CMP Endpoint Reliability Prediction Using Wear and Process States
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional polishing devices lack the capability to determine the reliability of endpoint detection results, leading to potential issues such as insufficient or excessive polishing, which can affect the integrity of semiconductor wafers.
Innovation Solution
An information processing device that utilizes a learning model trained through machine learning to predict the reliability of the polishing endpoint detection function by inputting reliability degradation factor state information, including wear state and processing state data, to accurately determine when the chemical mechanical polishing process has reached its endpoint.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional endpoint detection is used without reliability assessment, then the polishing process can be completed, but the reliability of the endpoint detection result is unknown leading to potential insufficient or excessive polishing
Solution Approach 1:
The system performs preliminary actions by acquiring wear state information and processing state information before endpoint detection, and uses a machine learning model to predict reliability information in advance, enabling assessment of detection reliability before the polishing endpoint is reached
Solution Approach 2:
The endpoint detection system is segmented into multiple independent components: wear state information acquisition, processing state information acquisition, machine learning model processing, and reliability information output. This segmentation allows each component to be optimized independently and combined to provide comprehensive reliability assessment
2Measurement precision
If machine learning model is introduced to predict reliability, then the accuracy of reliability prediction is improved, but the device complexity increases
Solution Approach 1:
A machine learning model serves as an intermediary between the acquired wear and processing state information and the final reliability prediction. This intermediary processes the input data through learned patterns and relationships to generate accurate reliability assessments without requiring direct complex physical measurements
Solution Approach 2:
The patent replaces traditional mechanical or physical endpoint detection methods with a machine learning-based prediction system. Instead of relying solely on physical sensors and mechanical thresholds, the system uses data-driven models to predict reliability, substituting mechanical complexity with computational intelligence
Data Source
AI summary
Information processing device that includes: an information acquisition unit that acquires reliability degradation factor state information in a chemical mechanical polishing process of a substrate performed by a substrate processing device, the reliability degradation factor state information including at least one of wear state information indicating a wear state of components of the substrate processing device and processing state information indicating a processing state during polishing; and a state prediction unit that predicts reliability information of a polishing endpoint detection function for the reliability degradation factor state information by inputting the reliability degradation factor state information acquired by the information acquisition unit into a learning model that has been trained through machine learning to learn a correlation between the reliability degradation factor state information and reliability information of the polishing endpoint detection function that indicates reliability of an endpoint detection function that detects that the chemical mechanical polishing process has reached an endpoint.