Stacked Pixel Structure With Isolated Floating Diffusion for HDR

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing CMOS image sensors face challenges in achieving high resolution with high dynamic range due to limitations in pixel density and conversion gain, particularly from increased capacitance from wiring connections to secondary dies.

Innovation Solution

A two- or three-device layer system is implemented, where the floating diffusion node is isolated from wiring to the second die in high conversion gain mode, with DCG transistor and source follower on the first die, and remaining photodetector circuit components on the second die, allowing for high pixel density and dynamic range through shared floating diffusion regions and reduced capacitance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If photodetector circuit components are integrated on a second die to increase pixel density, then resolution is improved, but capacitance from wiring connections increases and reduces conversion gain

Engineering Contradiction:
ImproveresolutionVSAvoidconversion gain
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The photodetector circuit is segmented into two parts: critical components (photosensitive area, transfer gate, floating diffusion node, source follower, and DCG transistor) remain on the first die, while non-critical components (row select transistor, reset transistor) are moved to the second die. This segmentation minimizes wiring connections and associated capacitance, preserving conversion gain while still achieving increased pixel density through the stacked architecture.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The DCG transistor and source follower are extracted and kept on the first die, isolated from the second die wiring. This extraction eliminates the capacitance burden that would result from connecting these high-impedance nodes to the second die, thereby maintaining high conversion gain in high conversion gain mode while allowing the second die to be used for increasing pixel density.

Inventive Principle:
Principle #2Taking out (Extraction)

2Adaptability or versatility

If floating diffusion node is connected to second die wiring, then circuit functionality is improved, but noise increases and dynamic range is reduced

Engineering Contradiction:
Improvecircuit functionalityVSAvoiddynamic range
Core Design Contradiction:
Adaptability or versatilityVSLoss of information

Solution Approach 1:

The system dynamically switches between two operational modes: high conversion gain mode for low-light conditions where the floating diffusion node remains isolated from second die wiring to minimize noise and maximize dynamic range, and low conversion gain mode for bright-light conditions where the node can be connected to second die components. This dynamic adaptation allows the system to optimize performance for different lighting conditions while maintaining full circuit functionality.

Inventive Principle:
Principle #15Dynamics

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 structure enables a large difference in conversion gain modes, increasing dynamic range and reducing noise, while maintaining high pixel density and resolution.

Implementation Method 1

A photodetector includes a photodiode in a first die

Methodology Applied
Scientific EffectPhotoelectric effect: Photoelectric Effect

Data Source

PatentUS20250275282A1Two layer pixel structure for high resolution with high dynamic range
Publication Date: 2025.08.28 TAIWAN SEMICONDUCTOR MANUFACTURING CO LTD
  • US20250275282A1 patent drawing
  • US20250275282A1 patent drawing
  • US20250275282A1 patent drawing

AI summary

An image sensor achieves high pixel density, and therefore high resolution, by offloading portions of a photodetector circuit to a separate device layer from the photodiodes. The photodetector uses a lateral overflow integration capacitor and a dual conversion gain transistor to increase dynamic range. The dynamic range is further increased by providing a high conversion gain mode in which the floating diffusion node is isolated from the second device layer and from the wiring that extends to the second device layer. This is accomplished by disposing the DCG transistor and the source follower in the first device layer which has the photodiodes, the transfer gates, and the floating diffusion regions. Isolating the floating diffusion node from the wiring to the second device layer in the high conversion gain mode reduces the capacitance of the floating diffusion node in the high conversion gain mode, and so increases the dynamic range.