CVD Layer Quality Control via Raw Data Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for depositing semiconductor layers lack efficient data processing and analysis, making it difficult to identify and characterize process steps and parameters without detailed formula knowledge, which can lead to inconsistencies in layer quality and process efficiency.
Innovation Solution
A method and device that utilize raw data from sensors and actuators to determine process parameters and steps through mathematical correlation, allowing for automated identification and analysis of process steps without relying on pre-defined formula commands, and using actuation data to compress and interpret data for improved process control and quality assessment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If detailed formula knowledge is used to control the deposition process, then manufacturing precision is improved, but device complexity increases
Solution Approach 1:
The patent creates a digital twin (virtual model) that copies and simulates the physical deposition process. This digital replica processes raw sensor data to reconstruct process steps and parameters, enabling quality control without requiring complex human expertise in formula interpretation. The digital twin serves as a simplified interface between raw data and manufacturing decisions.
Solution Approach 2:
The patent introduces an intermediate data processing layer that acts as a mediator between raw sensor data and process control. This intermediary system automatically identifies process steps, determines parameters, and generates control signals without requiring direct human intervention or complex formula knowledge, thereby simplifying the overall system architecture.
2Measurement precision
If manual analysis of raw data is performed to identify process steps, then measurement precision is improved, but loss of time increases
Solution Approach 1:
The system performs self-service by automatically analyzing raw sensor data to identify process steps and parameters without requiring manual human analysis. The digital twin autonomously processes the data stream, detects process transitions, and determines deposition parameters in real-time, eliminating time-consuming manual review while maintaining high identification accuracy.
Solution Approach 2:
The patent replaces manual mechanical analysis with automated computational analysis. Instead of human experts manually examining raw data to identify process steps, the system uses automated algorithms and digital twin technology to perform the same function faster and with consistent precision, substituting human cognitive processes with computational processes.
3Reliability
If comprehensive sensor data is collected during deposition, then reliability is improved, but quantity of substance increases
Solution Approach 1:
The patent extracts only the essential information from comprehensive sensor data through the digital twin analysis. Instead of storing and processing all raw sensor data, the system extracts key process parameters and step identifiers that are sufficient for reliable process control. This extraction approach maintains reliability by capturing critical information while discarding redundant data.
Solution Approach 2:
The patent segments the continuous stream of comprehensive sensor data into discrete process steps and parameters. By dividing the data into meaningful segments (individual deposition steps, parameter transitions, process anomalies), the system reduces the overall data volume while preserving the essential information needed for reliable process monitoring and control.
Data Source
AI summary
Information about a process for depositing at least one layer on a substrate in a process chamber is obtained via a method including the step of storing actuation data and sensor values as raw data in a log file, together with their time reference. Knowledge about the quality of the deposited layer is obtained by using the raw data. For this purpose, process parameters are obtained from the raw data by means of a computing apparatus. The beginning and the end of the process steps for processing the substrate and their respective types are identified by analyzing the time curve of the process parameters. For at least some of the process steps, characteristic process step quantities corresponding to the particular type of the process steps are calculated from the measured values, and the obtained process step quantities are compared with comparison quantities associated with one or more similar process steps.