First-principles calculations of the fragility and the thermodynamic properties of GeTe and antimony

by Prof. Riccardo Mazzarello (Universitá La Sapienza, Rome, Italy)

Thursday 9 Oct 2025, 16:00 → 17:00 Europe/Berlin

XHQ / E1.173 (European XFEL)

Phase-change materials (PCMs) undergo fast and reversible transitions between crystalline and amorphous states, exhibiting pronounced contrasts in reflectivity and resistivity. Under ambient conditions, both states are very stable. These properties entail a strong temperature dependence of the crystallization kinetics, which has been attributed to the high fragility of the supercooled liquid phase. Such characteristics are exploited in optical and electronic storage devices and make PCMs promising candidates for neuromorphic computing applications.

In the first part of this talk, I will discuss our molecular dynamics investigation of GeTe, a prototypical PCM and the parent compound of the GeSbTe alloys employed in non-volatile memories. We evaluate the configurational entropy of the supercooled phase by sampling the potential energy landscape and performing thermodynamic integration. We also calculate the viscosity and the α-relaxation time in the same temperature range. Finally, we employ the Adam–Gibbs relation to extrapolate the viscosity down to the glass transition temperature and to estimate the fragility index. We obtain a value on the order of 135–140, which confirms that liquid GeTe is a highly fragile system.

In the second part, I will focus on antimony. Bulk antimony crystallizes rapidly even under ambient conditions; however, its amorphous phase can be stabilized through alloying or nanoconfinement. Previous studies have suggested that Ge-alloyed antimony exhibits a liquid–liquid transition in the supercooled regime. Such a transition may benefit phase-change applications, as it is accompanied by a fragile-to-strong dynamical crossover that enhances glass stability.

We investigate the thermodynamic and kinetic properties of liquid antimony to assess whether a liquid–liquid transition occurs. To this end, we employ molecular dynamics simulations with neural-network-based interatomic potentials. We observe thermodynamic anomalies akin to those of water and compute viscosity, relaxation times, and configurational entropy as functions of temperature and pressure. Finally, we examine the nucleation mechanisms across a range of thermodynamic conditions.

Zoom info:

https://xfel.zoom.us/j/62165022943?pwd=VHIzpEwJgkkBcROvF4z2MJDhe5UqPw.1

Meeting ID: 621 6502 2943
Passcode: 213774