Serial Feature Tracking for Polishing Endpoint Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing optical monitoring techniques during chemical mechanical polishing (CMP) struggle to determine the polishing endpoint effectively due to variations in material removal rates and initial substrate thickness, as spectral features may move across the monitored wavelength band before completion, making it difficult to track a single feature throughout the process.
Innovation Solution
The method involves serially 'stitching' together the tracking of multiple spectral features by selecting a new peak once the initially tracked feature crosses a boundary, allowing continuous monitoring and adjusting polishing parameters based on the evolving position of the new feature.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a single spectral feature is tracked throughout the polishing process, then the endpoint detection is simple, but the feature may move across the monitored wavelength band before completion, making tracking impossible
Solution Approach 1:
The spectral monitoring range is divided into multiple wavelength bands, each with its own tracked spectral feature. When the polynomial fit indicates a feature will cross a band boundary, the system proactively switches to tracking a different feature in the adjacent band, ensuring continuous reliable monitoring throughout the polishing process.
Solution Approach 2:
The system performs preliminary polynomial fitting on sequential spectra to predict future feature positions before they actually cross band boundaries. This allows the system to switch tracking features in advance, maintaining continuous reliable tracking without interruption when features move across the wavelength spectrum.
2Loss of time
If polynomial fitting is used to predict feature position, then early switching can be achieved, but computational complexity increases
Solution Approach 1:
The system performs polynomial fitting only on selected spectral features at specific intervals rather than continuously analyzing all spectral data. This partial application of the computational method reduces processing complexity while still achieving timely detection of feature position trends and enabling proactive switching before boundaries are crossed.
3Adaptability or versatility
If the monitored wavelength band is expanded to cover all possible feature positions, then all features can be tracked, but the spectral resolution and sensitivity decrease
Solution Approach 1:
Instead of using a single wide wavelength band with reduced resolution, the system segments the spectrum into multiple narrower bands, each monitored with high spectral resolution. This allows precise measurement of feature positions within each band while collectively covering the full range of possible feature movements throughout the polishing process.
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
This approach improves endpoint control and wafer-to-wafer thickness uniformity by enabling effective tracking of spectral features across a wider range of layer structures and compositions, ensuring accurate determination of the polishing endpoint.
Implementation Method 1
measuring with an in-situ spectrographic optical monitoring system a first sequence of spectra of light reflected from the substrate
Data Source
AI summary
A method of controlling polishing includes polishing a substrate having a second layer overlying a first layer, detecting exposure of the first layer with an in-situ monitoring system, receiving an identification of a selected spectral feature and a characteristic of the selected spectral feature to monitor during polishing, measuring a sequence of spectra of light from the substrate while the substrate is being polished, determining a first value for the characteristic of the feature at the time that the first in-situ monitoring technique detects exposure of the first layer, adding an offset to the first value to generate a second value, and monitoring the characteristic of the feature and halting polishing when the characteristic of the feature is determined to reach the second value.