SRAM Cell Layout With GAA Transistors for Read Stability

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing SRAM technologies face challenges in achieving high-speed, low-power operation and stability as device densities shrink with scaling IC technology nodes, leading to tradeoffs between SRAM speed and power consumption, and limited design flexibility.

Innovation Solution

The implementation of gate-all-around (GAA) transistors with optimized layouts, where pull-down transistors and pass-gate transistors have a channel width ratio and adjusted gate lengths to enhance the β ratio, improving SRAM performance by increasing the on-current of pull-down transistors relative to pass-gate transistors.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If device density is increased through scaling, then functional density increases, but manufacturing complexity increases

Engineering Contradiction:
Improvefunctional densityVSAvoidmanufacturing complexity
Core Design Contradiction:
Quantity of substanceVSDevice complexity

Solution Approach 1:

The patent implements gate-all-around (GAA) transistors that extend vertically into the third dimension, allowing multiple gates to surround the channel from all directions. This vertical stacking approach increases functional density without proportionally increasing lateral manufacturing complexity, as the gates are formed through sequential deposition and etching processes rather than requiring proportionally more lithography steps.

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

2Area of moving object

If SRAM cell size is reduced, then chip footprint decreases, but read stability deteriorates

Engineering Contradiction:
Improvechip footprintVSAvoidread stability
Core Design Contradiction:
Area of moving objectVSReliability

Solution Approach 1:

The GAA transistor structure utilizes vertical channel stacking to maintain adequate channel volume and control resistance despite reduced lateral cell dimensions. The multi-layer gate structure provides enhanced electrostatic control over the channel, compensating for the reduced planar area and maintaining read stability in scaled-down SRAM cells.

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

Solution Approach 2:

The patent optimizes the β ratio by adjusting the channel width ratio between pull-down and pass-gate transistors, and by tuning gate lengths. These parameter changes enhance the on-current of pull-down transistors relative to pass-gate transistors, improving read stability even as cell area is reduced through scaling.

Inventive Principle:
Principle #35Parameter changes

3Quantity of substance

If transistor scaling continues, then functional density increases, but design flexibility decreases

Engineering Contradiction:
Improvefunctional densityVSAvoiddesign flexibility
Core Design Contradiction:
Quantity of substanceVSAdaptability or versatility

Solution Approach 1:

The GAA transistor structure segments the gate into multiple independent gates that can be controlled separately. This segmentation allows for differentiated control of pull-down and pass-gate transistors within the same SRAM cell, enabling optimized β ratios and improved read stability while maintaining scalability. The segmented gate structure provides additional design knobs for tuning performance without increasing lateral footprint.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS20260075789A1Semiconductor structure
Publication Date: 2026.03.12 TAIWAN SEMICONDUCTOR MANUFACTURING CO LTD
  • US20260075789A1 patent drawing
  • US20260075789A1 patent drawing
  • US20260075789A1 patent drawing

AI summary

A memory semiconductor includes a first memory cell including a first pull-up (PU) transistor and a first pull-down (PD) transistor sharing a first gate structure, a second memory cell including a second PU transistor and a second PD transistor sharing a second gate structure, a first active area shared by the first PD transistor and the second PD transistor, and a second active area shared by the first PU transistor and the second PU transistor. The first gate structure has a first portion for the first PU transistor and a second portion for the first PD transistor. The second gate structure has a third portion for the second PU transistor and a fourth portion for the second PD transistor. A distance between the second portion and the fourth portion is greater than a distance between the first portion and the third portion.