Linking acoustic emission signals to deformation mechanisms in magnesium

Understanding a material's behavior requires insight into how microscopic deformation mechanisms evolve, but identifying these processes at the level of individual microscopic events is a major challenge. Here, the authors present a physics-guided, data-driven spectral analysis of acoustic emission (AE) signals to classify individual deformation events in a magnesium single crystal. The analysis links AE frequency signatures to twinning and slip mechanisms and validates them through resonance ultrasound spectroscopy and modal calculations. Thus, the study achieves unsupervised classification of deformation events, uncovering the transition from twinning-dominant to slip-dominant behavior. This approach offers a new pathway for mechanism-specific monitoring of damage evolution.