On phase separation and crystallization of Ge-rich GeSbTe alloys from atomistic simulations with a machine learning interatomic potential
This paper presents a highly transferable machine learning interatomic potential for Ge-rich GeSbTe alloys, which was used to simulate nanosecond-scale crystallization at 600 K, revealing that kinetic effects during memory set operations lead to metastable phase separation into Sb-doped cubic GeTe and amorphous GeSb/Ge rather than thermodynamically stable products.