Polishing End Point Detection Using Spectral Waveform Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for detecting the polishing end point in chemical mechanical polishing (CMP) processes face challenges in accurately determining the distinctive point for processing end points due to noise interference from varying device patterns and structures, leading to inaccuracies in film thickness monitoring and increased manufacturing costs.
Innovation Solution
A method and apparatus that calculate a characteristic value using a spectral waveform of reflected light, selecting wavelengths for local maximum and minimum values, and integrating these values to detect the processing end point, while adjusting the rotational speeds of the top ring and polishing table to ensure even sensor path distribution across the substrate, reducing noise interference and enhancing signal accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional optical sensing methods are used to detect polishing end point, then the polishing process can be monitored, but noise interference from varying device patterns and structures reduces measurement precision
Solution Approach 1:
The patent changes the parameter of light wavelength by selecting specific wavelengths corresponding to local maximum and minimum values in the spectral waveform. This parameter change allows the system to detect polishing end points based on characteristic value changes at these specific wavelengths, which are less affected by noise from varying device patterns and structures
Solution Approach 2:
The patent performs preliminary action by calculating spectral waveforms in advance for reference workpieces with known film thicknesses. These pre-calculated spectral waveforms are stored and used as reference data to compare against during actual polishing, enabling more accurate end point detection by reducing the impact of noise from device patterns
2Productivity
If the sensor path is not evenly distributed across the substrate, then the polishing process is faster, but film thickness uniformity deteriorates
Solution Approach 1:
The patent applies dynamics by making the rotational speeds of the top ring and polishing table adjustable and variable. By dynamically changing the rotational speeds during the polishing process, the system optimizes both productivity and film thickness uniformity. The rotational speeds are set in a specific relationship to ensure even sensor path distribution while maintaining efficient polishing
Solution Approach 2:
The patent uses feedback by continuously monitoring the characteristic values during polishing and comparing them with reference values. Based on this feedback, the system can adjust the rotational speeds of the top ring and polishing table to maintain even sensor path distribution, ensuring both high productivity and uniform film thickness
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 allows for accurate detection of the polishing end point with improved signal-to-noise ratio, ensuring uniform film thickness and reducing manufacturing costs by minimizing errors in film thickness monitoring.
Implementation Method 1
measuring an intensity of a reflected light using an optical sensor and detecting an end point of the CMP process based on the measured intensity of the reflected light
Implementation Method 2
a spectroscope unit configured to separate the reflected light received by the light-receiving unit into a plurality of light rays according to wavelengths
Data Source
AI summary
A processing end point detection method detects a timing of a processing end point (e.g., polishing stop, changing of polishing conditions) by calculating a characteristic value of a surface of a workpiece (an object of polishing) such as a substrate. This method includes producing a spectral waveform indicating a relationship between reflection intensities and wavelengths at a processing end point, with use of a reference workpiece or simulation calculation, based on the spectral waveform, selecting wavelengths of a local maximum value and a local minimum value of the reflection intensities, calculating the characteristic value with respect to a surface, to be processed, from reflection intensities at the selected wavelengths, setting a distinctive point of time variation of the characteristic value at a processing end point of the workpiece as the processing end point, and detecting the processing end point of the workpiece by detecting the distinctive point during processing of the workpiece.