Wafer AVI Dynamic Sampling for Throughput and Quality Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The semiconductor industry faces inefficiencies in wafer inspection and metrology due to high costs and resource utilization associated with oversampling, which slows down throughput and increases manufacturing costs, despite the need for precise quality control in IC fabrication.
Innovation Solution
A continuous dynamic sampling scheme is implemented to optimize sampling rates based on acceptable quality levels, allowing for variable sampling rates and quality thresholds across different processing stages, adjusting sampling frequencies dynamically in response to inspection results and defect detection, and adjusting processing conditions to maintain quality control while reducing unnecessary data generation and processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If continuous sampling at high rates is used for wafer inspection, then quality control precision is improved, but manufacturing throughput and cost-effectiveness deteriorate
Solution Approach 1:
The patent implements dynamic sampling rate adjustment based on process stability. When the process is stable and within control limits, the sampling rate is reduced to maintain throughput. When process variations are detected, the sampling rate increases to ensure quality control. This dynamic adaptation resolves the contradiction by making the sampling intensity variable rather than fixed.
Solution Approach 2:
The system changes the sampling parameter (sampling rate) based on process conditions. By monitoring process control charts and statistical metrics, the system adjusts the sampling fraction dynamically - using lower sampling rates during stable periods and higher rates during unstable periods, thus maintaining quality control precision while optimizing manufacturing throughput.
2Reliability
If high sampling rates are used for inspection, then defect detection capability is improved, but inspection costs and resource utilization worsen
Solution Approach 1:
The patent applies partial sampling rather than continuous full-rate sampling. By using statistical process control to determine appropriate sampling fractions, the system performs just enough inspection to maintain defect detection capability while avoiding excessive sampling that would waste resources. The sampling rate is optimized to be sufficient for quality control but not excessive.
Solution Approach 2:
The system uses feedback from process monitoring and inspection results to dynamically adjust sampling rates. When defect rates are low and process is stable, sampling rate decreases reducing inspection costs. When defects are detected or process becomes unstable, sampling rate increases to maintain detection capability. This feedback loop resolves the contradiction between detection capability and resource consumption.
3Stability of the object's composition
If fixed high sampling rates are applied across all process stages, then quality control consistency is improved, but adaptability to different process stages deteriorates
Solution Approach 1:
The patent implements different sampling rates and quality thresholds for different process stages based on their specific characteristics and risk levels. Critical process stages with higher defect risks use higher sampling rates, while stable stages use lower rates. This local adaptation maintains quality control consistency across the entire manufacturing process while being adaptable to the specific needs of each process stage.
Solution Approach 2:
The system dynamically adjusts sampling parameters for each process stage based on real-time process stability and historical performance data. Rather than applying a fixed uniform sampling rate, the system adapts the sampling intensity to match the actual quality risks and process characteristics of each stage, maintaining consistency in quality outcomes while being versatile in implementation.
Data Source
AI summary
A wafer metrology system having a continuous dynamic sampling scheme configured to optimize a sampling rate for AVI of process wafers in an IC fabrication flow based on acceptable quality levels. For a stable process, the process wafers may be sampled at a lower rate without negatively affecting quality control.