Abstract:Due to the complexity and dynamics of the semiconductor manufacturing processes, wafer bin maps (WBM) present various defect patterns caused by various process faults. The defect type detection on wafer maps provides information about the process and equipment in which the defect occurred. Recently, automatic inspection has played a vital role in meeting the high-throughput demand, especially with deep convolutional neural networks (DCNN) which shows promising efficiency. At the same time, the need for a large amount of labeled and balanced datasets limits the performance of such approaches. In addition, complex DCNN in recognition tasks can provide redundant features that cause overfitting and reduce interpretability. In this paper, a new hybrid deep model for wafer map fault detection to get over these challenges is proposed. Firstly, a new convolutional autoencoder (CAE) is employed as a synthetization model to fix the high imbalance problem of the dataset. Secondly, for efficient dimensionality reduction, an embedding procedure is applied to the synthesized maps to get sparse encoded wafer maps by reinforcing a sparsity regularization in an encoder-decoder network to form a sparsity-boosted autoencoder (SBAE). The sparse embedding of wafer maps guarantees more discriminative features with 50% reduction in spatial size compared to the original wafer maps. Then, the 2D encoded sparse maps are converted to 1D sinograms to be fed later into another aggressive feature reduction stage using a new modified red deer algorithm with a new tinkering strategy. The resultant feature pool is reduced to ~ 25 1D feature bases, i.e., ~ 1.5% of the initial size of the 2D wafer maps. Finally, for the prediction stage, a simple 1DCNN model is introduced. The proposed inspection model is tested via different experiments on real-world wafer map dataset (WM-811K). Compared to state-of-the-art techniques, the proposed model outperforms their performance even with small-sized 1D feature pool. The average testing accuracy are 98.77% and 98.8% for 9 and 8 types of faults, respectively.

What problem does this paper attempt to address?

The paper mainly addresses the issue of wafer defect detection in the semiconductor manufacturing process by proposing a new hybrid deep learning model to improve the performance of automatic detection. Specifically, the paper tackles the following key issues: 1. **Data Imbalance Problem**: Wafer Bin Maps (WBM) in the semiconductor manufacturing process exhibit different defect patterns due to various process faults. The publicly available dataset WM-811K used in the paper has a severe class imbalance problem, where the number of samples for certain types of defects is much smaller than for other types. 2. **Feature Redundancy and Overfitting Problem**: Traditional Deep Convolutional Neural Networks (DCNN) can provide good recognition efficiency, but they tend to generate redundant features on small sample and imbalanced datasets, leading to overfitting issues and reducing the model's interpretability. 3. **High-Dimensional Feature Processing Problem**: When directly using traditional 2D convolutional neural networks for defect classification, the low resolution of wafer images can lead to unstable classification results. Additionally, high-dimensional features increase computational complexity, reduce memory efficiency, and further exacerbate the risk of overfitting. To address the above issues, the paper proposes the following main contributions: 1. **Using Convolutional Autoencoder (CAE)**: First, a convolutional autoencoder is constructed for data augmentation to solve the data imbalance problem. Second, by introducing a Sparsity-Boosted Autoencoder (SBAE), efficient feature dimensionality reduction is achieved, ensuring feature discriminability while reducing the dimensionality of the feature space. 2. **Optimized Feature Selection**: The Enhanced Red Deer Optimization (ERD) algorithm is used for further selection and optimization of features, compressing the feature pool to approximately 1.5% of its initial size, significantly reducing the number of features and avoiding the interference of redundant information. 3. **1D Feature Representation and Classification**: The 2D wafer maps are converted into a 1D representation, effectively extracting the most significant features through a series of steps, and finally using a simple 1D Convolutional Neural Network (1DCNN) for classification. This approach achieves high accuracy even on the highly compressed feature set. Experimental results show that the model achieves an average test accuracy of 98.77% to 98.8% on the WM-811K dataset, demonstrating better performance compared to existing techniques.

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

A Deep Convolutional Neural Network for Wafer Defect Identification on an Imbalanced Dataset in Semiconductor Manufacturing Processes

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

Wafer map defect pattern detection method based on improved attention mechanism

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

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

Convolutional Neural Network for Semiconductor Wafer Defect Detection

AI classification of wafer map defect patterns by using dual-channel convolutional neural network

A Deep convolutional neural network with residual blocks for wafer map defect pattern recognition

A Light-Weight Neural Network for Wafer Map Classification Based on Data Augmentation

A Deep Convolutional Neural Network-Based Multi-Class Image Classification for Automatic Wafer Map Failure Recognition in Semiconductor Manufacturing

Wafer map failure pattern classification using geometric transformation-invariant convolutional neural network

A Deep Learning-Based Surface Defect Inspection System Using Multiscale and Channel-Compressed Features

Contrastive deep clustering for detecting new defect patterns in wafer bin maps

A self-supervised learning framework based on masked autoencoder for complex wafer bin map classification

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

Spatio-Temporal 3-D Residual Networks for Simultaneous Detection and Depth Estimation of CFRP Subsurface Defects in Lock-In Thermography

Recognition and Classification of Mixed Defect Pattern Wafer Map Based on Multi-Path DCNN

Wafer Map Defect Patterns Semi-Supervised Classification Using Latent Vector Representation

Efficient Mixed-Type Wafer Defect Pattern Recognition Based on Light-Weight Neural Network