro-prosthetics

https://doi.org/10.1016/j.nanoen.2019.104035Get rights and content

Highlights

  • Synaptic decay of ion-gel gated organic synaptic transistors (IGOSTs) was easily modulated.

  • The relation between the crystallinity and electrochemical doping kinetics in the IGOSTs were discovered.

  • Artificial auditory nerve with short-term decay and artificial neural networks with long-term memory were demonstrated.

  • Versatile neuromorphic application from neuromorphic computing to neuro-prosthetics was achieved.

Abstract

Organic neuromorphic electronics are inspired by a biological nervous system. Bio-inspired computing mimics learning and memory in a brain (i.e., the central nervous system), and bio-inspired soft robotics and nervous prosthetics mimics the neural signal transmission of afferent/efferent nerves (i.e., the peripheral nervous system). Synaptic decay time of nerves differ among biological organs, so the decay time of artificial synapses should be tuned for their specific uses in neuro-inspired electronics. However, controlling a synaptic decay constant in a fixed synaptic device geometry for broad applications was not been achieved in previous research of neuromorphic electronic devices despite the importance to achieve broad applications from neuromorphic computing to neuro-prosthetics. Here, we tailored the synaptic decay constant of organic synaptic transistors with fixed materials and devices structure rather than changing the form of presynaptic spikes, which enabled broad applications from neuromorphic computing to neuro-prosthetics. To achieve this, the relation between crystallinity of the polymer semiconductor film and the synaptic decay constant was revealed. The crystallinity of the polymer controlled electrochemical-doping kinetics and resultant synaptic behaviors of artificial synaptic transistors. In this way, we demonstrated not only long-term retention for learning and memory that is useful for neuromorphic computing in ion-gel gated organic synaptic transistor (IGOST) but also the short-term retention for fast synaptic transmission that is useful for emulating peripheral nerves such as sensory and motor nerve. To prove the feasibility of our approach in a two different ways, we first simulated pattern recognition on the MNIST dataset of handwritten digits using an IGOST with long-term retention due to increased crystallinity and then, developed artificial auditory sensory nerves that combines an IGOST with short term retention due to disordered chain morphology in a polymer semiconductor, with a triboelectric acoustic sensor. We expect that our approach will provide a universal strategy to realize wide neuromorphic electronic applications.

Introduction

A biological nervous system is an extremely compact and well--organized collection of a large number of neurons and synapses [1,2]. Neural signals (i.e., action potentials) are transmitted through the neurons and synapses, which result in learning and memory in the brain (i.e., central nerves) or sensing and action in the body (i.e., peripheral nerves). Biological synapses have different decay-time constants depending on their roles and locations [[3], [4], [5], [6]]. Emulating the various decay time of synapses may be an important step toward developing various neuromorphic devices that mimic different nerves in central and peripheral nervous systems.

Development of neuromorphic devices has mostly focused on mimicking learning and memory of a biological brain [[7], [8], [9]]. On the other hand, organic neuromorphic electronics for brain-inspired computing and bio-inspired soft robotics can emulate both rules of neural signal transmission and neuroplasticity in a brain and peripheral nervous system [7,[10], [11], [12], [13], [14], [15]]. Furthermore, organic synaptic transistors have advantages of low-cost solution fabrication, low-voltage-driven operation, low energy consumption, and precise emulation of synaptic plasticity, and can therefore be used universally for various neuromorphic electronic devices [9,13,16,17].

Recently-developed organic synaptic transistors exploit ion migration in the electrolyte and their electrochemical doping to the conjugated polymers [9,13,16]. Short and long-term synaptic behaviors of ion-gel gated organic synaptic transistor (IGOST) are caused primarily by migration of ions in the electrolyte dielectric to form electrostatic electric double layers (EDLs), and by electrochemical doping of ions into organic semiconductor (OSC), respectively [16]. Stable and heavy doping of ions into an OSC maintains increased conductance (i.e., synaptic weight) for a long time; the result is long-term retention and slow synaptic decay, which are important for neuromorphic computing and non-volatile memory. In contrast, short-term retention and fast synaptic decay are important for instant signal transmission of artificial sensory and motor nervous systems [18,19]. At this infant stage of organic neuromorphic devices, understanding the effect of OSC morphology on electrochemical-doping of ions and its synaptic plasticity is a critical importance to make effective devices for an intended use in neuromorphic computing and neuro-prosthetics. But until now, this has not been studied in organic synaptic transistors and there was no attempt to control the synaptic decay of single IGOST to make them applicable in diverse neuromorphic comupting and bioelectronic device systems: previous researches on organic synaptic transistors targeted only on single neuromorphic system that mimic either brains or biological afferent(sensory) nerves for either neuromorphic computing or neuro-prosthetics [9,18,20]. In conventional neuromorphic devices, to modulate the synaptic decay constant of the device from the short-term potentiation (STP) to long-term potentiation (LTP), the form of presynaptic spike (i.e., amplitude of spike, number of spikes etc.) was adjusted. Instead, tailoring of the synaptic properties of the devices in a fixed device using the same active material are necessary for broad applications.

