Image Sensor Separation Structure for Low Cross-Talk Pixels
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image sensors face challenges in achieving high resolution and reducing dark current and cross-talk, which affect the signal-to-noise ratio and overall image quality.
Innovation Solution
The image sensor incorporates a separation structure with a lower and upper separation pattern, where the upper surface of the lower separation pattern and/or the lower surface of the upper separation pattern has a wavy or sawtooth shape, and the vertical length of one pattern is significantly greater than the other in intersecting regions, to reduce dark current and prevent cross-talk.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If a conventional separation structure with flat surfaces and uniform depth is used, then the manufacturing process is simple, but dark current and cross-talk cannot be effectively reduced
Solution Approach 1:
The separation structure employs wavy or sawtooth-shaped surfaces instead of flat surfaces. The wavy shape creates multiple reflection interfaces that scatter and block carrier movement paths, effectively reducing dark current and cross-talk between pixel regions while maintaining manufacturability through standard semiconductor processing techniques.
Solution Approach 2:
The invention introduces vertical depth variation to the separation structure by creating intersecting regions where separation lines cross. At these intersection points, the separation structure extends deeper into the substrate (1.5-10 μm), forming a three-dimensional network that more effectively blocks carrier diffusion paths compared to a two-dimensional flat separation pattern.
2Object-affected harmful factors
If the separation structure depth is increased uniformly throughout, then cross-talk reduction improves, but manufacturing precision requirements increase significantly
Solution Approach 1:
The separation structure implements depth variation at specific locations rather than uniformly throughout. The intersecting regions where separation lines cross have increased depth (1.5-10 μm), while non-intersecting regions maintain standard depth. This localized deepening provides enhanced cross-talk blocking at critical intersection points without imposing stringent depth control requirements across the entire sensor area.
3Measurement precision
If photoelectric conversion device regions are placed closer together to increase resolution, then pixel density improves, but cross-talk between adjacent pixels increases
Solution Approach 1:
The wavy or sawtooth-shaped separation structure creates multiple internal reflection interfaces that scatter and block carrier diffusion paths. This curved geometry provides effective electrical isolation between closely-spaced pixel regions, enabling high pixel density while maintaining low cross-talk levels through physical barrier formation rather than increased spacing.
Data Source
AI summary
An image sensor includes a substrate having first and second surfaces. A separation structure penetrates the substrate. Photoelectric conversion device regions are spaced apart from each other in the substrate. Color filters are disposed on the second surface of the substrate. Microlenses are disposed on the color filters. The separation structure includes lower and upper separation patterns, first line portions that run parallel to each other, and second line portions that perpendicularly intersect the first line portions. An upper surface of the lower separation pattern or a lower surface of the upper separation pattern has a wavy or sawtooth shape. In intersecting regions in which the first line portions and the second line portions intersect, a vertical length of one of the lower separation pattern and the upper separation pattern is about 2 to 10 times greater than a vertical length of the other.