Metrology Algorithm Optimization for Semiconductor Specimens
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current semiconductor manufacturing processes face challenges in optimizing metrology algorithms due to the need for manual tuning of parameters, which is time-consuming and error-prone, especially with the increasing complexity of semiconductor devices and shrinking feature dimensions, requiring tighter specifications for precision and matching.
Innovation Solution
A computerized metrology system that varies tool and algorithm parameters to generate a dataset representing expected tool variations, optimizing the metrology algorithm to meet specific metrics such as tool matching, precision, correlation, and sensitivity, thereby automating the process and improving system throughput.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual tuning of metrology algorithm parameters is performed, then expertise and precision can be applied, but the process becomes time-consuming and error-prone
Solution Approach 1:
The system performs self-optimization by automatically tuning metrology algorithm parameters using machine learning models. The system collects measurement data, trains models to identify optimal parameters, and applies them without manual intervention, enabling the system to serve itself in the parameter optimization process while maintaining high measurement precision
Solution Approach 2:
The system performs preliminary optimization by pre-training machine learning models on historical measurement data to determine optimal parameter settings before actual metrology measurements are taken. This advance preparation eliminates the need for time-consuming manual tuning during production while ensuring measurement precision requirements are met
2Reliability
If multiple tool setups are performed to optimize metrology algorithms, then comprehensive data can be collected, but manufacturing efficiency decreases
Solution Approach 1:
The machine learning model serves multiple functions: it analyzes measurement data from various tool settings, determines optimal parameters across different scenarios, and provides universal optimization guidance that applies across multiple tool configurations without requiring physical reconfiguration of equipment
Solution Approach 2:
The system creates virtual representations of tool settings and measurement outcomes through machine learning models. Instead of physically setting up multiple tools, the system uses computational models to simulate and analyze various configurations, identifying optimal parameters through data analysis rather than physical experimentation
3Measurement precision
If stringent metrology specifications are imposed, then measurement quality improves, but the complexity of optimization increases
Solution Approach 1:
The system implements feedback loops where measurement results are continuously analyzed against stringent specifications, and the machine learning model adjusts parameter recommendations based on performance outcomes. This automated feedback mechanism handles the complexity of meeting tight specifications without requiring complex manual optimization procedures
Data Source
AI summary
There is provided a metrology system and method. The method includes obtaining a set of tool parameters selected from multiple tool parameters characterizing the examination tool, varying a value of each tool parameter from the set a number of times, giving rise to a plurality of tool settings corresponding to a plurality of combinations of varying values of the set of tool parameters, configuring an examination tool with each given tool setting of the plurality of tool settings; and in response to receiving, from the examination tool, a plurality of sets of images corresponding to the plurality of tool settings and representing expected tool variations over time in a single tool or between different tools, optimizing a metrology algorithm using the plurality of sets of images so as to meet at least one metrology metric including tool matching.