• Embed Doc
  • Copy Link
  • Readcast
  • Collections
  • CommentGo Back
Download
 
1
An organic-nanoparticle transistor  behaving as a biological spiking synapse
Fabien Alibart,
1
Stéphane Pleutin,
1
David Guerin,
1
Christophe Novembre,
2
Stéphane Lenfant,
1
Kamal Lmimouni,
1
 Christian Gamrat,
2
and Dominique Vuillaume.
1,*
1. Molecular Nanostructures and Devices group, Institute for Electronics Microelectronics andNanotechnology, CNRS, University of Lille, BP60069, avenue Poincaré, F-59652cedex, Villeneuved'Ascq, France.2. CEA, LIST/LCE (Advanced Computer technologies and Architectures), Bat. 528, F-91191, Gif-sur-Yvette, France.* corresponding author: dominique.vuillaume@iemn.univ-lille1.fr
 
2ABSTRACT.We demonstrate a device made of molecules and nanoparticles, a nanoparticle organic memory filed-effect transistor (NOMFET), which exhibits the main behavior of a biological spiking synapse. Short-term plasticity, facilitating and depressing behaviors can be programmed. The synaptic plasticity for realtime computing is evidenced and described by a simple model. We can tailor the dynamic behavior inthe frequency/time domain by adjusting the size of both the NPs and the NOMFET. These results openthe way to use NOMFET in dynamical neuromorphic computing circuits, or as a building block of neuroelectronics for interfacing biologic neurons or neuronal logic devices with solid-state devices andcircuits.
 
3It is now well recognized that electronic circuits based on von Neuman paradigm are unable to catchthe complex, real-world environment, behaviors as do a biological neural system (e.g. human brain).One of the reasons is the so-called von Neuman bottleneck 
1
due to the physical separation of computingunits and memories. In the brain, memory and computation are mixed together and allow the processingof information both in time and in space via the time dependent properties of interconnected neurons.This challenge for the development of a new generation of computers has induced a lot of efforts inneuroscience computation activities and the framework for the spatiotemporal processing of informationseems to be theoretically achievable.
2
One key element that is still a limitation concerns the integrationof neurons and synaptic connections in order to realize a brain-like computer. Even if silicon CMOSchips have been designed and fabricated to emulate the brain behaviors,
3, 4
this approach is limited tosmall system because it takes several (at least 7) silicon transistors to build an electronic synapse.
5
Asthe human brain contain more synapses than neurons (by a factor of ~ 10
7
), it is mandatory to develop anano-scale, low power, synapse-like device if we want to scale neuromorphic circuits towards thehuman brain level. This feature has recently prompted the research for nano-scale synaptic devices.Proposals for such programmable memory devices include optically-gated CNTFET,
6
organic/hybrid Sinanowire transistor,
7
memristor
8-11
. Moreover, in neuroscience, experiments coupling neurons or neuraltissues with CMOS circuits are a corner stone in the quest for understanding the brain functioning.
12
Thedevelopment of a micro- or nano-sized artificial synapse may help for interfacing neurons or neuronallogic devices made from patterned neuronal cultures with solid-state devices and circuits.
13, 14
 Here, we show that mixing nanoparticles (NPs) and molecules to implement computation andmemory in a single synapse-like device is a powerful approach towards such objectives. NPs andmolecules are nano-size objects suitable for nano-device fabrication, they can be manipulated andassembled by low-cost, bottom-up, techniques (e.g. self-assembly),
15, 16
they are prone to work onflexible, plastic, substrates (see the tremendous efforts on plastic, printable, organic electronics
17-19
).
of 00

Leave a Comment

You must be to leave a comment.
Submit
Characters: ...
You must be to leave a comment.
Submit
Characters: ...