Etch Pattern Sampling for Reliable Semiconductor Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing etch-modeling systems for semiconductor devices face challenges in accurately modeling etching processes without requiring the costly and inefficient creation of new masks for pattern formation, especially when trying to replicate the surroundings of actual patterns to be etched.
Innovation Solution
An etch-modeling system that increases reliability and efficiency by sampling actual patterns and performing etch modeling based on these samples, eliminating the need for new mask fabrication and allowing for more accurate representation of etching environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If new masks are fabricated for modeling purposes, then etching model reliability is improved, but manufacturing cost and time increase
Solution Approach 1:
The patent uses existing actual patterns from layout data as copies for modeling instead of creating new physical masks. The system extracts pattern information from existing semiconductor device layouts and uses these digital copies to generate etching models, eliminating the need for physical mask fabrication while maintaining modeling accuracy.
Solution Approach 2:
The patent replaces the mechanical/physical mask fabrication process with a computational system. Instead of physically creating masks through manufacturing steps, the system uses computer-based pattern extraction, density calculation, and simulation to generate etching models directly from layout data.
2Reliability
If new masks are fabricated for modeling purposes, then etching model reliability is improved, but manufacturing cost increases
Solution Approach 1:
The patent uses existing actual patterns from layout data as copies for modeling instead of creating new physical masks. The system extracts pattern information from existing semiconductor device layouts and uses these digital copies to generate etching models, eliminating the need for physical mask fabrication while maintaining modeling accuracy.
Solution Approach 2:
The patent replaces the mechanical/physical mask fabrication process with a computational system. Instead of physically creating masks through manufacturing steps, the system uses computer-based pattern extraction, density calculation, and simulation to generate etching models directly from layout data.
3Measurement precision
If all unique patterns are used for etch modeling, then model accuracy is improved, but processing complexity increases
Solution Approach 1:
The patent applies partial action by selecting a representative subset of unique patterns for modeling rather than processing all possible patterns. The system calculates density values and populations to identify and select the most significant patterns that adequately represent the overall etching conditions, reducing processing complexity while maintaining sufficient modeling accuracy.
Solution Approach 2:
The patent introduces density value and population parameters to characterize and prioritize different unique patterns. By calculating these parameters, the system can transform the complex set of all unique patterns into a manageable subset based on density thresholds and population weights, simplifying the modeling process.
4Productivity
If sampling is performed on patterns, then processing efficiency is improved, but modeling comprehensiveness may be reduced
Solution Approach 1:
The patent introduces density value and population parameters to characterize and prioritize different unique patterns. By calculating these parameters, the system can transform the complex set of all unique patterns into a manageable subset based on density thresholds and population weights, simplifying the modeling process.
Solution Approach 2:
The system uses population calculations as a feedback mechanism to ensure that sampled patterns represent the overall pattern distribution. By weighting the selection of unique patterns based on their population frequencies, the system provides feedback that maintains modeling comprehensiveness even when processing only a subset of patterns.
Data Source
AI summary
Provided is a method of manufacturing a semiconductor device. the method comprises receiving layout data including a plurality of pieces of pattern data, the plurality of pieces of pattern data having through first to Nth unique patterns (N is a natural number greater than or equal to two), calculating first to Nth density values of the first to Nth unique patterns from the layout data and calculating first to Nth populations of the first to Nth unique patterns from the layout data, performing sampling by selecting some unique patterns among the first to Nth unique patterns, the selecting based on the first to Nth density values and the first to Nth populations, and performing etch modeling on sampled patterns of the plurality of pieces of pattern data, the sampled patterns corresponding to the selected unique patterns.