Adaptive State Estimation for Semiconductor Wafer Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
State-of-the-art semiconductor manufacturing processes face inaccuracies in process control due to equal weighting of diverse metrology data, leading to erroneous assessments and reduced quality or yield, as they fail to adequately represent data with special properties such as source, age, and credibility.
Innovation Solution
An adaptive state estimation process is implemented, where manufacturing data is weighted based on metadata including source, age, goodness-of-fit, tool identification, and credibility, allowing for differential treatment of metrology data from test and production wafers, and adjusting state estimation parameters accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If equal weighting is applied to all metrology data, then the processing is simple and uniform, but the state estimation accuracy deteriorates due to failure to represent data with special properties
Solution Approach 1:
The patent applies local quality by assigning different weights to different metrology data points based on their specific properties. Instead of uniform treatment, each data point receives a weight reflecting its credibility, source reliability, and goodness-of-fit characteristics, thereby improving state estimation accuracy while maintaining manageable processing complexity through systematic differentiation.
Solution Approach 2:
The patent changes the parameter of data weighting from uniform to variable based on metadata characteristics. By introducing weight parameters that reflect data credibility, source quality, and goodness-of-fit, the system transforms the state estimation process to prioritize high-quality data while downweighting or excluding unreliable data, thus resolving the contradiction between accuracy and complexity.
2Productivity
If all metrology data is treated uniformly, then the processing approach is consistent, but the quality and yield deteriorate due to erroneous assessments from unrepresentative data
Solution Approach 1:
The patent improves data representativeness by applying local quality assessment to each metrology data point. Through credibility scoring and goodness-of-fit evaluation, the system identifies and weights data that truly represents process conditions, thereby improving manufacturing yield by basing decisions on reliable, representative data rather than uniform treatment of all data.
3Measurement precision
If metadata analysis is performed to assess data credibility, then the state estimation accuracy improves, but the processing time and complexity increase
Solution Approach 1:
The patent applies preliminary action by assessing metadata credibility and assigning weights to metrology data points before the state estimation process. This pre-weighting approach consolidates the complexity of credibility assessment into an upfront step, allowing the subsequent state estimation to proceed efficiently with pre-determined weights, thus reducing overall processing time while maintaining high accuracy.
Data Source
AI summary
A method, apparatus, and a system for performing an adaptive state estimation process for processing semiconductor wafers. A processing of a first workpiece is controlled using a process controller and a processing tool. Manufacturing data relating to the processing of the first workpiece is acquired. Status data relating to the manufacturing data is acquired. The status data includes data relating to the source of the manufacturing data. A state of a process controller or a processing tool is determined based upon the status data and the manufacturing data.