Automatic Reference Spectra Library Generation for CMP Endpoint Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Chemical mechanical polishing (CMP) processes face challenges in achieving consistent material removal rates due to variations in substrate thickness, slurry composition, polishing pad conditions, and load, making it difficult to determine the polishing endpoint and achieve a desired profile.
Innovation Solution
A computer-implemented method generates reference spectra during CMP using in-situ optical monitoring, determining a best matching reference spectrum and adjusting polishing parameters to ensure consistent endpoint detection across substrates, allowing for automatic generation of new reference libraries when existing ones do not provide a good fit.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing optical monitoring techniques are used, then endpoint detection is possible, but reliability and accuracy are insufficient for increasing demands
Solution Approach 1:
The system dynamically adapts the reference spectrum based on real-time measurements from the substrate being polished. Instead of using a static reference spectrum, the system continuously updates the reference spectrum to match the actual substrate conditions, thereby improving reliability while maintaining adaptability across different substrate types and conditions
Solution Approach 2:
The system implements feedback by measuring the actual substrate spectrum during polishing and using this information to update the reference spectrum. This closed-loop approach ensures that the reference spectrum accurately reflects current substrate conditions, enhancing both reliability and adaptability of endpoint detection
2Manufacturing precision
If multiple preset algorithms are used for different device/mask patterns, then endpoint detection can be tailored, but time required to begin polishing new substrates increases
Solution Approach 1:
The system performs self-service by automatically generating and updating its own reference spectrum based on real-time substrate measurements. This eliminates the need for manual configuration of different preset algorithms for various device/mask patterns, allowing the system to adapt autonomously and reduce setup time while maintaining precision
Solution Approach 2:
The system performs preliminary action by proactively updating the reference spectrum before actual polishing begins. By pre-adapting the reference spectrum to the specific substrate being processed, the system eliminates the need for time-consuming selection and configuration of preset algorithms, thereby reducing startup time while ensuring precision
3Ease of operation
If constant polishing pressure is applied, then process simplicity is maintained, but within-wafer non-uniformity increases due to variations in material removal rate
Solution Approach 1:
The system dynamically adjusts polishing parameters in real-time based on spectral measurements of material removal rate variations across the wafer. This enables the system to maintain simple operation while achieving high precision by automatically compensating for within-wafer non-uniformity through adaptive pressure or speed adjustments
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 the reliability and accuracy of endpoint determination, reduces within-wafer non-uniformity, and shortens the time required to begin polishing new substrates, eliminating the need for different preset algorithms for each device/mask pattern.
Implementation Method 1
measuring a sequence of spectra from the first substrate during polishing with an in-situ optical monitoring system
Data Source
AI summary
A method of generating reference spectra includes polishing a first substrate in a polishing apparatus, measuring a sequence of spectra from the first substrate during polishing with an in-situ optical monitoring system, for each spectrum in the sequence of spectra, determining a best matching reference spectrum from a first plurality of first reference spectra to generate a sequence of reference spectra, calculating a value of a metric of fit of the sequence of spectra to the sequence of reference spectra, comparing the value of the metric of fit to a threshold value and determining whether to generate a second library based on the comparison, and if the second library is determined to be generated, storing the sequence of spectra as a second plurality of reference spectra.