Press

Emulating the plasticity of synapses between neurons in a physical system by light stimuli

This is the first case of a 2D electron system, in which the electron population can be increased and reduced in a controlled way by light. The study, published in Physical Review Letters, has been carried out by Gervasi Herranz and co-workers from the Laboratory of Multifunctional Thin Films and Complex Sructures (MULFOX) group at the ICMAB. 

In a system where electrons can only move in 2 dimensions (2D), at the interface between two insulating materials, researchers at the Institute of Materials Science of Barcelona (ICMAB-CSIC) have managed to increase and decrease in a plastic way the conductance of the system by means of light stimuli of different wavelengths.

It has been observed that applying a short wavelength, corresponding to blue light, the conductance is increased, i.e. the amount of electrons present is higher, while applying a longer wavelength, corresponding to red light, the conductance decreases, i.e. the amount of electrons is reduced.

This fact is very interesting because it manages to emulate the nervous synapses between neurons, which can be strengthened or weakened. In addition, like the synapses, the changes of conductance observed are plastic, since the conductance is always modified, never reaches the initial state.

The main novelty of the study is the way in which the changes in conductance have been achieved.

"Until now, the plastic synapse in conductance in non-biological physical systems had been achieved by means of electrical stimuli, in the so-called memristors. Now we have proved for the first time that stimuli generated with optical pulses, i.e. with small pulses of light achieved with monochromatic lasers, can plastically increase or reduce the conductance with pulses on the scale from seconds to milliseconds" points out Gervasi Herranz, researcher of the MULFOX group, at the ICMAB.

The sensitivity of conductance to these light pulses opens up fascinating perspectives on the use of optical synapses for neuromorphic devices based on these photoconductive systems. This fact is especially interesting for future applications in optoelectronic devices, and also for future artificial vision. One of the next challenges of Gervasi Herranz is to achieve artificial neurons that emit impulses (spikes) in response to variations in synaptic connections (represented by the conductance of the 2D electronic gas).

 imatge highlight Gervasi photoinduced

Figure: The image describes this research very well. The graph below shows the conductivity over time for two different sequences of light pulses of different wavelengths. In the blue sequence, peaks in conductivity are seen every time there is a blue light pulse for 5 ms (blue arrow), and the conductivity increases. During the time when there is no light, the conductivity slowly decreases, but never reaches zero. In the orange sequence, you see peaks of conductivity every time there is a pulse of blue or red light for 5 ms (blue and red arrows), but you see that when the light pulse is red, the conductivity decreases, instead of increasing. That is to say, by means of different wavelengths, the conductivity of the system is increased or decreased, which is equivalent to saying that the electron population is increased or reduced in a controlled way by means of light.

Cover Figure: The sensitivity of conductance to these light pulses opens up fascinating perspectives on the use of optical synapses for neuromorphic devices based on these photoconductive systems.

Reference article:

Photoinduced persistent electron accumulation and depletion in LaAlO3/SrTiO3 quantum wells
Yu Chen, Yoann Lechaux, Blai Casals, Bruno Guillet, Albert Minj, Jaume Gázquez, Laurence Méchin, and Gervasi Herranz.
Phys. Rev. Lett. 124, 246804 – Published 19 June 2020
DOI: 10.1103/PhysRevLett.124.246804

Download here the press release in CAT, ESP, ENG.

Press Release, Top Science Results

  • Hits: 335

INSTITUT DE CIÈNCIA DE MATERIALS DE BARCELONA, Copyright © 2020 ICMAB-CSIC | Privacy Policy | This email address is being protected from spambots. You need JavaScript enabled to view it.