CMP Vibration Monitoring for Real-Time Wafer Scratch Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Chemical mechanical polishing (CMP) processes in semiconductor manufacturing introduce mechanical defects like scratches on wafers, and current methods of monitoring these processes require stopping the process for optical inspection, leading to substantial downtime and reduced yield.
Innovation Solution
Implementing vibration sensors to monitor CMP processes in real-time, analyzing vibration data to detect anomalies such as micro-scratches, and using a signal processor to determine the severity of these defects, allowing for immediate process control without stopping the operation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If periodic optical observations are used to monitor wafers during CMP processes, then defect detection capability is improved, but process downtime increases and yield decreases
Solution Approach 1:
The patent replaces optical observation systems with acoustic emission sensing. Instead of using optical sensors to visually inspect wafers during CMP, the system uses acoustic sensors to detect sound waves generated by scratching events. This substitution eliminates the need to stop the CMP process for optical inspection, as acoustic monitoring can occur in real-time during operation, thus maintaining productivity while achieving defect detection.
Solution Approach 2:
The patent introduces acoustic emission signals as an intermediary indicator of wafer scratching. Rather than directly observing the wafer surface optically, the system detects acoustic waves that are generated when scratches occur. These acoustic signals serve as a mediator that provides indirect but real-time information about scratching events, enabling continuous monitoring without process interruption.
2Measurement precision
If periodic optical observations are implemented for real-time monitoring, then measurement accuracy is improved, but process continuity deteriorates due to substantial downtime
Solution Approach 1:
The patent replaces the mechanical/optical inspection system with an acoustic emission detection system. Optical observation requires stopping the CMP process to capture clear images, while acoustic sensing can detect scratching events continuously during operation. The acoustic sensors monitor sound waves in real-time, eliminating the need for process interruption and reducing time loss while maintaining detection accuracy.
Solution Approach 2:
The patent enables continuous monitoring of scratching events throughout the entire CMP process. Instead of periodic interruptions for optical inspection, the acoustic emission system provides uninterrupted detection of scratches from start to finish of the polishing operation. This continuous action ensures no scratching events are missed while maintaining process continuity and minimizing time loss.
3Productivity
If vibration sensors are used for continuous monitoring, then productivity is improved by reducing downtime, but device complexity increases due to additional sensing and signal processing components
Solution Approach 1:
The patent extracts and isolates the acoustic emission signals from the complex CMP process environment. By using band-pass filters tuned to specific frequency ranges where scratching events generate acoustic waves, the system separates the relevant defect detection signals from background noise and other process signals. This extraction approach simplifies the monitoring system by focusing only on the critical frequency bands, reducing overall system complexity while maintaining high productivity.
Solution Approach 2:
The patent changes the monitoring parameter from optical intensity to acoustic frequency. Instead of measuring light reflection or absorption properties of the wafer, the system measures acoustic wave frequencies generated by scratching events. This parameter change simplifies the sensing mechanism, as acoustic sensors and signal processing are more straightforward to implement continuously during CMP operation, thereby improving productivity without excessive complexity increase.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables continuous monitoring of CMP processes, reducing downtime and improving yield by detecting and responding to defects in real-time, thus optimizing manufacturing efficiency and reducing costs.
Implementation Method 1
a first and a second vibration sensors configured to respectively collect first and second vibration data corresponding to the CMP process
Implementation Method 2
analyzing vibration data to detect anomalies such as micro-scratches
Data Source
AI summary
An apparatus for monitoring a CMP process on a wafer includes vibration sensors to collect vibration data corresponding to the CMP process and to transmit electric signals, a signal processor to obtain digital signals by converting the electric signals into a frequency domain, and filters to filter out noise signals from the digital signals to obtain noise reduced digital signals. The signal processor obtains one or more frequency spectrums from the noise reduced digital signals, and determines a micro-scratch occurrence on the wafer by analyzing the obtained one or more frequency spectrums. The vibration sensors are in rigid contact with at least a tool such as a head holding a carrier of the wafer or a platen holding a polishing pad. Each vibration sensor includes at least two sub-frequency-ranges respectively corresponding to at least two materials to be polished by the polishing pad.