SEM FOV Fingerprint Error Decomposition for Lithography Placement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The lithography process in semiconductor manufacturing faces challenges in distinguishing between real errors and SEM distortions, leading to variability in the formation and placement of small structures on substrates, which can result in failed chips due to misalignment or misplacement of features like contact holes.
Innovation Solution
A method is developed to analyze SEM images to separate real errors from SEM distortions, allowing for the adjustment of the SEM to reduce apparent errors and modify the lithography process to correct real errors, using techniques such as error decomposition, edge detection algorithms, and data processing to generate commands for improving the lithography process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If SEM distortion is not separated from real errors, then measurement precision deteriorates, but if separation methods are applied, then device complexity increases
Solution Approach 1:
The patent segments the total measurement error into two distinct components: SEM distortion error and real structural error. This is achieved by capturing multiple images at different positions and decomposing the error into a first error component (SEM distortion) and a second error component (real error), allowing each to be analyzed and corrected independently to improve measurement precision.
Solution Approach 2:
The patent introduces an intermediary data processing system that acts as a mediator between the SEM imaging system and the final measurement results. This intermediary system performs error decomposition, separates distortion from real errors, and provides corrected measurements, thereby improving precision without requiring physical modification of the SEM hardware.
2Manufacturing precision
If multiple images are captured and processed to separate errors, then manufacturing precision improves, but loss of time increases
Solution Approach 1:
The patent performs preliminary error characterization by capturing multiple images at different positions before final measurement. The error decomposition and separation of SEM distortion from real errors is done in advance, creating a corrected measurement model that can be applied to subsequent measurements, thereby improving manufacturing precision while managing time investment through upfront analysis.
Solution Approach 2:
The patent captures more images than the minimum single image, using multiple images at different positions to perform error decomposition. This excessive action (capturing additional images) enables accurate separation of error components, improving manufacturing precision. The time cost is managed by processing these images efficiently through automated error decomposition algorithms.
3Reliability
If SEM distortion is corrected without identifying real errors, then reliability improves, but loss of information occurs
Solution Approach 1:
The patent extracts and isolates the SEM distortion component from the total measurement error through error decomposition. By separating the first error component (SEM distortion) from the second error component (real structural error), the system can correct for distortion while preserving and identifying real error information, thereby improving reliability without losing critical measurement data.
Solution Approach 2:
The patent implements a feedback mechanism where the decomposed error components are analyzed and used to generate commands that modify the lithography process. The separated real error information provides feedback about actual structural deviations, allowing corrective actions to be taken while the SEM distortion feedback enables ongoing calibration and correction of the imaging system itself.
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 reduces variability in the lithography process by accurately identifying and correcting real errors, preventing false corrections that could introduce new errors and improving the precision and efficiency of the manufacturing process.
Implementation Method 1
determining, based on an image (or images) obtained based on a scan of the substrate by a scanning electron microscope (SEM)
Data Source
AI summary
A method of reducing variability of an error associated with a structure on a substrate in a lithography process is disclosed. The method includes determining, based on one or more images obtained based on a scan of the substrate by a scanning electron microscope (SEM), a first error due to a SEM distortion in the image. The method also includes determining, based on the image, a second error associated with a real error of the structure, where the error associated with the structure includes the first error and the second error. A command is generated by a data processor that enables a modification of the lithography process and an associated reduction of the variability of the error based on reducing any of the first error or the second error.