Prediction of properties from first principles with quantitative accuracy: six representative ice phases

Abstract

High-pressure ice is an important part of the Earth's crust and an important research target for the exploration of outer space. The significant changes of intermolecular distance in ice at high pressure will lead to essential changes in structure, morphology, symmetry, etc. Yet the atomic-level structures of ice are still far from well understood, which results in many aspects of ice phases remaining uncertain or unknown. Ice phase transition under high pressure is a problem that has not been solved for more than 100 years, due to the low accuracies in describing the hydrogen bonding and van der Waals interactions. By using the embedded-fragment approach at the second-order Møller–Plesset perturbation (MP2) level, we study the structures and Raman spectra of six representative ice phases (ices II, VI, VII, VIII, IX, XV) at high pressures and predict their phase transitions up to 12 GPa quantitatively. The experimentally determined crystal structural parameters, Raman spectra, and ice phase boundaries are reproduced accurately by our simulation, identifying that our approach offers accurate and quantitative predictions of observable properties of different ice phases over a broad range of pressure. The proposed work not only determines the ice stability in different high pressure ranges, but also motivates scientists to explore new molecules in the interdisciplinary fields and will significantly advance scientific understanding.