CMOS Image Sensor Trench Isolation for Phase Detection Auto-Focus
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current isolation methods for adjacent photodiodes in CMOS image sensors, such as implant isolation, are insufficient to eliminate cross-talk effects, compromising the sensitivity of phase detection auto-focus (PDAF) CMOS image sensors.
Innovation Solution
The implementation of shallow trench isolation (STI) and deep trench isolation (DTI) structures in the substrate, along with a shielding grid layer, to separate and reduce cross-talk between image sensing units and phase detection units, enhancing sensitivity by minimizing interference.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If implant isolation is used to separate adjacent photodiodes, then manufacturing is simplified, but cross-talk effect between photodiodes is not sufficiently eliminated, compromising sensitivity
Solution Approach 1:
The isolation structure is divided into two segments: shallow trench isolation (STI) extending from the first surface to a first depth, and deep trench isolation (DTI) extending to a second depth greater than the first depth. This segmented approach allows the STI to provide manufacturing simplicity while the DTI segment provides enhanced cross-talk elimination, resolving the contradiction between ease of manufacture and cross-talk prevention.
Solution Approach 2:
The isolation structure transitions from a single-depth approach to a multi-depth approach by introducing deep trench isolation that extends deeper into the substrate than the shallow trench isolation. This dimensional change in the depth direction enables superior cross-talk elimination while maintaining the manufacturing advantages of the shallow trench portion.
2Device complexity
If shallower trench isolation is used, then manufacturing complexity is reduced, but cross-talk elimination between adjacent photodiodes is insufficient
Solution Approach 1:
The isolation structure is segmented into two functional parts: the shallow trench isolation portion that can be formed using standard manufacturing processes (maintaining low complexity), and the deep trench isolation portion that provides enhanced cross-talk elimination. This segmentation allows each part to optimize for its specific function while working together as a unified structure.
Solution Approach 2:
The shallow trench isolation and deep trench isolation are nested within the same isolation region, with the STI forming an upper portion and the DTI forming a deeper portion. This nested configuration allows the simpler STI structure to be combined with the more effective DTI structure, achieving high cross-talk elimination without proportionally increasing manufacturing complexity.
3Reliability
If deep trench isolation is implemented, then cross-talk elimination is improved, but manufacturing complexity increases
Solution Approach 1:
The isolation system is segmented so that the deep trench isolation handles the critical cross-talk elimination function while the shallow trench isolation handles the bulk of the isolation and manufacturing-friendly functions. This segmentation allows the DTI to be optimized for performance without requiring the entire isolation structure to be complex.
Solution Approach 2:
The deep trench isolation extends beyond the depth of the shallow trench isolation only where needed for cross-talk elimination, rather than requiring the entire isolation structure to be deep. This partial action approach applies the more complex DTI only to the extent necessary to solve the cross-talk problem, minimizing overall manufacturing complexity.
Data Source
AI summary
An image sensor including a substrate, a trench isolation, a plurality of image sensing units, at least one phase detection unit, and an interconnection layer is provided. The trench isolation is in the substrate, and a plurality of active areas of the substrate are separated from each other by the trench isolation. The image sensing units and the at least one phase detection unit are in the active areas arranged in an array, and a sensing area of the at least one phase detection unit is smaller than a sensing area of each of the image sensing units. The interconnection layer is disposed on the image sensing units and the at least one phase detection unit. In addition, a method of fabricating an image sensor is also provided.