Stacked Pixel Wiring Layout for Dark-Noise-Limited Vision Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Dynamic vision sensors face bottlenecks due to wiring congestion, reduced design flexibility, increased dark noise, and limited sensitivity and speed, primarily because of the 2-D architecture where detection and readout circuit configurations share the same substrate, leading to reduced occupation ratio of light-receiving elements.
Innovation Solution
The imaging device incorporates a first pixel with a photoelectric conversion region and a readout circuit that converts electric charges into a logarithmic voltage signal, featuring a specific wiring arrangement to optimize capacitance ratios and reduce noise, while a second substrate with a logic circuit processes the signal, and includes an isolation structure between pixels to enhance sensitivity and speed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a 2-D architecture is used where detection and readout circuit configurations share the same substrate, then integration is achieved, but wiring congestion occurs causing reduced design flexibility, increased dark noise, and limited sensitivity and speed
Solution Approach 1:
The patent transitions from a planar 2-D architecture to a three-dimensional stacked architecture where the photoelectric conversion unit and readout circuit unit are positioned on different substrates vertically. This dimensional change eliminates wiring congestion by routing connections through vertical vias rather than horizontal traces, thereby reducing dark noise while maintaining integration.
Solution Approach 2:
The imaging device is segmented into distinct functional units: a photoelectric conversion unit on a first substrate and a readout circuit unit on a second substrate. This segmentation separates noise-generating circuits from light-receiving elements, reducing electromagnetic interference and dark noise while allowing independent optimization of each unit.
2Device complexity
If a 2-D architecture is used where detection and readout circuit configurations share the same substrate, then integration is achieved, but wiring congestion occurs causing reduced design flexibility
Solution Approach 1:
By stacking substrates vertically, the design gains an additional degree of freedom for circuit layout and signal routing. This enables greater design flexibility as circuits can be optimized independently on each substrate layer without being constrained by planar wiring restrictions.
Solution Approach 2:
Separating detection and readout functions onto different substrates allows independent design optimization. Each unit can be designed and tuned separately, enhancing adaptability and design flexibility while maintaining system integration through controlled inter-substrate connections.
3Device complexity
If a 2-D architecture is used where detection and readout circuit configurations share the same substrate, then integration is achieved, but wiring congestion occurs causing limited sensitivity and speed
Solution Approach 1:
The vertical stacking architecture shortens signal transmission paths by eliminating long horizontal wiring traces. Signals travel directly between photoelectric conversion and readout circuits through vertical connections, reducing transmission delay and improving processing speed.
Solution Approach 2:
By placing readout circuits immediately adjacent to photoelectric conversion units in the vertical dimension, the patent minimizes signal path length. This segmentation strategy reduces signal degradation and transmission time, enhancing both sensitivity and processing speed.
4Device complexity
If a 2-D architecture is used where detection and readout circuit configurations share the same substrate, then the occupation ratio of light-receiving elements is reduced
Solution Approach 1:
Moving readout circuits to a separate substrate layer frees up the photoelectric conversion substrate, allowing light-receiving elements to occupy the entire active area without circuit encroachment. This maximizes the light-receiving element occupation ratio while maintaining full integration functionality.
Solution Approach 2:
Separating photoelectric conversion elements from readout circuits onto different substrates eliminates the need to share the same physical space. This allows light-receiving elements to be positioned at the maximum possible density on their substrate without being constrained by circuit layout requirements.
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 configuration reduces noise, improves design flexibility, and enhances sensitivity and speed by optimizing the wiring layout and capacitance ratios, addressing the limitations of traditional dynamic vision sensors.
Implementation Method 1
a first photoelectric conversion region disposed in a first substrate and that converts incident light into first electric charges
Data Source
AI summary
An imaging device includes a first pixel including a first photoelectric conversion region disposed in a first substrate and that converts incident light into first electric charges, and a first readout circuit including a first converter that converts the first electric charges into a first logarithmic voltage signal. The first converter includes a first transistor coupled to the first photoelectric conversion region and a second transistor coupled to the first transistor. The imaging device includes a wiring layer on the first substrate and includes a first level of wirings arranged in a first arrangement overlapping the first photoelectric conversion region and in a second arrangement overlapping the first and second transistors, the second arrangement being different than the first arrangement.