Crystal-Oriented Image Sensor Trenches for Fewer White Pixels
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional CMOS image sensor formation processes face issues with white pixels due to the accumulation of dislocations at the upper part of trenches, leading to defects and reduced image quality.
Innovation Solution
The method involves forming an image sensor with columnar structures having specific crystal orientations and filling trenches with doped epitaxial layers, where the growth rate in the first trench exceeds that in the cross trench, preventing the formation of (111) crystal faces that cause dislocations, and using an intrinsic epitaxial layer to repair defects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional trench filling process is used, then manufacturing process is simple, but dislocations accumulate at upper part of cross trench causing white pixels and reduced image quality
Solution Approach 1:
The trench filling process is segmented into multiple stages: first filling the first trench with doped epitaxial layer, then filling the cross trench with intrinsic epitaxial layer. This segmentation prevents dislocation accumulation by avoiding the formation of (111) crystal faces in the cross trench, thereby improving image quality while maintaining reasonable process complexity
Solution Approach 2:
The first trench is filled with doped epitaxial layer before filling the cross trench. This preliminary action establishes a growth pattern that prevents dislocation accumulation in subsequent steps, effectively eliminating white pixels while keeping the overall process manageable
2Productivity
If doped epitaxial layer is formed in all trenches simultaneously, then process is efficient, but (111) crystal faces form causing dislocation accumulation
Solution Approach 1:
The simultaneous filling process is divided into sequential steps: first trench filling with doped epitaxial layer, then cross trench filling with intrinsic epitaxial layer. This segmentation prevents (111) crystal face formation and dislocation accumulation while maintaining acceptable productivity
Solution Approach 2:
Different types of epitaxial layers are used in different trenches: doped epitaxial layer in the first trench and intrinsic epitaxial layer in the cross trench. This local differentiation prevents dislocation accumulation in the cross trench while maintaining overall filling efficiency
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 approach effectively reduces white pixels by minimizing defects in the cross trench, thereby improving image sensor performance.
Implementation Method 1
forming a doped epitaxial layer in the first trench, in the second trench and in the cross trench
Implementation Method 2
a side wall of each of the plurality of columnar structures perpendicular to the first direction is a (110) crystal face, and an oblique section of each of the plurality of columnar structures perpendicular to the third direction is a (100) crystal face
Data Source
AI summary
An image sensor and a method for forming the image sensor are provided. The method includes: providing a substrate; patterning the substrate to form a plurality of columnar structures configured in an array, wherein a first trench, a second trench, and a third trench are configured between adjacent columnar structures and respectively along a first direction, a second direction, and a third direction, side walls of the columnar structures perpendicular to the first direction are (110) crystal faces, and oblique sections of the columnar structures perpendicular to the third direction are (100) crystal faces; and forming a doped epitaxial layer in the first trench, the second trench and the cross trench. Therefore, for the image sensor, an upper part of the cross trench is improved with little defects after the cross trench is full filled, which can effectively reduce white pixels and thus improve the performance of the image sensor.