Reverse Monte Carlo investigations concerning recent isotopic substitution neutron diffraction data on liquid water

TL;DR

This study uses Reverse Monte Carlo modeling to analyze recent neutron diffraction data with isotopic substitution on liquid water, revealing consistent local hydrogen bond structures despite data inconsistencies.

Contribution

It applies Reverse Monte Carlo methods to interpret recent neutron diffraction data with isotopic substitution, highlighting structural details of hydrogen bonds in water.

Findings

Hydrogen bond angles are predominantly straight.

Oxygen neighbors form approximately tetrahedral arrangements.

Data inconsistencies reveal different modeling challenges.

Abstract

Although liquid water has been studied for many decades by (X-ray and neutron) diffraction measurements, new experimental results keep appearing, virtually every year. The reason for this is that neither X-ray, nor neutron diffraction data are trivial to correct and interpret for this essential substance. Since X-rays are somewhat insensitive to hydrogen, neutron diffraction with (most frequently, H/D) isotopic substitution is vital for investigating the most important feature in water: hydrogen bonding. Here, the two very recent sets of neutron diffraction data are considered, both exploiting the contrast between light and heavy hydrogen, $^1$H and $^2$H, in different ways. Reverse Monte Carlo structural modeling is applied for constructing large structural models that are as consistent as possible with all experimental information, both in real and reciprocal space. The method has also proven to be useful for revealing where possible small inconsistencies appear during primary data processing: for one neutron data set, it is the molecular geometry that may not be maintained within reasonable limits, whereas for the other set, it is one of the (composite) radial distribution functions that cannot be modeled at the same (high) level as the other three functions. Nevertheless, details of the local structure around the hydrogen bonds appear very much the same for both data sets: the most probable hydrogen bond angle is straight, and the nearest oxygen neighbours of a central oxygen atom occupy approximately tetrahedral positions.