Semiconductor Yield Improvement via Customer Data Integration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Semiconductor manufacturing processes lack effective methods to determine quality affecting factors based on real-world customer evaluations, leading to inefficiencies in product yield and reliability.
Innovation Solution
A method involving a supplier server connected to customer servers through a communication network, which preprocesses customer evaluation data using statistical models to identify critical quality factors, classify products as good or failed, and generate quality affecting factors to improve yield by iterative feedback loops.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If semiconductor manufacturing companies rely only on internal evaluation results to determine quality affecting factors, then the manufacturing process can be managed with existing data infrastructure, but the quality assessment does not reflect real-world customer usage conditions and leads to inaccurate quality predictions
Solution Approach 1:
The patent combines internal manufacturing evaluation data with external customer usage evaluation data into a unified quality assessment system. The server receives and integrates both data sources, applying statistical models to collectively determine quality affecting factors, thereby achieving more accurate quality predictions that reflect both manufacturing conditions and real-world usage.
Solution Approach 2:
The patent introduces a server as an intermediary component that collects, processes, and analyzes quality data from both manufacturing and customer usage environments. This intermediary server applies statistical models to determine quality affecting factors, acting as a bridge between internal manufacturing processes and external customer feedback without requiring direct complexity in either end system.
2Productivity
If customer evaluation data is collected and processed in real-time through statistical models, then quality affecting factors can be accurately determined to improve yield, but data processing time and computational resources increase
Solution Approach 1:
The patent applies statistical models to customer evaluation data to pre-determine quality affecting factors before actual manufacturing decisions are made. By analyzing customer usage data in advance and identifying critical quality factors, the system prepares quality assessment criteria that can be quickly applied to manufacturing yield improvement without requiring time-consuming analysis during production.
Solution Approach 2:
The patent replaces manual or mechanical quality assessment methods with automated statistical modeling and data analysis systems. The server automatically processes customer evaluation data using statistical models to determine quality affecting factors, substituting computational automation for traditional time-consuming quality analysis procedures, thereby reducing processing time while improving accuracy.
3Reliability
If comprehensive quality data from customer usage is integrated into manufacturing management, then quality affecting factors become more accurate, but the system complexity and data management burden increase
Solution Approach 1:
The patent extracts only the essential quality evaluation data from customer usage environments that are relevant to manufacturing quality factors. The server selectively receives and processes specific quality-related data from customer evaluations, filtering out unnecessary information, thereby integrating meaningful quality data into manufacturing management without requiring complex management of all customer data.
Solution Approach 2:
The patent creates a multi-functional server system that handles multiple tasks: collecting customer evaluation data, applying statistical models, determining quality affecting factors, and providing recommendations for yield improvement. This universal system performs multiple quality management functions through a single integrated platform, reducing the need for separate complex systems for each function.
Data Source
AI summary
A quality affecting factor generation method for a semiconductor manufacturing process is provided. The method includes receiving data of a customer evaluation result obtained by a real use of shipped semiconductor products, preprocessing the data of the customer evaluation result, determining critical quality factors that affect a quality of the semiconductor products by applying a statistical model to the preprocessed data of the customer evaluation result; and determining a semiconductor product to be shipped to a customer company as a good product and a failed product using the determined critical quality factors and generating quality affecting factors of the semiconductor products to be managed to improve yield in a semiconductor manufacturing process.