Ferroelectric FET Synapse Array for Neuromorphic Integration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current neuromorphic devices face challenges in achieving high integration and efficient synaptic operations due to limitations in their design and materials, particularly in mimicking the complex synaptic functions of the human brain.
Innovation Solution
The use of ferroelectric field effect transistors in a synapse array of a neuromorphic device, which includes a common floating gate electrode, individual ferroelectric films, and control gate electrodes, allows for parallel and series connections of transistors, enabling excitatory and inhibitory synaptic operations and high integration levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional neuromorphic device designs are used, then device simplicity is maintained, but integration level and synaptic operation efficiency are insufficient
Solution Approach 1:
The synapse is divided into multiple ferroelectric field effect transistors (FEFETs) connected in parallel, with each FEFET representing a separate synaptic weight element. This segmentation enables high integration by allowing multiple synapses to be packed into a compact array structure while maintaining independent control of each synaptic connection.
Solution Approach 2:
Multiple FEFETs are merged into a single synapse unit through parallel connection, sharing common source and drain regions. This merging reduces the overall area required for each synapse while preserving the ability to independently modulate each synaptic weight through individual gate control.
2Use of energy by moving object
If simple transistor configurations are used, then device simplicity is maintained, but power efficiency and ability to perform complex synaptic operations are limited
Solution Approach 1:
The FEFET configuration enables dynamic control of synaptic weights through voltage application to the gate, allowing the synapse to switch between different conductance states. This dynamic capability allows complex synaptic operations (excitatory and inhibitory) to be performed with controlled power consumption, improving power efficiency for neuromorphic computing tasks.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This configuration enhances the integration and power efficiency of neuromorphic devices, enabling them to perform complex synaptic operations with multiple resistance levels, thereby improving their ability to mimic brain-like learning and recognition processes.
Implementation Method 1
a synapse array of a neuromorphic device including a synapse array having a plurality of ferroelectric field effect transistors
Data Source
AI summary
A neuromorphic device having a synapse array is provided. The synapse array of the neuromorphic device may include an input neuron; an output neuron; and a synapse. The synapse may include a plurality of ferroelectric field effect transistors electrically connected to each other in parallel.