In this study, we realized a synaptic-property-tunable IGOST using single intrinsic semiconducting polymer that emulate universal synaptic behaviors of both brain and peripheral nervous systems. We manipulated the morphology of OSC thin films to engineer electrochemical interaction between ions and OSC thin films in the IGOST. This process results in different synaptic decay behaviors so that the device mimics both cerebral computing and memory, and the responses of peripheral nerves depending on the morphologies of the polymer semiconductor films.

Depending on the crystallinity and grain size of OSC films, the IGOST showed a clear transition of synaptic responses between short-term potentiation (STP) and long-term potentiation (LTP) without adjusting the form of presynaptic spikes. The IGOST also demonstrated various synaptic properties such as paired-pulse facilitation (PPF), spike number dependent potentiation (SNDP), spike duration dependent potentiation (SDDP), spike frequency dependent potentiation (SFDP) and spike voltage dependent potentiation (SVDP). To prove the validity of our strategy, we performed simulations of pattern recognition of handwritten digits by highly crystalline IGOST with long synaptic decay and suggested that they have an accuracy of up to 94% and also developed an artificial auditory nerve that combines a triboelectric acoustic sensor and fast responsive IGOST with short synaptic decay.

Section snippets

Device fabrication

Thin poly(thienoisoindigo-naphthalene) (molecular weight Mn = 36 kDa, PDI = 2.06) films were spincoated from a solution of 1,2-dichlorobenzene (~5 mg⋅ml-1) on Si/SiO2 (100 nm) substrate. Subsequently, the films were thermally annealed at 80, 150, 200, 250, or 310 ℃ for 10 min. Au source-drain electrodes (30 nm) were deposited by thermal evaporation. Ion-gel with poly(styrene--block-methyl methacrylate-block-styrene) (PS-PMMA-PS) and 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide

Film morphology control and I–V characterization

IGOST were fabricated with poly(thienoisoindigo-naphthalene) (PTIIG-Np) thin-film as an active layer, and 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ([EMIM][TFSI]) as an ion-gel dielectric (Fig. 1a). PTIIG-Np is a thienoisoindigo polymer with bicyclic lactam unit and thiophene derivative acceptor incorporated with a centrosymmetric acene naphthalene donor (Fig. 1b) that improves structural ordering, molecular packing and charge transport [21]. Also, alternative donor-acceptor

Conclusion

We achieved different synaptic decay times in single IGOST without changing organic semiconducting polymer and device geometry but by controlling the morphological properties of the polymer film. As polymer crystallinity changed, the synaptic decay properties of the devices changed from STP-dominant to LTP-dominant. Furthermore, engineering the synaptic decay time constant by changing the crystallinity revealed the relationship between the synaptic property and morphological property in IGOST.

Acknowledgements

D.-G. S. and Y. L. equally contributed to this work. This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (Ministry of Science, ICT & Future Planning) (NRF-2016R1A3B1908431, NRF2017R1A2B4009313). This work was also supported by the Center for Advanced Soft-Electronics funded by the Ministry of Science and ICT as Global Frontier Project (2013M3A6A5073175), and Creative-Pioneering Researchers Program through Seoul National University (SNU)

Dae-Gyo Seo received his B.S. (2018) in Department of Materials Science and Engineering from Korea University, Korea. He is studying for his Ph.D. course in the Department of Materials Science and Engineering of Seoul National University. His research interest includes the organic transistors, neuromorphic electronics, and artificial nerve systems.

References (43)

  • B.-Y. Kim et al.

    NPG Asia Mater.

    (2017)
  • S. Bleeck et al.

    Hear. Res.

    (2006)
  • V.M. Ho et al.

    Science

    (2011)
  • L.F. Abbott et al.

    Nature

    (2004)
  • J.F. Otto et al.

    J. Neurosci.

    (2006)
  • M.A. Ungless et al.

    J. Neurophysiol.

    (2002)
  • A.L. Hodgkin et al.

    J. Physiol.

    (1952)
  • E.M. Izhikevich

    IEEE Trans. Neural Netw.

    (2003)
  • M. Hu et al.

    Adv. Mater.

    (2018)
  • P. Yao et al.

    Nat. Commun.

    (2017)
  • Y. Van De Burgt et al.

    Nat. Mater.

    (2017)
  • W. Wang et al.

    Sci. Adv.

    (2018)
  • M. Prezioso et al.

    Nat. Commun.

    (2018)
  • W. Xu et al.

    Adv. Mater.

    (2016)
  • P. Gkoupidenis et al.

    Adv. Mater.

    (2015)
  • G. Wu et al.

    Sci. Rep.

    (2016)
  • W. Xu et al.

    Sci. Adv.

    (2016)
  • S.-I. Kim et al.

    Adv. Electron. Mater.

    (2019)
  • Y. Kim et al.

    Science

    (2018)
  • Y. Lee et al.

    Sci. Adv.

    (2018)
  • T.-Y. Wang et al.

