Ice phonon spectra and Bayes inference: a gateway to a new understanding of terahertz sound propagation in water

TL;DR

This paper uses Bayesian inference to analyze and compare phonon spectra of ice and water, offering new insights into molecular collective motions and challenging existing theories of density fluctuations in networked liquids.

Contribution

It introduces Bayesian inference as a novel approach to interpret phonon spectra, providing a more detailed understanding of density fluctuations in ice and water.

Findings

Bayesian inference enhances spectral analysis accuracy.

Comparison reveals differences in phonon responses between ice and water.

Results challenge existing models of density fluctuations in liquids.

Abstract

Understanding how molecules engage in collective motions in a liquid where a network of bonds exists has both fundamental and applied relevance. On the one hand, it can elucidate the ``ordering" role of long-range correlations in an otherwise strongly dissipative system; on the other hand, it can inspire new avenues to control such order to implement sound manipulation. Water represents an ideal investigation case to unfold these general aspects and, across the decades, it has been the focus of thorough scrutiny. Despite this investigative effort, the spectrum of terahertz density fluctuations of water largely remains a puzzle for Condensed Matter physicists. To unravel it, we compare previous scattering measurements of water spectra with new ones on ice. Thanks to the unique asset of Bayesian inference, we draw a more detailed portrayal of the phonon response of ice. The comparison with the one of liquid water challenges the current understanding of density fluctuations in water, or more in general, of any networked liquid.