Metal Gate Fill Etching with ML Feedback for Thin-Film Precision
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing etching techniques in semiconductor fabrication face challenges in ensuring the proper formation of features, leading to inconsistencies in thin-film thickness and composition, which can result in performance issues in integrated circuits.
Innovation Solution
The use of machine learning techniques to adjust thin-film etching process parameters dynamically, both between and during etching processes, to train an analysis model that determines optimal process parameters for achieving target specifications in thin-film thickness and composition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional etching techniques are used to form thin-film features, then manufacturing complexity is reduced, but manufacturing precision deteriorates due to inconsistencies in thin-film thickness and composition
Solution Approach 1:
The patent implements dynamic adjustment of etching parameters during the etching process based on real-time monitoring of thin-film thickness and composition. The system continuously modifies etch rate, gas flow rates, and power levels to maintain precise control over thin-film formation, transitioning from static conventional etching to dynamic adaptive etching that responds to actual process conditions
Solution Approach 2:
The patent employs a feedback control system that monitors thin-film thickness and composition during etching and uses this information to adjust process parameters. Spectroscopic ellipsometry and mass spectrometry provide real-time feedback on film properties, which are then fed back to the process control system to modify etching conditions and maintain target specifications
2Productivity
If feature size is decreased to increase computing power, then productivity is improved by packing more transistors, but manufacturing precision deteriorates due to difficulty in ensuring proper feature formation
Solution Approach 1:
The patent replaces conventional mechanical and thermal etching methods with plasma-based etching technology. The plasma process uses reactive species and ion bombardment to achieve precise material removal at sub-10nm scales, enabling better control over feature formation as dimensions decrease while maintaining high productivity
Solution Approach 2:
The patent utilizes changes in plasma process parameters including gas composition, pressure, power density, and temperature to optimize etching performance at reduced feature sizes. By dynamically adjusting these parameters, the system maintains precise control over etch rate and anisotropy even as feature dimensions shrink to increase transistor density
3Reliability
If conventional etching processes are used, then device complexity is reduced, but reliability deteriorates due to performance issues from improper feature formation
Solution Approach 1:
The patent implements real-time feedback monitoring during etching using spectroscopic ellipsometry to track thin-film thickness and composition. This feedback enables immediate detection and correction of process deviations that could lead to defective features, ensuring high reliability of the resulting integrated circuits
Solution Approach 2:
The patent introduces plasma as an intermediary medium between the etching equipment and the thin-film material. The plasma provides controlled chemical reactions and ion bombardment that enable precise and repeatable feature formation, improving reliability while the complexity is managed through automated process control systems
Data Source
AI summary
A semiconductor process system etches gate metals on semiconductor wafers. The semiconductor process system includes a machine learning based analysis model. The analysis model dynamically selects process conditions for an etching process. The process system then uses the selected process conditions data for the next etching process.