Hybrid Resistive Memory Devices for Synaptic Plasticity

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current technologies face challenges in simulating the complex information processing mechanisms of neurons, particularly in replicating short-term and long-term plasticities, which are essential for efficient data storage and signal transmission, due to their intricate processes.

Innovation Solution

The development of hybrid resistive memory devices comprising at least two resistive memory units, one configured for short-term plasticity and another for long-term plasticity, using perovskite oxide and transition metal oxides respectively, with specific electrode materials, allowing for serial connection and voltage-controlled operation to achieve multi-level resistance states.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a single resistive memory device is used, then the device structure is simple, but it cannot simulate both short-term and long-term plasticity states of neurons

Engineering Contradiction:

Improveability to simulate synaptic plasticityVSAvoidstructure complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent divides the memory system into multiple resistive memory devices (first RRAM device for short-term plasticity, second RRAM device for long-term plasticity) that are serially connected. Each device is specialized for a specific plasticity state, allowing the system to simulate both STP and LTP functions separately while maintaining manageable complexity in each individual component.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent combines multiple resistive memory devices with different plasticity characteristics into a single hybrid memory system. By merging the first RRAM device (optimized for STP) and the second RRAM device (optimized for LTP) in series, the system achieves comprehensive synaptic simulation capability that neither device could provide alone.

Inventive Principle:

Principle #5Merging (Combining)

2Adaptability or versatility

If multiple resistive memory devices are used to simulate both short-term and long-term plasticity, then the synaptic simulation capability is improved, but the device complexity increases

Engineering Contradiction:

Improvesynaptic plasticity simulation capabilityVSAvoidnumber of memory units

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent designs a universal memory system architecture where the same basic RRAM device structure can serve multiple functions. By configuring identical or similar RRAM devices in series with different weight assignments (first weight for STP, second weight for LTP), the system achieves multi-functionality without requiring fundamentally different device structures, thus reducing overall complexity.

Inventive Principle:

Principle #6Universality (Multi-functionality)

3Reliability

If perovskite oxide is used for short-term plasticity memory layer, then the short-term plasticity performance is improved, but the manufacturing process becomes more complex

Engineering Contradiction:

Improveshort-term plasticity performanceVSAvoidmanufacturing process complexity

Core Design Contradiction:

ReliabilityVSEase of manufacture

Solution Approach 1:

The patent employs composite material structures in the RRAM devices, specifically using perovskite oxide materials (such as Pr0.5Ca0.5MnO3 or Pb(Zr1-xTix)O3) for the memory layer to achieve superior short-term plasticity performance. These composite material choices are integrated into a multi-layer device structure that balances performance enhancement with manufacturability through established thin-film deposition techniques.

Inventive Principle:

Principle #40Composite materials

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 approach enables the hybrid resistive memory device to mimic synaptic plasticity, offering improved data retention properties with both short-term and long-term memory capabilities, similar to neuronal functions, by gradually changing resistances in response to voltage, thus enhancing data storage and processing efficiency.

Implementation Method 1

The resistive memory unit configured to operate in the short-term plasticity state may have a resistance that changes due to movement of oxygen holes or charge trapping

Methodology Applied

Scientific EffectOxygen hole movement: Holes

Implementation Method 2

The resistive memory unit configured to operate in the short-term plasticity state may have a resistance that changes due to movement of oxygen holes or charge trapping

Methodology Applied

Scientific EffectCharge trapping:

Implementation Method 3

The resistive memory unit configured to operate in the long-term plasticity state may have a resistance that changes when filaments are formed

Methodology Applied

Scientific EffectFilament formation:

Data Source

PatentUS8902632B2Hybrid resistive memory devices and methods of operating and manufacturing the same

Publication Date: 2014.12.02 SAMSUNG ELECTRONICS CO LTD

AI summary

Hybrid resistive memory devices and methods of operating and manufacturing the same, include at least two resistive memory units. At least one of the at least two resistive memory units is a resistive memory unit configured to operate in a long-term plasticity state.