3D Memory Stack Row Transfer Circuit With Split Pass Transistors

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The increasing demand for higher-integration semiconductor devices is limited by the need for ultra-high-cost devices for finer pattern miniaturization, and existing 2D semiconductor devices have limited applications due to their size constraints, prompting the development of 3D semiconductor devices with memory cells arranged in a three-dimensional structure.

Innovation Solution

A semiconductor device design that distributes pass transistors between a memory chip and a circuit chip, allowing for reduced size and improved high-speed performance by dividing transfer circuits between the chips, thereby reducing the number of pass transistors required per stack, regardless of the increasing number of stacks.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If memory cells are arranged in a three-dimensional structure to increase integration degree, then integration density is improved, but device complexity increases

Engineering Contradiction:
Improveintegration densityVSAvoiddevice complexity
Core Design Contradiction:
Quantity of substanceVSDevice complexity

Solution Approach 1:

The transfer circuit is divided into multiple segments distributed across different chips (first chip and second chip). Each chip contains a portion of the transfer circuit, which connects row lines to global row lines. This segmentation reduces the complexity burden on any single chip while maintaining the overall 3D integrated structure.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent transitions from 2D to 3D stacking architecture, arranging memory cells in a three-dimensional structure with multiple chips stacked vertically. This dimensional change enables higher integration density by utilizing the vertical dimension, accommodating more memory cells without increasing the planar footprint.

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

2Quantity of substance

If the number of stacks is increased to improve integration, then capacity is improved, but the number of pass transistors per stack increases

Engineering Contradiction:
ImproveintegrationVSAvoidnumber of pass transistors per stack
Core Design Contradiction:
Quantity of substanceVSDevice complexity

Solution Approach 1:

The transfer circuit comprising pass transistors is segmented and distributed to multiple chips in the stack. Instead of concentrating all pass transistors in a single chip, the invention divides them between the first chip and second chip, reducing the number of pass transistors that must be managed on each individual chip while supporting an increased total number of stacks.

Inventive Principle:
Principle #1Segmentation

3Device complexity

If pass transistors are concentrated in one chip to simplify design, then device complexity is reduced, but the size of the semiconductor device increases

Engineering Contradiction:
Improvedesign simplicityVSAvoiddevice footprint
Core Design Contradiction:
Device complexityVSArea of stationary object

Solution Approach 1:

The patent utilizes 3D stacking to vertically distribute the transfer circuit across multiple chips rather than expanding it horizontally on a single chip. This vertical distribution in the third dimension reduces the planar footprint while managing the complexity through structured segmentation across the stack.

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

Data Source

PatentUS11751387B2Semiconductor device
Publication Date: 2023.09.05 SK HYNIX INC
  • US11751387B2 patent drawing
  • US11751387B2 patent drawing
  • US11751387B2 patent drawing

AI summary

A semiconductor device having a three-dimensional (3D) structure is disclosed. The semiconductor device includes a first chip configured to include a peripheral circuit, and a second chip stacked on the first chip that is configured to include a first memory cell array and a second memory cell array. A plurality of transfer circuits are configured to connect a plurality of row lines of the first memory cell array and a plurality of row lines of the second memory cell array to respective global row lines is divided between the first chip and the second chip.