Expand this Topic clickable element to expand a topic
Skip to content
Optica Publishing Group

Optoelectronic neurons using MSM detectors in GaAs

Open Access Open Access

Abstract

We present a new GaAs integrated optoelectronic neuron for use in neural networks. We have previously demonstrated that the integration of photodetectors, thresholding transistors, and a light source on a single substrate allows one to have high neuron density with acceptable power dissipation. In this paper we report a circuit in which we used a double heterostructure LED (light emitting diode) as the light source. We use an LED rather than laser diodes because LEDs can be operated with small currents, due to the lack of a threshold current. To minimize the total mesa height and maintain high LED quantum efficiency, MSM (metal–semiconductor–metal) photodetectors were used. The thresholding transistor was a MESFET (metal–semiconductor field-effect transistor). The total mesa height is less than 3 μm. Since no two devices share an epitaxial layer, each device is individually optimized. Our circuit uses two photodetectors, one to set the threshold voltage and the other to detect the signal. With this circuit, we can also build both excitatory and inhibitory neurons. Results from our experimental studies are presented.

© 1991 Optical Society of America

PDF Article
More Like This
GaAs optoelectronic neurons

Steven Lin and Demetri Psaltis
MK1 OSA Annual Meeting (FIO) 1991

GaAs-Based Optoelectronic Neurons

Steven Lin, Francis Ho, Jae Kim, and Demetri Psaltis
WB3 Optical Computing (IP) 1991

High-gain GaAs optoelectronic thresholding devices for neural network implementation

Steven Lin, Demetri Psaltis, and Jae Kim
TuC1 Integrated Photonics Research (IPR) 1991

Select as filters


Select Topics Cancel
© Copyright 2024 | Optica Publishing Group. All rights reserved, including rights for text and data mining and training of artificial technologies or similar technologies.