Semiconductor Chamber Drift Detection for Real-Time Recipe Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Semiconductor wafer processing faces challenges due to process parameter drifts caused by chamber degradation and changes in processing conditions, which require continuous tuning and correction, especially as the complexity of semiconductor devices increases and feature sizes decrease.
Innovation Solution
A processing tool architecture that includes a drift detection module utilizing machine learning and hybrid models to identify and correct process parameter drifts in real-time, featuring witness sensors and control loop sensors to monitor and adjust settings automatically, and a prediction module to anticipate and prevent drifts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If chamber components are used for extended processing, then productivity increases, but process parameter drift occurs due to erosion and degradation
Solution Approach 1:
The patent implements a feedback mechanism using witness sensors to continuously monitor process parameters and compare them against reference values. When drift is detected, the system automatically adjusts processing parameters to compensate, enabling extended chamber usage while maintaining process stability. This resolves the contradiction by providing real-time correction that allows both high productivity and reliable process control.
Solution Approach 2:
The system performs self-diagnosis and self-correction by automatically detecting process parameter drift through witness sensors and adjusting processing parameters without external intervention. This enables the chamber to maintain reliable operation over extended periods, resolving the contradiction between productivity and reliability by allowing continuous operation with automatic compensation for component degradation.
2Manufacturing precision
If extensive recipe development and manual tuning are performed, then manufacturing precision improves, but loss of time increases
Solution Approach 1:
The patent establishes reference process parameters through preliminary processing that captures optimal settings. These reference values are stored and used for automatic comparison during subsequent processing, eliminating the need for extensive manual tuning for each new production run. This resolves the contradiction by performing the necessary calibration work once in advance, then automatically maintaining precision through ongoing automated monitoring.
Solution Approach 2:
The system creates a digital copy of optimal process parameters as reference values that can be replicated across multiple processing runs. Instead of manually recreating optimal settings each time, the system copies and applies the stored reference parameters, automatically adjusting for any detected drift. This dramatically reduces the time required to achieve manufacturing precision while maintaining accuracy.
3Measurement precision
If additional witness sensors and drift detection modules are added, then measurement precision improves, but device complexity increases
Solution Approach 1:
The witness sensors are designed to monitor multiple process parameters simultaneously, and the drift detection module serves multiple functions including drift detection, reference value comparison, and automatic parameter adjustment. This multi-functionality reduces the need for separate dedicated components for each function, improving measurement precision while limiting the increase in overall system complexity.
Data Source
AI summary
Embodiments disclosed herein include a processing tool for semiconductor processing. In an embodiment, the processing tool comprises a chamber, and a plurality of witness sensors integrated with the chamber. In an embodiment, the processing tool further comprises a drift detection module. In an embodiment, data from the plurality of witness sensors is provided to the drift detection module as input data. In an embodiment, the processing tool further comprises a dashboard for displaying output data from the drift detection module.