Polishing Recipe Determination via Irregularity Screening
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The determination of a polishing recipe for semiconductor devices is currently limited by the reliance on human expertise, which can lead to inaccuracies due to the limited amount of past data that can be learned by a single engineer and the inability to efficiently handle irregularities in area response data, making it difficult to achieve high accuracy and efficiency in polishing processes.
Innovation Solution
A polishing recipe determination device utilizing machine-learning models to estimate irregularities in area response data, correct and refine the data, and simulate optimal polishing recipes, enabling automated and accurate recipe determination through an irregularity-presence-or-absence estimation unit, screening unit, simulation unit, acceptance evaluation unit, and response data correction unit.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If a skilled engineer manually determines the polishing recipe by screening area response data, then high accuracy in polishing uniformity is achieved, but the process is time-consuming and limited by the engineer's capacity to learn from past data
Solution Approach 1:
The patent replaces the manual mechanical process of data screening and analysis by engineers with an automated information processing system. The system includes a data acquisition unit that collects area response data, a machine learning unit that automatically screens and analyzes the data using learned models, and a recipe determination unit that generates polishing recipes. This substitution eliminates the time constraint of manual engineering analysis while maintaining high accuracy through automated pattern recognition in the area response data.
2Productivity
If conventional programs with conditional branches are used to automate recipe determination, then processing speed is improved, but accuracy deteriorates because irregular values in area response data cannot be properly handled
Solution Approach 1:
The patent replaces conventional conditional-branch programming with a machine learning-based system. The machine learning unit is trained on historical area response data and can automatically identify and handle irregular values through pattern recognition rather than rigid conditional logic. This allows the system to process data with varying irregularities adaptively, maintaining high accuracy while achieving automated processing speed.
Solution Approach 2:
The system changes the approach from fixed conditional parameters to adaptive learned parameters. The machine learning unit learns optimal thresholds and criteria from historical data, allowing it to dynamically adjust to different patterns of irregularities in area response data. This enables accurate handling of varied irregularity types without requiring explicit conditional programming for each scenario.
3Manufacturing precision
If more past area response data is learned, then polishing recipe accuracy is improved, but the complexity of manual data processing increases beyond human capacity
Solution Approach 1:
The patent replaces human cognitive processing with an automated machine learning system that can handle large volumes of historical area response data. The system includes data acquisition, storage, and machine learning units that automatically process and learn from extensive historical datasets without the complexity burden falling on human operators. This enables the system to leverage large amounts of past data for improved accuracy while keeping the operational interface simple.
Data Source
AI summary
An information processing apparatus is an information processing apparatus that determines a polishing recipe based on area response data acquired by changing a pressure for each area in a polishing head, the apparatus including an irregularity-presence-or-absence estimation unit that estimates and outputs whether an irregularity is present using new area response data as an input, a screening unit estimates and outputs, when the irregularity-presence-or-absence estimation unit estimates that an irregularity is present, area response data after the removal of the irregularity using area response data estimated that an irregularity is present as an input, and a simulation unit that determines a polishing recipe by simulation based on area response data estimated by the irregularity-presence-or-absence estimation unit that no irregularity is present or a response for each area after the removal of the irregularity estimated by the screening unit.