Abstract:<p>Defective wafer detection is essential to avoid loss of yield due to process abnormalities in semiconductor manufacturing. For most complex processes in semiconductor manufacturing, various sensors are installed on equipment to capture process information and equipment conditions, including pressure, gas flow, temperature, and power. Because defective wafers are rare in current practice, supervised learning methods usually perform poorly as there are not enough defective wafers for fault detection (FD). The existing methods of anomaly detection often rely on linear excursion detection, such as principal component analysis (PCA), k-nearest neighbor (kNN) classifier, or manual inspection of equipment sensor data. However, conventional methods of observing equipment sensor readings directly often cannot identify the critical features or statistics for detection of defective wafers. To bridge the gap between research-based knowledge and semiconductor practice, this paper proposes an anomaly detection method that uses a denoise autoencoder (DAE) to learn a main representation of normal wafers from equipment sensor readings and serve as the one-class classification model. Typically, the maximum reconstruction error (MaxRE) is used as a threshold to differentiate between normal and defective wafers. However, the threshold by MaxRE usually yields a high false positive rate of normal wafers due to the outliers in an imbalanced data set. To resolve this difficulty, the Hampel identifier, a robust method of outlier detection, is adopted to determine a new threshold for detecting defective wafers, called MaxRE without outlier (MaxREwoo). The proposed method is illustrated using an empirical study based on the real data of a wafer fabrication. Based on the experimental results, the proposed DAE shows great promise as a viable solution for on-line FD in semiconductor manufacturing.</p>

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: the detection of defective wafers in the semiconductor manufacturing process due to process abnormalities. Specifically, the paper focuses on the following points: 1. **Rarity of Defective Wafers**: In current semiconductor manufacturing practices, the number of defective wafers is very small, which makes it difficult for supervised learning methods to work effectively because there are not enough defective samples for fault detection (FD). 2. **Limitations of Traditional Methods**: - Traditional anomaly detection methods (such as Principal Component Analysis (PCA), k - Nearest Neighbor (kNN) classifier, etc.) usually rely on linear models or manual inspection of device sensor data, but these methods cannot effectively identify key features or statistical information to detect defective wafers. - When conventional methods directly observe device sensor readings, they are often unable to accurately identify defective wafers. 3. **High - Dimensional Non - Linear Data**: Sensor data generated in the semiconductor manufacturing process is usually multivariate, highly correlated, and non - normally distributed, which poses a challenge for constructing effective multivariate models. To solve these problems, the paper proposes an anomaly detection method based on Denoising Autoencoder (DAE) and combines Hampel identifier to optimize the detection threshold. The following are the core objectives and solutions of the paper: ### Core Objectives - Propose a data - driven method that does not rely on domain - knowledge - based feature extraction for real - time detection of defective wafers in the semiconductor manufacturing process. - Solve the problem of high false - positive rate or low detection rate caused by data imbalance (more normal wafers and fewer defective wafers). ### Solutions 1. **Use DAE to Extract Features**: - DAE generates a one - class classification model for anomaly detection by learning the device sensor readings of normal wafers. - DAE can extract important features from partially damaged or interfered input data and reconstruct the undamaged input data. 2. **Optimize Detection Threshold**: - Use the Maximum Reconstruction Error (MaxRE) as a preliminary threshold to distinguish between normal wafers and defective wafers. - Introduce Hampel identifier to filter outliers in the reconstruction error, thereby determining a new threshold - Maximum Reconstruction Error without Outliers (MaxREwoo). 3. **Real - Time Monitoring**: - This method can directly monitor the wafer state from device sensor readings without relying on post - hoc measurement data (such as metrology data) for quality assessment. ### Summary The paper aims to solve the challenges of defective wafer detection in semiconductor manufacturing, especially in view of the problems of rare defective wafers, data imbalance, and deficiencies in traditional methods. By combining DAE and Hampel identifier, the paper proposes an efficient, real - time anomaly detection method, which significantly reduces the false - positive rate and improves the detection performance.

Defective wafer detection using a denoising autoencoder for semiconductor manufacturing processes

Autoencoder-based detector for distinguishing process anomaly and sensor failure

Two-Dimensional Principal Component Analysis-Based Convolutional Autoencoder for Wafer Map Defect Detection

Machine Learning Based Wafer Defect Detection

Machine Learning-Based Detection Method for Wafer Test Induced Defects

Elevating Wafer Defect Inspection with Denoising Diffusion Probabilistic Model

Wafer Map Defect Classification Using Autoencoder-Based Data Augmentation and Convolutional Neural Network

A voting-based ensemble feature network for semiconductor wafer defect classification

Deep Learning-Based Detection of Defects in Wafer Buffer Zone During Semiconductor Packaging Process

Data Visualization of Anomaly Detection in Semiconductor Processing Tools

Wafer Bin Map Recognition with Autoencoder-based Data Augmentation in Semiconductor Assembly Process

DSCU-Net: MEMS Defect Detection Using Dense Skip-Connection U-Net

Fault Diagnosis of Wafer Acceptance Test and Chip Probing Between Front-End-of-Line and Back-End-of-Line Processes

An unknown wafer surface defect detection approach based on Incremental Learning for reliability analysis

Towards Robust and Understandable Fault Detection and Diagnosis Using Denoising Sparse Autoencoder and Smooth Integrated Gradients

Robust Monitoring and Fault Isolation of Nonlinear Industrial Processes Using Denoising Autoencoder and Elastic Net

Hybrid Semiconductor Wafer Inspection Framework via Autonomous Data Annotation

Auto-Labeling for Pattern Recognition of Wafer Defect Maps in Semiconductor Manufacturing

Sparse deep encoded features with enhanced sinogramic red deer optimization for fault inspection in wafer maps

Advances in machine learning and deep learning applications towards wafer map defect recognition and classification: a review

Clear Memory-Augmented Auto-Encoder for Surface Defect Detection