CMSP-SISSA Atomistic Seminar by Dr. Pablo Piaggi, 3 July 11:00hrs

[CMSP Seminars Secretariat]

---------------------------
CMSP-SISSA Atomistic Seminar
---------------------------

** * * Monday 3 July**2023, 11:00***** * **
In person: *Luigi Stasi Seminar Room **(ICTP Leonardo Building, second 
floor)***
**
/Zoom: 
https://zoom.us/meeting/register/tJ0vcuGqrT4oHtGPWExh6dAn5LoFDB0MLNvY/<https://zoom.us/meeting/register/tJIsfuuhqjsrE9EZY2loNxNobg8Lf39NUVHJ>

Speaker:*Pablo Piaggi *(Princeton University)

Title:*Understanding the crystallization of ice polymorphs from first 
principles
**
*
Abstract:
The vast and complex phase diagram of water, with at least 18 different 
ice polymorphs, is a rich playground for the study of crystallization. 
Moreover, the equilibrium picture provided by the phase diagram is 
enriched further by the possible existence of a metastable liquid-liquid 
critical point at deeply supercooled conditions. Over the years, 
considerable attention has been devoted to the study of water and ices, 
and a vast literature has amassed on studies of this system using 
molecular simulations based on empirical potentials. In spite of the 
many merits of empirical models, they are often not able to describe 
important physical effects, such as polarization and chemical reactions. 
A possible strategy to overcome these limitations is the use of 
quantum-mechanical ab initio simulations to derive the potential energy 
surface and use the associated forces to drive the dynamics of the 
nuclei. For many years, this strategy was severely hampered by the sheer 
computational cost of direct ab initio simulations. Recently, the use of 
machine learning algorithms to learn the potential energy surface has 
mitigated the cost of these calculations, paving the way to more 
realistic studies of many systems.  In this talk, I will discuss recent 
developments on the use of first principles simulations and rare-event 
techniques to study the crystallization of ice polymorphs. I will first 
present results on the calculation of homogeneous ice nucleation rates 
from first principles [1].  Then, I will describe the development of a 
machine learning potential for the study of heterogeneous ice nucleation 
at feldspar minerals, one of the most potent ice nucleating particles in 
the atmosphere [2]. Afterwards, I will present evidence of the existence 
of a liquid-liquid transition in ab initio water [3], and I will discuss 
the exotic behavior of the melting lines of several ice polymorphs in 
the vicinity of the liquid-liquid critical point [4]. I will conclude 
the talk with some thoughts about how these techniques are 
revolutionizing our ability to understand and predict the 
crystallization of materials.

[1] Piaggi, Weis, Panagiotopoulos, Debenedetti, and Car, Proc. Natl. 
Acad. Sci. 119, 33 (2022)
[2] Piaggi, Selloni, Panagiotopoulos, Car, and Debenedetti, 
arXiv:2305.10255 (2023)
[3] Gartner, Piaggi, Car, Panagiotopoulos, and Debenedetti, Phys. Rev. 
Lett. 129, 25 (2022)
[4] Piaggi, Gartner, Car, and Debenedetti, arXiv:2302.08540 (2023)