Direct Measurement Model for Semiconductor Process Calibration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Accurate simulation and calibration of semiconductor manufacturing processes are hindered by the complexity of modern integrated circuit manufacturing, which involves numerous interacting characteristics, leading to high computational demands and imprecise edge detection algorithms sensitive to noise and low-contrast features, especially at feature sizes below 65 nm.
Innovation Solution
A direct measurement model is implemented that simulates manufacturing processes without relying on edge detection algorithms, using empirical or non-empirical parameters, which are iteratively calibrated to ensure satisfactory measurement information, reducing reliance on inverse problem solutions and noise-sensitive edge extraction methods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If empirical manufacturing process models are used to simulate manufacturing processes, then computing resources required are reduced, but measurement precision and manufacturing precision deteriorate due to rough parameter estimation and low-contrast features
Solution Approach 1:
The patent performs preliminary action by simulating the manufacturing process to generate a preliminary measurement model before actual measurement. This preliminary simulation provides initial parameter estimates and measurement expectations, which guide the subsequent actual measurement process. The preliminary action includes simulating the manufacturing process using process models, extracting preliminary measurement information from simulated images, and comparing with actual measurement information to iteratively refine parameters.
Solution Approach 2:
The patent implements feedback by comparing preliminary measurement information from simulation with actual measurement information from real measurements. The differences between simulated and actual measurements are used to adjust and refine the manufacturing process parameters. This feedback loop continues iteratively, improving measurement precision while managing computational resources through targeted refinements rather than full recomputation.
2Ease of manufacture
If conventional calibration methodologies with edge detection algorithms are used, then manufacturing process calibration can be performed, but measurement precision deteriorates due to sensitivity to noise and low-contrast features
Solution Approach 1:
The patent replaces the mechanical edge detection approach with a simulation-based measurement model. Instead of using edge detection algorithms that are sensitive to noise and low-contrast features, the patent simulates the entire manufacturing process and measurement process to generate expected measurement information. This substitution of the measurement approach with simulation-based modeling eliminates the sensitivity to edge detection parameters and noise, providing more robust and precise measurements.
Solution Approach 2:
The patent creates a virtual copy of the manufacturing and measurement process through simulation. By generating simulated images and measurement information that replicate the actual process, the patent can compare simulated measurements with actual measurements to refine parameters. This copying approach allows calibration without relying on noisy edge detection, as the simulated measurements provide a clean reference for comparison and parameter adjustment.
3Manufacturing precision
If accurate semiconductor manufacturing models are built to correctly simulate manufacturing processes, then manufacturing precision improves, but device complexity and computing resources required increase enormously
Solution Approach 1:
The patent applies partial action by focusing computational resources on the most critical aspects of the manufacturing process. Rather than attempting to model every detail with maximum accuracy, the patent identifies key parameters that most influence measurement outcomes and concentrates computational effort on accurately modeling those specific aspects. This selective approach achieves sufficient manufacturing precision for measurement calibration without the enormous complexity of complete high-fidelity modeling.
Solution Approach 2:
The patent implements dynamics by making the model complexity adaptive. The simulation model starts with initial process models and measurement models, then iteratively refines parameters based on comparisons with actual measurements. The model complexity and computational effort are dynamically adjusted through iterations, increasing detail only where needed to improve measurement precision while managing overall system complexity through iterative refinement rather than static high-complexity modeling.
Data Source
AI summary
Various embodiments of the present invention are generally directed to a method, system, and computer program product for implementing direct measurement model with simulation and calibration of manufacturing process model in the manufacturing of precision devices such as electronic integrated circuits. The method and the system determine the measured measurement result and the direct measurement information and compare the direct measurement information against the other to determine whether to adjust the process models and the empirical parameters thereof.