Diffusion-based Generative Machine Learning Model for Predicting Crack Propagation in Aluminum Nitride at the Atomic Scale
This work addresses the challenge of rapid crack prediction for semiconductor reliability optimization, representing a novel method for a known bottleneck in materials science.
The paper tackles the problem of predicting atomic-scale crack propagation in aluminum nitride, which is critical for semiconductor reliability but computationally expensive with molecular dynamics, by developing a diffusion-based generative machine learning model that achieves significant speedup and accurately forecasts dynamic fracture processes, including crack initiation and branching, while generalizing to unseen configurations.
Predicting atomic-scale crack propagation in aluminum nitride (AlN) is critical for semiconductor reliability but remains prohibitively expensive via molecular dynamics (MD). We develop a diffusion-based generative machine learning model to predict atomic-scale crack propagation in AlN, a critical semiconductor material, by conditioning solely on initial microstructure embeddings. Trained on MD simulations of single-crack systems, the model achieves a significant speedup while accurately forecasting dynamic fracture processes, including stress-driven crack initiation, crack branching, and atomic-scale bridging ligaments. Crucially, it demonstrates inherent physical fidelity by reproducing material-intrinsic mechanisms while disregarding periodic boundary artifacts, and generalizes to unseen multi-crack configurations. Validation against MD ground truth confirms the capability of the model to capture complex fracture physics without auxiliary stress or energy data, enabling rapid exploration of crack-mediated failure for semiconductor reliability optimization.