Backside-Illuminated Avalanche Image Sensor for Low-Noise Inspection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image sensors, such as silicon CMOS and CCD sensors, face challenges in detecting low levels of UV, DUV, and VUV light at high data rates due to high noise levels and false defect detection issues, which hinder efficient inspection of photomasks and semiconductor wafers.
Innovation Solution
The development of backside-illuminated avalanche image sensors with a thin highly doped p-type layer, a pure boron layer, and epitaxial silicon structure that minimizes electron recombination and uses a negative potential to enhance quantum efficiency and detect near-IR, visible, UV, DUV, and VUV light.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If high data rates are used for high-speed inspection, then productivity is improved, but noise levels increase and detection precision deteriorates
Solution Approach 1:
The patent transitions from front-side illumination to backside illumination, fundamentally changing the dimensional approach to sensor operation. This allows photons to enter through the back surface and travel through the silicon substrate to reach the photosensitive region, avoiding surface-related noise mechanisms that dominate at high data rates. The backside-illuminated architecture enables high-speed operation while maintaining low noise levels equivalent to 1-2 electrons RMS.
Solution Approach 2:
The patent employs multiple parameter changes including: (1) changing the illumination direction from front to back, (2) adjusting the doping concentration profile with a thin highly doped p-type layer followed by a pure boron layer, (3) operating the sensor in avalanche mode with high electric fields to achieve internal gain, and (4) optimizing the epitaxial silicon layer thickness and doping levels. These parameter changes collectively enable simultaneous achievement of high data rates and low noise.
2Measurement precision
If cooling is applied to reduce noise, then detection precision is improved, but device complexity and operating cost increase
Solution Approach 1:
The sensor structure itself provides the noise reduction function that would otherwise require external cooling systems. The backside-illuminated architecture combined with the specific doping profile (thin highly doped p-type layer and pure boron layer) and avalanche multiplication mechanism inherently suppresses noise sources, allowing the sensor to achieve low noise performance at room temperature without requiring complex cryogenic cooling infrastructure.
3Ease of manufacture
If conventional silicon sensors are used for UV detection, then manufacturing is simplified, but quantum efficiency deteriorates due to electron recombination
Solution Approach 1:
The patent inverts the conventional sensor architecture by illuminating from the backside rather than the front. This inversion allows UV photons to enter through the back surface and travel through the optimized silicon structure to reach the photosensitive region, avoiding surface recombination losses that occur in conventional front-side illuminated sensors. The backside illumination approach maintains compatibility with standard silicon fabrication processes while dramatically improving quantum efficiency for UV, DUV, and VUV wavelengths.
4Productivity
If high data rates are used, then productivity is improved, but false defect detection increases
Solution Approach 1:
The patent replaces the mechanical/electronic readout system that generates noise and false detections with a physics-based solution: avalanche photodetection with internal gain. The high electric fields in the avalanche region provide signal amplification that maintains signal-to-noise ratio at high data rates, enabling reliable defect detection without false positives even during high-speed inspection operations.
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
These sensors achieve high quantum efficiency and low noise levels, enabling the detection of single photons and reducing false defect rates, thus improving the sensitivity and accuracy of inspection systems for photomasks and semiconductor wafers.
Implementation Method 1
a pure boron layer on the light-sensitive surface to minimize electron recombination
Implementation Method 2
backside-illuminated avalanche image sensors
Implementation Method 3
when a photon is absorbed in silicon, usually only a single electron-hole pair is created
Data Source
Figure 1
Figure 2(A)~2(B)
Figure 3
AI summary
A high sensitivity image sensor comprises an epitaxial layer of silicon that is intrinsic or lightly p doped (such as a doping level less than about 1013 cm-3). CMOS or CCD circuits are fabricated on the front-side of the epitaxial layer. Epitaxial p and n type layers are grown on the backside of the epitaxial layer. A pure boron layer is deposited on the n-type epitaxial layer. Some boron is driven a few nm into the n-type epitaxial layer from the backside during the boron deposition process. An anti-reflection coating may be applied to the pure boron layer. During operation of the sensor a negative bias voltage of several tens to a few hundred volts is applied to the boron layer to accelerate photo-electrons away from the backside surface and create additional electrons by an avalanche effect. Grounded p-wells protect active circuits as needed from the reversed biased epitaxial layer.