Optoelectronic black phosphorus used for neural networking

July 19, 2019 //By Peter Clarke
Optoelectronic black phosphorus used for neural networking
Researchers at Royal Melbourne Institute of Technology (RMIT) have used thin layers of black phosphorous to develop neural network circuit that can be controlled optically.

Layered black phosphorous is a 2D material – similar to graphene – that tends to oxidize under ambient conditions. These intrinsic defect states give rise to novel photoelectronic  effects including the ability to distinguish between UV-A and UV-B light. The researchers were able to generate a photocurrent that could be switched in polarity thus simulating synaptic behaviour. Switching between colors causes the current to reverse direction from positive to negative.

The research team also demonstrated that serially connected devices can be used to perform digital logic operations under optical control.

"We're able to simulate the brain’s neural approach simply by shining different colours onto our chip," said research leader Sumeet Walia. Taimur Ahmed is lead author of the study, which was published in Advanced Functional Materials.

The technology was developed at the Micro Nano Research Facility and is compatible with existing electronics and has also been demonstrated on a flexible platform, for integration into wearable electronics.

Related links and articles:

www.rmit.edu.au

News articles:

Experimental optical chip works like a brain

Organic transistor mimics synapse, suitable for ML/AI

Spiking NN core for unsupervised learning available as IP

Photonics-based AI chips run at attojoule levels

Intel scales up self-learning neuromorphic computing


Vous êtes certain ?

Si vous désactivez les cookies, vous ne pouvez plus naviguer sur le site.

Vous allez être rediriger vers Google.