Python-Based Wafer Yield Analysis and Root Cause Identification for Semiconductor Manufacturing

---

Project Overview

This project presents a Python-based semiconductor wafer yield analysis system that simulates a realistic semiconductor manufacturing environment using synthetic wafer process data. The project demonstrates how statistical data analytics can be applied to evaluate manufacturing yield, identify dominant process defects, perform root cause analysis, and support yield improvement in semiconductor fabrication.

A synthetic database containing 250 wafer records across 10 manufacturing lots was developed using Python. The generated dataset includes important semiconductor manufacturing parameters such as particle contamination, oxide thickness, overlay error, implant dose variation, and metal void formation. Statistical analysis and visualization techniques were then applied to understand the relationship between process parameters and manufacturing yield.

---

Objectives

• Develop a synthetic semiconductor wafer manufacturing database.
• Simulate realistic semiconductor process variations.
• Calculate wafer yield using a weighted yield model.
• Identify dominant manufacturing defects.
• Perform root cause analysis.
• Apply Statistical Process Control (SPC).
• Visualize semiconductor manufacturing data using Python.
• Demonstrate engineering-oriented yield analysis techniques.

---

Project Features

• Synthetic wafer database generation
• Yield calculation model
• Root cause identification
• Process status classification
• Statistical Process Control (SPC)
• Correlation analysis
• Yield distribution analysis
• Major defect analysis
• Engineering interpretation of manufacturing data

---

Technologies Used

• Python
• Pandas
• NumPy
• Matplotlib
• OpenPyXL
• Jupyter Notebook
• Microsoft Excel

---

Dataset Overview

Parameter	Description
Manufacturing Lots	10
Total Wafers	250
Total Dies per Wafer	1200
Average Yield	97.09%
Yield Range	93.00% – 99.39%

Process Parameters

• Particle Count
• Oxide Thickness
• Overlay Error
• Implant Dose Error
• Metal Voids
• Good Dies
• Bad Dies
• Yield Percentage
• Major Defect
• Dominant Root Cause
• Process Status

---

Workflow

Semiconductor Process Parameters
                │
                ▼
Synthetic Wafer Database Generation
                │
                ▼
Yield Calculation
                │
                ▼
Defect Identification
                │
                ▼
Root Cause Analysis
                │
                ▼
Statistical Analysis
                │
                ▼
Visualization
                │
                ▼
Engineering Interpretation

---

Project Files

Wafer-Yield-Analysis-using-Python/

├── generate_wafer_database.ipynb
├── Yield_Analysis.ipynb
├── Generated_Wafer_Database.xlsx
├── Wafer_Yield_Analysis_Report.pdf
├── README.md
└── LICENSE

---

Analysis Performed

The project includes:

• Exploratory Data Analysis (EDA)
• Yield Distribution Analysis
• Yield by Manufacturing Lot
• Particle Count Analysis
• Overlay Error Analysis
• Correlation Matrix
• Major Defect Frequency
• Statistical Process Control (SPC)
• Root Cause Identification

---

Key Results

• Generated a realistic semiconductor wafer manufacturing dataset containing 250 wafers.
• Achieved an average simulated wafer yield of 97.09%.
• Identified particle contamination as the dominant manufacturing defect.
• Performed statistical correlation analysis between process parameters and wafer yield.
• Demonstrated how manufacturing process variations influence production yield.

---

Engineering Applications

This project demonstrates concepts used in:

• Semiconductor Yield Engineering
• Process Engineering
• Manufacturing Data Analytics
• Statistical Process Control
• Root Cause Analysis
• Semiconductor Process Optimization

---

Future Improvements

Potential future enhancements include:

• Machine Learning based Yield Prediction
• Wafer Map Visualization
• Predictive Maintenance
• Power BI Dashboard
• Streamlit Web Application
• SQL Database Integration
• Real Semiconductor Fab Dataset Integration

---

How to Run

1. Clone the repository.
2. Install the required Python libraries:

   pip install pandas numpy matplotlib openpyxl jupyter

3. Run 01_Database_Generation.ipynb to generate the wafer database.
4. Run 02_Yield_Analysis.ipynb to perform statistical analysis and generate visualizations.

---

Author

Marut Mahija Roy

M.Tech – Semiconductor Materials and Devices
Indian Institute of Technology Hyderabad (IIT Hyderabad)

LinkedIn: https://www.linkedin.com/in/marut-mahija-roy-3696a01ba/

GitHub: https://github.com/Marut-777

---

License

This project is licensed under the MIT License.

---

⭐ If you found this project useful, consider giving the repository a star.