Accurate and efficient parameterization of an atomic cluster expansion (ACE) potential for ammonia under extreme conditions
JT Willman and R Perriot and C Ticknor, JOURNAL OF CHEMICAL PHYSICS, 162, 144316 (2025).
DOI: 10.1063/5.0252373
We present a machine learning interatomic potential for ammonia designed
to capture its complex multiphase behavior, including both molecular and
superionic phases. The potential is based on the atomic cluster
expansion (ACE) formulation and has been parameterized to facilitate
high-fidelity molecular dynamics simulations of ammonia under extreme
conditions, for pressures up to 100 GPa and for temperatures above 500 K
and up to 6000 K. A diverse range of configurations was generated
through high-quality ab initio molecular dynamics simulations, covering
insulating and superionic ice phases, liquid ammonia, molecular nitrogen
(N-2) and hydrogen (H-2), and metastable compounds that form upon
dissociation, including NH4+, H-3(+), N2H4, and N3H. We demonstrate that
the ammonia ACE potential accurately reproduces experimental and density
functional theory predicted isotherms and Hugoniots. Crucially, the
potential is able to capture the intricate phase behavior of ammonia,
including the transition from insulating molecular fluid to the
superionic phase. This work provides a robust interatomic potential that
can be used for large-scale, accurate simulations of ammonia under
extreme thermodynamic conditions, offering a powerful tool for
investigating its behavior in various phases and applications.
(c)
2025 Author(s). All article content, except where otherwise noted, is
licensed under a Creative Commons Attribution-NonCommercial
4.0International (CC BY-NC) license
(https://creativecommons.org/licenses/by-nc/4.0/).
Return to Publications page