Graph Attention-Based Virtual Metrology for Film Deposition Processes in Semiconductor Manufacturing

2026-05-30

AI Summary

This work addresses the challenge in semiconductor manufacturing where physical metrology is constrained by latency, cost, and limited sampling rates, hindering real-time, high-precision control required for high-throughput production. To overcome this, the authors propose a graph attention–based virtual metrology framework that models process parameters and thin-film layers as nodes in a directed heterogeneous graph. A convolutional encoder captures high-frequency temporal features from sensor data, while a graph attention network explicitly learns the structured dependencies from process parameters to film properties. Evaluated on real-world deposition line data, the proposed method significantly outperforms baseline models in prediction accuracy. Moreover, the learned attention weights provide interpretable insights by highlighting key process factors and their temporally resolved dependencies in alignment with known physical principles.

Abstract

Artificial intelligence-driven semiconductor manufacturing increasingly operates at nanometer and angstrom scales, where precise process control depends on accurate and timely metrology. However, physical metrology is limited by measurement latency, cost, and sampling constraints, restricting its scalability in high-volume production. Virtual metrology (VM) has emerged as an effective alternative by predicting wafer-level characteristics from equipment sensor data. Despite recent advances, many existing VM models remain correlation-driven and lack the ability to capture structured dependencies among heterogeneous process variables, while providing limited interpretability. This study presents a graph attention-based VM framework for film deposition processes that integrates temporal feature learning with structured parameter-layer dependency modeling. The proposed approach represents each step-parameter pair as a node and extracts temporal embeddings from high-frequency equipment traces using convolutional feature encoders. A parameter-to-layer graph attention mechanism is employed to model directional dependencies, enabling each film layer to aggregate relevant process information. The framework is evaluated using industrial deposition data collected from production wafers, where the model predicts film thickness from multivariate sensor signals. Experimental results demonstrate improved predictive performance compared to baseline models. In addition, analysis of the learned attention weights reveals interpretable parameter-layer relationships consistent with physical process behavior, capturing dominant process factors and temporal dependencies across deposition stages. These results indicate that the proposed framework enhances prediction accuracy and provides meaningful insight into process dynamics, supporting effective monitoring and optimization in semiconductor manufacturing.

Problem

• Virtual Metrology
• Semiconductor Manufacturing
• Film Deposition
• Process Monitoring
• Graph Attention

Innovation

• Graph Attention
• Virtual Metrology
• Semiconductor Manufacturing
• Temporal Feature Learning
• Interpretable AI