Polarization-sensitive neuromorphic vision sensing enabled by pristine black arsenic-phosphorus.

Area of Science:

• Materials Science
• Neuroscience
• Optoelectronics

Background:

• Polarization-sensitive vision systems offer advantages in glare reduction and clarity.
• Neuromorphic computing aims to mimic human neural functions for advanced processing.
• Existing systems face challenges in complex lighting and data interpretation.

Purpose of the Study:

• To develop a polarization-sensitive neuromorphic phototransistor for advanced vision applications.
• To investigate the synaptic behaviors and plasticity of the phototransistor.
• To demonstrate its capabilities in image classification, reconstruction, and polarization-resolved imaging.

Main Methods:

• Fabrication of a phototransistor using black arsenic-phosphorus (BaP) 2D nanosheets.
• Characterization of optoelectronic properties, including responsivity and polarization ratio.
• Simulation of synaptic behaviors and plasticity through gate voltage and polarization control.
• Development of a hybrid optical-electronic neural network for image tasks.
• Implementation of polarization-resolved imaging for target reconstruction.

Main Results:

• The BaP phototransistor exhibited high responsivity (2.88 A/W), polarization ratio (4.7), and dynamic range (40 dB).
• Device successfully simulated synaptic behaviors with paired-pulse facilitation up to 201% and demonstrated gate-tunable plasticity.
• Achieved >90% classification accuracy on Fashion-MNIST and 71.38% reconstruction accuracy on Yale Face Database.
• Demonstrated high-fidelity reconstruction of hidden targets using polarization-resolved imaging.

Conclusions:

• The BaP phototransistor serves as a foundational platform for high-performance neuromorphic vision systems.
• Integrated polarization imaging, computation, and communication functionalities are enabled.
• Addresses critical challenges in scalable brain-inspired optoelectronic technologies.