    ACS Appl. Mater. Interfaces

    (2018)
  • Cited by (116)

    • Resistive switching of two-dimensional NiAl-layered double hydroxides and memory logical functions

      2023, Journal of Alloys and Compounds
      Citation Excerpt :

      Therefore, no current jump happens regardless of the state of NiAl-LDHs memristor. Flip-flop is a logic device with storage function that carries on logic operation and can store the result of operation [71–75]. Here, memtransistor makes use of its unique programmability and modulability to build a D-type trigger and verify its logic function.

    View all citing articles on Scopus

    Dae-Gyo Seo received his B.S. (2018) in Department of Materials Science and Engineering from Korea University, Korea. He is studying for his Ph.D. course in the Department of Materials Science and Engineering of Seoul National University. His research interest includes the organic transistors, neuromorphic electronics, and artificial nerve systems.

    Yeongjun Lee is a postdoctoral researcher in Materials Science and Engineering at Seoul National University, Korea. He received his M.S. (2014) and Ph.D. (2018) in Materials Science and Engineering from Pohang University of Science and Technology (POSTECH), Korea. His research interests include printed electronics, nanowire electronics, stretchable polymer electronics, and organic neuromorphic electronics.

    Gyeong-Tak Go is studying for his Ph.D Course in the Department of Materials Science and Engineering of Seoul National University. His research interests include the organic neuromorphic electronics based on artificial synapse devices.

    Mingyuan Pei is currently studying for Ph.D. degree under the supervision of Prof. Hoichang Yang from Department of Chemical Engineering at Inha University, Korea. He received his B.S. and M.S. degree from the Department of Materials Engineering at Tianjin Polytechnic University in 2011 and 2014, respectively. His research focuses on the structure and performance of organic thin film transistors.

    Sungwoo Jung is in the combined master's and doctorate course for Energy Engineering in Ulsan National Institute of Science and Technology (UNIST). He received B.S. in 2017 Energy Engineering from UNIST. He joined Professor Changduk Yang's group in 2017 and his main research focus is on synthesizing organic materials which will be applied for optoelectronic devices and triboelectric nanogenerator.

    Yo Han Jeong is a Master with Prof. Sang-Woo Kim at SKKU Advanced Institute of Nanotechnology (SAINT), Sungkyunkwan University (SKKU). His research interests are DFT simulation related with new materials for nanogenerators and 2D materials.

    Wanhee Lee is an undergraduate studying in the Department of Physics and Astronomy at Seoul National University. His research interests include device simulation for neuromorphic computing and first principle description of atoms and molecules on two-dimensional materials.

    Hea-Lim Park has joined Prof. Lee's group since 2018 as a postdoctoral researcher. She received her Ph.D. at the School of Electrical Engineering, Seoul National University, Republic of Korea in 2018. She received her B.S. degree in electrical and electronic engineering from Yonsei University in 2012. Her current research interests include organic transistors, sensors, and bio-inspired neuromorphic devices.

    Sang-Woo Kim is an Associate Professor in School of Advanced Materials Science and Engineering at Sungkyunkwan University (SKKU). He received his Ph.D. from Kyoto University in Department of Electronic Science and Engineering in 2004. After working as a postdoctoral researcher at Kyoto University and University of Cambridge, he spent 4 years as an assistant professor at Kumoh National Institute of Technology. He joined the School of Advanced Materials Science and Engineering, SKKU Advanced Institute of Nanotechnology (SAINT) in 2009. His recent research interest is focused on piezoelectric/triboelectric nanogenerators, photovoltaics, and two-dimensional nanomaterials including graphene and hexagonal boron nitride nanosheets.

    Hoichang Yang is currently a professor at the Department of Chemical Engineering, Inha University since 2009. He received his Ph.D. from Chemical Engineering, Pohang University of Science and Technology (POSTECH) in 2002. He had been a senior researcher at the Rensselaer Nanotechnology Center in Rensselaer Polytechnic Institute (2003–2008). His research interest are in the field of surface/interface engineering, self-assembly related to energy and organic electronic applications, specifically, organic thin film transistors.

    Changduk Yang obtained his Ph.D. degree from the Max Plank Institute for Polymer Research (Germany) in 2006 under Prof. Klaus Mullen. He finished his postdoctoral training in 2009 at the University of California Santa Barbara with Prof. Fred Wudl. In March 2009, he joined the Ulsan National Institute of Science and Technology (UNIST), Korea. He serves on the editorial board for Polymer (Korea) and works as a full-time professor at UNIST. His research focus is on the development of organic semiconducting materials and multifunctional molecule-related self-assembly for various applications, including organic solar cells, organic optoelectronics, triboelectric generations, and secondary batteries.

    Tae-Woo Lee is a professor in Materials Science and Engineering at Seoul National University, Korea. He received his Ph.D in Chemical Engineering from KAIST, Korea in 2002. He joined Bell Laboratories, USA as a postdoctoral researcher and worked in Samsung Advanced Institute of Technology as a member of the research staff (2003–2008). He was an associate professor in Materials Science and Engineering at Pohang University of Science and Technology (POSTECH), Korea until Aug 2016. His research focuses on printed flexible electronics based on organic and organic-inorganic hybrid materials for flexible displays, nerve-inspired artificial electronic devices, solid-state lightings, and solar-energy-conversion devices.

    1

    These authors equally contributed to this work.

    View full text