Functional Model of a Nanoparticle Organic Memory Transistor for Use as a Spiking Synapse

Abstract

Emerging nanocomponents are of great interest to provide adaptability, high density and robustness for the development of new bio-inspired circuits or systems. Although CMOS Neuromorphic circuit was one of the most intense researches to bring the adaptability and robustness in the circuit beyond the conventional Von Neumann architecture in early 1990', CMOS technology could not provide the huge capacity to be scalable to biological levels because a great number of transistors are required to emulate the dynamical behaviors of biological synapse [1]. Nanoscale components are therefore of great interest to develop new neuromorphic circuits by replacing CMOS technology based synapse. The NanoparticleOrganic Memory transistor (NOMFET) is one of the most promising candidates as it can exhibit dynamical behaviors similar to a biological synapse [2]. It could be very suitable to implement some natural synaptic learning mechanisms such as Spike-Time-Dependent-The NOMFET is composed of three terminals as the conventional MOSFET, Drain (D), Source (S) and Gate (G). The device is fabricated using a bottom-gate electrode configuration. Gold nanoparticles (NP) are immobilized on the surface of the inter-electrode gap before pentacene deposition. The conduction in the device is assured by holes, created in the thin film at the interface with the silicon oxide when a negative gate voltage is applied. In addition to rather classical transistor behavior, a negative gate voltage also positively charges the Au NPs. This has the effect of diminishing the channel conductivity of the device, because the charged NPs cause a repulsive electrostatic interaction between the holes trapped in the NPs and the ones created in the pentacene. The NOMFET therefore exhibits a short term memory and the charge retention time in the NPs can be as high as several thousand seconds [4].A functional model is very useful for the design of hybrid Nano/CMOS circuit and architecture as it provides the interface between the fundamental physics and the electrical behaviors. We established the functional model of NOMFET for a two-terminal device configuration, as shown Fig. 1. The gate and the drain electrodes are driven by the same input voltage, a pulse train of a variable frequency. In this configuration, the dynamical behaviors of the NOMFET is a lot as a biological synapse: there is a competition between the charges provided to the NPs by the gate voltage pulses (resulting in a decrease of the drain-to-source current) and the natural NPs charge relaxation (which increases the current intensity until the NPs are fully discharged). When a new input pulse sequence occurs, the NOMFET exhibits either a depressing or a facilitating behavior, depending both on the duration between the pulses and on the charge level of the NPs. The current in the device is thus dependent on the history of the input signal; the maximum variation on our devices is close to 25% with 30V pulses amplitude.Our model is iterative: the drain-to-source current response to a voltage pulse...

Low voltage and time constant organic synapse-transistor

“…This device (which is memristor-like) 8 mimics short-term plasticity (STP) 7 and temporal correlation plasticity (STDP, spike-timing dependent plasticity) 8 of biological spiking synapses, two "functions" at the basis of learning processes. A compact model was developed 9 and we demonstrated an associative memory, which can be trained to exhibit a Pavlovian response. 10 However, these devices were limited to work with spikes in the range of few tens of volts and time scale of 1-100s.…”

Simulation Study on Silicon-Based Floating Body Synaptic Transistor with Short- and Long-Term Memory Functions and Its Spike Timing-Dependent Plasticity

“…Recently, there are great interests in realizing artificial synapses based on the Hebbian learning for the brain-like computing architectures [1][2][3][4][5][6][7][8][9][10]. This is because a synapse is believed to make a great contribution to many cognitive functions such as perception and memory in a biological system [11][12][13][14][15][16].…”

On-Chip Universal Supervised Learning Methods for Neuro-Inspired Block of Memristive Nanodevices

“…In this context, neuro-inspired networks are considered as a relevant candidate to implement adaptive architectures, the nanodevices have been used as synapses [Pershin and Di Ventra 2010;Versace and Chandler 2010;Snider 2007; This work was supported by the French National Research Agency under the MOOREA program. Bichler et al 2010;Lee and Likharev 2007;Liao et al 2011;Hasegawa et al 2010;Lai et al 2010;Bichler et al 2013]. Suitable learning methods for the nanodevices are required.…”