Sequence-Aware Inline Measurement Attribution for Good-Bad Wafer Diagnosis

2025-07-27

AI Summary

In semiconductor manufacturing, tracing wafer defects across hundreds of sequential process steps remains highly challenging; conventional Shapley value-based attribution methods neglect temporal ordering and inter-step dependencies, leading to inaccurate root-cause localization. To address this, we propose Trajectory Shapley Attribution—a novel framework that explicitly incorporates process sequence into Shapley value computation. By introducing sequence-aware marginal contribution evaluation, it precisely quantifies the causal influence of upstream process parameters on downstream defects. Unlike existing approaches, our method requires no predefined reference point and explicitly models the temporal dependencies inherent in fabrication workflows. Evaluated on real-world front-end and back-end process data from Albany NanoTech’s experimental line, the framework significantly improves yield diagnostic efficiency and accurately identifies critical metrology features strongly correlated with defects. This work establishes a new, interpretable, and scalable paradigm for root-cause analysis in high-complexity semiconductor manufacturing.

Technology Category

• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Causality

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Security and Privacy: Data transparency and provenance
• Responsible Web: Algorithmic accountability and transparency on the web

Abstract

How can we identify problematic upstream processes when a certain type of wafer defect starts appearing at a quality checkpoint? Given the complexity of modern semiconductor manufacturing, which involves thousands of process steps, cross-process root cause analysis for wafer defects has been considered highly challenging. This paper proposes a novel framework called Trajectory Shapley Attribution (TSA), an extension of Shapley values (SV), a widely used attribution algorithm in explainable artificial intelligence research. TSA overcomes key limitations of standard SV, including its disregard for the sequential nature of manufacturing processes and its reliance on an arbitrarily chosen reference point. We applied TSA to a good-bad wafer diagnosis task in experimental front-end-of-line processes at the NY CREATES Albany NanoTech fab, aiming to identify measurement items (serving as proxies for process parameters) most relevant to abnormal defect occurrence.

Problem

• Identify problematic upstream processes for wafer defects
• Analyze root causes across complex semiconductor manufacturing steps
• Overcome limitations of standard Shapley values in sequential processes

Innovation

• Extends Shapley values for sequential processes
• Overcomes reference point reliance in SV
• Applies TSA to wafer defect diagnosis