Defence Lorenzo La Rosa: "Unravelling Atomistic Mechanisms of Martensitic Transformations in Shape-Memory Alloys"

Promotors: Prof Francesco Maresca and Prof Antonis Vakis

Abstract: Shape memory alloys (SMAs) are widely used because they can recover large deformations through reversible martensitic transformations. However, functional fatigue degrades shape recovery over repeated cycles and compromises their long-term reliability. This thesis investigates the atomistic mechanisms governing martensitic transformations in NiTi, using advanced simulations to bridge gaps in experimental and theoretical understanding. In this work, we first establish a benchmarking method for interatomic potentials to ensure the accurate reproduction of experimental martensite microstructure. This enables precise modeling of twin interfaces and self-accommodation mechanisms. A key finding challenges conventional theories: reversible twinning follows a hierarchical selection process governed by twin boundary mobility rather than twin interface energy. Then, to improve modeling accuracy, we develop a high-fidelity interatomic potential using the Performant Atomic Cluster Expansion method. This method achieves ab initio accuracy in twin interface simulations. This advancement enhances phase-field models, enabling more accurate predictions of microstructural evolution. Finally, functional fatigue in SMAs is linked to irreversible deformation from slip activity in the austenite phase. While slip selection in B2 alloys remains an open question, we introduce a new hypothesis integrating atomistic and experimental insights to clarify its governing mechanisms. Overall, this work advances the fundamental understanding of twin interface motion, slip selection, and functional fatigue in SMAs, paving the way for the design of alloys with improved stability and long-term reliability.

Dissertation