Describing hydrogen nuclei as quantum particles leads to an increase in the density of ice.

What is it about?

In this work, we use a machine learning interatomic potential to simulate the properties of normal ice. Interatomic potentials describe the interaction between atoms and molecules in such a way that we can describe their movement in space. The use of machine learning methods ensures we can perform simulations for large systems and long periods of time with a lot of precision. Here, we also included the fact that nuclei, particularly hydrogen, must be described quantum mechanically.

Featured Image

Why is it important?

In this work, we show some interesting results. First, by considering different exchange and correlation functionals within density functional theory, we note there is a tendency to overestimate the density of ice in all cases. This means that we are not correctly capturing the interaction between water molecules. The inclusion of quantum nuclear effects increases the density, which is an important piece of information for those looking at the development of novel exchange and correlation potentials. Secondly, we identify that the inclusion of quantum nuclear effects increases the strength of the hydrogen bonds in ice, differently from water clusters.

Perspectives