Spectral Library for CMP Endpoint Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Chemical mechanical polishing (CMP) processes face challenges in determining the polishing endpoint accurately and efficiently, leading to variations in substrate layer thickness due to motion-related inaccuracies and noise in in-situ optical monitoring systems.
Innovation Solution
A method involving the creation of a library from a reference substrate, where spectra are measured at multiple points and associated with well-defined substrate characteristics, allowing for in-situ monitoring of product substrates to determine polishing endpoints with greater speed and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If in-situ optical monitoring is used to determine polishing endpoint, then real-time monitoring capability is improved, but measurement precision deteriorates due to motion-related inaccuracies and noise
Solution Approach 1:
The patent applies preliminary action by creating a library of spectra from a reference substrate before processing product substrates. The library contains spectra measured at multiple well-defined locations with known substrate characteristics, establishing a reference framework in advance that enables accurate real-time monitoring without being affected by motion-related inaccuracies during actual polishing operations
Solution Approach 2:
The patent uses copying by creating spectral copies from a reference substrate and storing them in a library. These spectral copies represent known substrate characteristics at various locations and can be compared against spectra from product substrates during polishing, allowing accurate endpoint detection through pattern recognition rather than direct measurement under motion conditions
2Measurement precision
If in-line metrology station is used for thickness measurement, then measurement precision is improved, but productivity deteriorates due to substrate removal and transportation time
Solution Approach 1:
The patent replaces the mechanical system of substrate removal and physical transportation to metrology stations with an optical-based spectral analysis system. By substituting mechanical handling with optical measurement and library comparison, the system achieves high-precision thickness measurement while maintaining continuous substrate processing flow, thereby eliminating throughput bottlenecks
Solution Approach 2:
The system creates spectral copies that can be analyzed computationally instead of requiring physical substrate handling. The library of reference spectra allows virtual comparison and measurement without mechanical substrate manipulation, maintaining both precision and productivity
3Productivity
If in-situ optical sensor is used during polishing, then productivity is improved through continuous processing, but measurement precision deteriorates due to polishing environment noise and substrate motion
Solution Approach 1:
The patent performs preliminary spectral measurements on a reference substrate under controlled conditions before polishing operations begin. This pre-established library of spectra captures the optical characteristics without polishing environment interference, providing a stable reference for comparison during continuous processing despite noise and motion during actual polishing
Solution Approach 2:
By creating spectral copies from a reference substrate and storing them in a library, the system enables comparison-based measurement during polishing. The measured spectra from product substrates are compared against the pre-recorded library spectra, allowing accurate determination of substrate characteristics even in the noisy polishing environment through pattern matching rather than direct absolute measurement
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 enables precise and rapid determination of substrate characteristics, reducing variations in post-polishing thickness and increasing throughput by allowing for real-time, in-situ monitoring without the need for precise alignment of measurement tools.
Implementation Method 1
measuring spectra at plurality of measurement points of the reference substrate
Implementation Method 2
measuring spectra at plurality of measurement points of the reference substrate
Data Source
AI summary
A method of generating a library from a reference substrate for use in processing product wafers is described. The method includes measuring substrate characteristics at a plurality of well-defined points of a reference substrate, measuring spectra at plurality of measurement points of the reference substrate, there being more measurement points than well-defined points, and associating measured spectra with measured substrate characteristics.


