Lithography Mark Detection Model Selection for Alignment Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In lithography processes, especially for semiconductor integrated circuits, alignment accuracy is compromised due to substrate distortion and measurement errors caused by film thickness variations and incomplete detection of alignment marks, particularly in peripheral regions where mark arrangement restrictions are strict, leading to incorrect correction values and reduced alignment precision.
Innovation Solution
A measuring method that detects multiple marks on a substrate, calculates an evaluation value for their distribution state, determines a model formula representing deformation components, and specifies the shape of the substrate region using this information to improve alignment accuracy by selecting appropriate correction methods based on the distribution state of detected marks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If alignment is performed using conventional methods with fixed mark detection, then the alignment process is simple, but alignment accuracy deteriorates due to substrate distortion and non-uniform mark distribution
Solution Approach 1:
The alignment method dynamically adapts to the actual mark distribution state by calculating an evaluation value and selecting different model formulas based on whether marks are uniformly distributed. This dynamic selection allows the system to optimize alignment accuracy for each specific case rather than using a fixed approach, resolving the contradiction between maintaining simplicity and achieving high accuracy under varying conditions.
Solution Approach 2:
The system changes the parameter of model formula selection based on the evaluation of mark distribution. By calculating an evaluation value from detected mark positions and comparing it against thresholds, the system switches between different alignment models (first model for non-uniform distribution, second model for uniform distribution), thereby adapting to different substrate conditions and achieving accurate alignment despite substrate distortion.
2Reliability
If marks are strictly arranged in peripheral shot regions, then mark arrangement restrictions are satisfied, but mark distribution becomes non-uniform leading to incorrect correction values
Solution Approach 1:
The system performs preliminary detection of mark positions and calculates an evaluation value before performing alignment. This preliminary assessment of mark distribution uniformity allows the system to anticipate the need for different model formulas, ensuring that the correct alignment method is selected in advance, thereby preventing incorrect correction values even when marks are non-uniformly distributed due to peripheral region restrictions.
Solution Approach 2:
The system uses feedback from the evaluated mark distribution state to select the appropriate model formula. By continuously monitoring whether marks are uniformly distributed and adjusting the alignment approach accordingly, the system compensates for the non-uniform distribution caused by strict mark arrangement restrictions in peripheral regions, maintaining both detection reliability and alignment precision.
3Ease of manufacture
If a single alignment method is used for all pattern regions, then the process is simple, but alignment accuracy deteriorates due to varying substrate distortion and mark distribution
Solution Approach 1:
The alignment process dynamically adapts to different pattern regions by evaluating mark distribution and selecting appropriate model formulas. This dynamic approach allows the system to maintain simplicity in the overall process flow while achieving high accuracy in each specific region, resolving the contradiction between process simplicity and measurement precision.
Solution Approach 2:
The system applies different alignment methods (different model formulas) to different regions based on their specific characteristics. By locally adapting the alignment approach to match the actual mark distribution in each region, the system achieves high alignment accuracy throughout the substrate without requiring a completely complex multi-step process for every region.
Data Source
AI summary
A measuring method is provided. The method includes detecting a plurality of marks including at least three marks existing on a substrate using a scope configured to capture images of the marks, calculating an evaluation value indicating a distribution state of positions of the plurality of detected marks, determining a type of a model formula representing a deformation component in a predetermined region of the substrate based on the calculated evaluation value, and specifying a shape of the predetermined region using the model formula of the determined type.