Investigation of rare events and slow dynamics in atomic and molecular liquids and glasses using ascent dynamics

Zhai, Yanqin

Permalink

https://hdl.handle.net/2142/117557 Copy

Description

Title

Investigation of rare events and slow dynamics in atomic and molecular liquids and glasses using ascent dynamics

Author(s)

Zhai, Yanqin

Issue Date

2022-11-22

Director of Research (if dissertation) or Advisor (if thesis)

Zhang, Yang

Doctoral Committee Chair(s)

Zhang, Yang

Committee Member(s)

• Schweizer, Kenneth S.
• Stubbins, James F.
• Kozlowski, Tomasz

Department of Study

Nuclear, Plasma, & Rad Engr

Discipline

Nuclear, Plasma, Radiolgc Engr

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Supercooled Liquid
• Slow Dynamics
• Accelerated Simulation
• Ascent Dynamics

Abstract

Far-from-equilibrium phenomena commonly exist in nature, for example, the crystallization and the glass transition, which are typically driven by rate events and can evolve for an extremely long timescale. As classical simulation techniques are limited in both spatial and temporal scales for the investigation of the rare event sampling and the long timescale phenomenon, we recently proposed that the Ascent Dynamics (AD) algorithm enables an efficient sampling of the potential energy landscape with an adjustable resolution by tuning a steepness parameter. We implemented the AD algorithm in LAMMPS, a widely utilized molecular dynamics simulation package supporting abundant categories of force fields, to study the dynamics in realistic systems--the Zr50Cu50 metallic supercooled liquid and the water. For the metallic glass, with pure AD simulations, we observed a gradual change from local to global structural rearrangement with dynamic heterogeneity decreasing as the steepness parameter increases. It identified a string-like local structural rearrangements serving as the possible mechanism for the secondary relaxation. We further combined the nudged elastic band method with the AD algorithm to study the activation energy and entropy of activation processes. The result unveiled the importance of the entropy effects in determining the transition rate of local activation. Finally, we probed the equilibrated dynamics of the system in a timescale greatly exceeding the accessibility of the traditional simulation techniques by evaluating the time correlation function. The thermodynamic properties, including the structural relaxation time and the glass transition temperature, demonstrated a good consistency with previous experimental observations and theoretical predictions, suggesting the preeminent efficacy of the AD method. We then employed a flexible water model to enable the application of the AD method in molecular systems. By performing the pure AD sampling with various steepness parameters, we identify the decoupling of the translational, rotational, and intramolecular motions. Different types of motion are coupled with different inherent activation energies, which are triggered by the AD sampling with different steepness parameters. Meanwhile, the structural relaxation introduced by the AD sampling is inclined to not disrupt formed hydrogen bonds, indicating the stability of the hydrogen-bond networks in water. Furthermore, by applying the MDAD sampling in water, we successfully retrieved the microscopic dynamics and macroscopic transport properties in the deeply supercooled regime which has never been reached by either experiment or simulation. The successful application of the AD sampling method in realistic systems marks the emergence of an efficient accelerating sampling method for simulating rare events and slow dynamics. This method can provide insights into the behaviors of phenomena far from equilibrium, for example, the crystallization, the glass transition, and even the protein folding, covering a wide range of disciplines, including physics, chemistry, biology, materials, etc. We anticipate this method to play a significant role in simulating rare events and slow dynamics beyond the accessible temporal regime of conventional simulation techniques.

Graduation Semester

2022-12

Type of Resource

Thesis

Copyright and License Information

Copyright 2022 Yanqin Zhai

Owning Collections