Hyperspace Mapping for Semiconductor Process Endpoint Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing complexity of semiconductor manufacturing processes, with thousands of variables to manage, makes it difficult for operators to identify process states and control manufacturing processes effectively, especially as device geometries shrink to the nanometer scale, as simpler approaches like univariate statistical process control are inadequate for controlling multiple interrelated variables.
Innovation Solution
The method involves monitoring and mapping manufacturing process variables into hyperspace, identifying predetermined states by determining when points reach a closed volume within a predefined distance from a location in hyperspace, and using transformation methods to project these points onto a line or lower-order hyperspace for control, utilizing sensors like Fourier transform infrared spectrometers and pressure sensors to adjust process parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple manufacturing variables are monitored simultaneously, then process control capability is improved, but data complexity and analysis difficulty increase
Solution Approach 1:
The patent transforms the monitoring approach by moving from traditional univariate analysis to multivariate analysis in hyperspace. Each manufacturing variable becomes a dimension in this hyperspace, allowing simultaneous monitoring of multiple variables while managing complexity through geometric representation. The endpoint detection is achieved by defining a target region in this multidimensional space, converting a complex multivariate problem into a geometric proximity problem that is more manageable.
2Manufacturing precision
If traditional univariate statistical process control is used, then single parameter control is effective, but multiple interrelated variables cannot be controlled
Solution Approach 1:
The patent merges multiple univariate control approaches into a unified multivariate control system. By combining multiple manufacturing variables into a single hyperspace representation, the system can control interrelated variables simultaneously while maintaining the precision of individual parameter monitoring. The geometric distance calculation in hyperspace provides a unified metric that accounts for correlations between variables.
3Measurement precision
If operators manually analyze raw data traces for each variable, then detailed inspection is possible, but operator workload and time consumption increase significantly
Solution Approach 1:
The patent replaces the manual mechanical analysis process with an automated computational system. Instead of operators visually inspecting data traces, the system automatically calculates the position of the manufacturing process in hyperspace and determines when the endpoint is reached by computing distances to the target region. This substitution of automated calculation for manual inspection dramatically reduces time consumption while maintaining or improving detection accuracy.
Data Source
AI summary
A method for identifying a predetermined state of a manufacturing process by monitoring the process which involves, monitoring a plurality of variables that vary in value during the manufacturing process. The method also involves mapping as points in a hyperspace the values of each variable at a plurality of times, where the hyperspace has a number of dimensions equal to the number of variables and the number of points is equal to the plurality of times. The method also involves identifying that a manufacturing process has reached the predetermined state when one of the points reaches a closed volume in the hyperspace that is located within a predefined distance from a predefined location in the hyperspace.