Semiconductor Recipe Simulation Using Similarity Search and Defect Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current semiconductor design tools lack precision in simulating semiconductor device manufacturing processes, leading to unintended electrical characteristics due to high integration and miniaturization, and do not provide accurate product specifications.
Innovation Solution
An automated simulation method using a database to predict defect probabilities and generate target script sets for simulating semiconductor manufacturing processes, determining the suitability of recipe sets through deep learning and similarity analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If current software tools are used to simulate semiconductor devices, then simulation capability is provided, but precision of product specifications is insufficient
Solution Approach 1:
The system performs preliminary actions by automatically generating simulation decks before actual manufacturing, using database searches to find similar historical recipes and pre-configuring simulation parameters. This preliminary simulation and validation process identifies potential issues before production, improving both specification precision and electrical characteristic accuracy.
Solution Approach 2:
The system creates copies of successful historical manufacturing recipes from the database and uses them as templates for new simulations. By copying and adapting proven process parameters and conditions, the system achieves higher precision in product specifications and more reliable electrical characteristics without starting from scratch.
2Productivity
If manual simulation processes are used, then flexibility is maintained, but automation and efficiency are reduced
Solution Approach 1:
The system performs self-service by automatically searching the database for similar recipes, generating simulation decks, and executing simulations without requiring manual intervention for each step. The automated workflow manages the entire simulation process from recipe selection to result analysis, significantly improving productivity while maintaining high extent of automation.
Solution Approach 2:
The system automatically adjusts simulation parameters based on database search results and historical data, changing parameters such as process conditions, material properties, and device geometries dynamically. This automated parameter optimization improves both simulation efficiency and accuracy without manual tuning.
3Adaptability or versatility
If high integration and miniaturization are implemented, then device functionality is improved, but unintended electrical characteristics increase
Solution Approach 1:
The system applies preliminary anti-action by simulating high-integration and miniaturized device structures before manufacturing, identifying potential unintended electrical characteristics such as parasitic effects, crosstalk, and manufacturing variations. The simulation results guide design modifications to prevent these harmful effects before they manifest in actual devices.
Solution Approach 2:
The system replaces physical prototyping and trial manufacturing with automated TCAD simulations to analyze electrical characteristics. By substituting computational modeling for physical experimentation, the system can efficiently evaluate high-integration designs and identify unintended electrical effects without the cost and time of repeated manufacturing cycles.
4Measurement precision
If database search and deep learning are performed, then defect prediction accuracy is improved, but computational time increases
Solution Approach 1:
The system performs preliminary actions by pre-processing and storing manufacturing data in the database with extracted features and patterns before actual defect prediction is needed. When a new case requires analysis, the system quickly retrieves pre-processed data and applies deep learning models, significantly reducing computational time while maintaining high defect prediction accuracy.
Solution Approach 2:
The system applies partial action by using database search to find the most similar historical cases first, then applying deep learning only to the most critical or uncertain predictions. This selective approach maintains high defect prediction accuracy for important cases while reducing overall computational time by avoiding exhaustive analysis of all possible scenarios.
Data Source
AI summary
A method for determining suitability of a target receipe set for manufacturing a semiconductor device includes: obtaining a reference recipe set by searching a database based on the target recipe set, the reference recipe set has a similarity with a threshold to the target recipe set; performing deep learning based on the database, the target recipe set and the reference recipe set to predict a probability of defect occurring in the semiconductor device when manufactured using a manufacturing process based on the target recipe set; generating a target script set corresponding to the target recipe set by comparing the target recipe set with the reference recipe set; simulating the manufacturing process of the semiconductor device using the target script set; and determining the suitability of the target recipe set based on the probability of the defect and a result of the simulating of the manufacturing process.