Mixed-Dimensional Floating Gate Phototransistors for Mixed-Modal In-Sensor Reservoir Computing.

Area of Science:

• Materials Science
• Neuromorphic Engineering
• Quantum Computing

Background:

• Low-dimensional materials like quantum dots (QDs) and 2D materials show promise for neuromorphic devices.
• Combining QDs as floating gates and 2D materials as channels is an underexplored area.
• Existing devices often lack the capacity for mixed-modal sensory input processing.

Purpose of the Study:

• To introduce a novel floating-gate phototransistor based on a mixed-dimensional heterostructure of 0D-CsPbBr3 QDs and 2D-MoS2.
• To realize mixed-modal in-sensor reservoir computing (RC) by leveraging the unique properties of QDs and MoS2.
• To demonstrate the device's capability for audio-visual fusion and recognition tasks.

Main Methods:

• Fabrication of a heterostructure device using 0D-CsPbBr3 QDs and few-layer 2D-MoS2.
• Characterization of the device's optoelectronic properties, including on/off ratio and multistate behavior.
• Implementation of mixed-modal reservoir computing with optical and electrical inputs for data processing and recognition.

Main Results:

• The device exhibits a high on/off ratio of 10^7 and over 7-bit multistates.
• Demonstrated nonlinear memory decay and tunable dynamic time scales.
• Successfully achieved mixed-modal reservoir computing with mixed optical and electrical signals.
• Accurate recognition of endangered species under challenging environmental conditions via audio-visual fusion.

Conclusions:

• The developed mixed-dimensional heterostructure device is a promising platform for advanced neuromorphic computing.
• The device enables efficient mixed-modal information fusion, mimicking biological sensory systems.
• This work opens new avenues for creating sophisticated sensory processing systems by integrating materials of different dimensions.