Here is a seemingly simple question: if you cool water to just below zero degrees Celsius will you see it turn into ice? The answer may surprise you. It is actually surprisingly difficult to observe ice crystallization from pure liquid water, it can be cooled down to −40 °C without freezing [1]. Our everyday experience of ice forming near zero degrees is actually caused by a process called "heterogeneous ice nucleation" whereby a foreign particle promotes the formation of ice. From glaciers, to cryopreservation, to climate modelling – the ability of materials to promote ice nucleation is at the heart of a myriad of technologies and natural phenomena. Understanding how materials affect ice nucleation thus presents great opportunity; however, it has presented an equally great challenge: despite over 75 years of research we lack a fundamental understanding of what makes a material good or bad at promoting ice. In our group we work to understand how ice nucleation proceeds using computational techniques such as molecular dynamics which allow us to directly observe nucleation at the atomic scale. Recent highlights include: (i) using deep learning to accurately predict a materials ice nucleation ability rapidly and cheaply from room temperature water; (ii) exploring routes to the formation of cubic ice – an elusive polymorph of ice with the structure of diamond that impacts atmospheric phenomena; (iii) showing that ice is born in the immobile regions of supercooled water.
On the other hand, we can also ask the question of another everyday observation – how do these crystals dissolve? Life on earth depends upon the dissolution of ionic salts in water, while understanding this process is also vital for tackling pressing technological challenges including battery science and water desalination. However despite its ubiquity, we have a very poor understanding of dissolution at a microscopic level. Due to the long time scales and accuracy required in modelling these systems, it presents many challenges to our current computational methods. The group has addressed fundamental questions about the mechanism of NaCl dissolution in water. Early first principles based studies on dissolution provided important insight into the process, however computational limitations meant that only the initial stages could be explored [2,3]. This leaves many unresolved questions, from the nature of the mechanism of dissolving crystals to the behaviour of these crystals under confinement. We are currently exploiting developments in machine learning potential methodology to study the entire process of dissolution accessing long time scale trajectories with an accuracy equivalent to ab initio methods.
References:
[1] Murray, B. J.; O'Sullivan, D.; Atkinson, J. D.; Webb, M. E. Chem. Soc. Rev. 2012, 41, 6519– 6554 DOI: 10.1039/c2cs35200a
[2] Li-Min Liu, Alessandro Laio, and Angelos Michaelides. Initial stages of salt crystal dissolution determined with ab initio molecular dynamics. Phys. Chem. Chem. Phys., 13(29):13162–13166
[3] Jirí Klimeš, David R Bowler, and Angelos Michaelides. Understanding the role of ions and water molecules in the NaCl dissolution process. J. Chem. Phys., 139(23):234702