Vertical VGAA SRAM Stacking With Oxide Channels for BEOL Scaling

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current silicon-based transistors are reaching performance and scalability limits, and alternative semiconductor materials like germanium and III-V materials are costly and face challenges in ultra-thin body performance scalability.

Innovation Solution

The use of vertical gate all around (VGAA) transistors with channels formed from semiconductive-behaving oxide materials, allowing for vertical stacking of SRAM cells and integration with Back-End-Of-Line processes, utilizing 2D materials for improved transport properties and compatibility with BEOL processes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If silicon-based transistors are continuously scaled down to improve integration density, then more components can be integrated into a given area, but performance and scalability limits are reached

Engineering Contradiction:
Improveintegration densityVSAvoidperformance scalability
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent transitions from planar transistor architectures to vertical channel structures, moving the scaling direction from lateral (2D) to vertical (3D). This dimensional change allows continued improvement of integration density while avoiding the performance degradation associated with excessive lateral scaling of silicon-based devices.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Solution Approach 2:

The patent changes the material parameter from conventional silicon to semiconductive-behaving oxide materials, which offer different electrical characteristics and scalability properties. This material parameter change enables improved performance scalability while maintaining integration density benefits.

Inventive Principle:
Principle #35Parameter changes

2Reliability

If alternative semiconductor materials like germanium and III-V materials are used to overcome silicon limits, then performance can be improved, but manufacturing cost increases

Engineering Contradiction:
ImproveperformanceVSAvoidmanufacturing cost
Core Design Contradiction:
ReliabilityVSEase of manufacture

Solution Approach 1:

The patent employs semiconductive-behaving oxide materials that can be formed using existing BEOL processes, effectively replacing expensive alternative materials like germanium and III-V compounds. These oxide materials provide comparable performance benefits at lower manufacturing cost by utilizing standard industrial processes.

Inventive Principle:
Principle #27Cheap short-living objects (Disposable)

Solution Approach 2:

The patent develops a universal manufacturing approach using BEOL-compatible processes that can form vertical channel structures with semiconductive oxide materials. This universal method replaces multiple specialized manufacturing processes required for different alternative materials, reducing overall manufacturing complexity and cost.

Inventive Principle:
Principle #6Universality (Multi-functionality)

3Area of stationary object

If vertical stacking of SRAM cells is implemented to improve area efficiency, then device footprint is reduced, but manufacturing complexity increases

Engineering Contradiction:
Improvedevice footprintVSAvoidstacking structure
Core Design Contradiction:
Area of stationary objectVSDevice complexity

Solution Approach 1:

The patent divides the memory device into multiple stacked SRAM cell layers, with each layer containing vertically oriented channels. This segmentation allows the device to achieve high area efficiency while managing complexity through modular layering, where each layer can be processed using similar BEOL-compatible techniques.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent stacks SRAM cells in the vertical dimension rather than expanding laterally, transforming the device architecture from a 2D planar layout to a 3D stacked configuration. This dimensional change achieves superior area efficiency while the use of standard vertical processing techniques keeps manufacturing complexity manageable.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

4Reliability

If 2D materials are used for channel formation to improve transport properties, then electrostatic control is enhanced, but process compatibility challenges arise

Engineering Contradiction:
Improveelectrostatic controlVSAvoidBEOL process compatibility
Core Design Contradiction:
ReliabilityVSEase of manufacture

Solution Approach 1:

The patent employs semiconductive-behaving oxide materials that can be formed in-situ using standard BEOL deposition and processing techniques. These materials self-organize into vertical channel structures with excellent electrostatic control, eliminating the need for separate 2D material transfer and integration processes that would compromise BEOL compatibility.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent modifies the material parameters by using oxide materials with semiconductive behavior that can be deposited as thin films using conventional BEOL processes. This parameter change maintains the electrostatic control benefits of 2D-like materials while achieving full process compatibility with existing manufacturing workflows.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentUS20250374510A1Stackable memory devices with vertical channels and methods of manufacturing thereof
Publication Date: 2025.12.04 TAIWAN SEMICONDUCTOR MANUFACTURING CO LTD
  • US20250374510A1 patent drawing
  • US20250374510A1 patent drawing
  • US20250374510A1 patent drawing

AI summary

A memory device includes a first n-type transistor and a second n-type transistor formed of a first channel extending along a vertical direction and wrapped by first, second, third, fourth, and fifth metal tracks; a third n-type transistor and a fourth n-type transistor formed of a second channel extending along the vertical direction and is wrapped by fourth, sixth, seventh, eighth, and ninth metal tracks; a first p-type transistor formed of a third channel extending along the vertical direction and is wrapped by second, third, and tenth metal tracks; and a second p-type transistor formed of a fourth channel extending along the vertical direction and is wrapped by sixth, seventh, and tenth metal tracks.