We have investigated the electronicstructure of water and ice
using a combination of experimental and theoretical techniques [1]. The
X-ray Absorption (XA) spectrum of the liquid is distinctly different from
that of tetrahedrally coordinated bulk ice, where the liquid shows a
distinct pre-edge feature and a strong enhancement of the intensity at the
edge.
Through spectrum simulations and model experiments (bulk and surface of
ice) we show that the specific features in the liquid spectrum are due
exclusively to asymmetric configurations with only two strong hydrogen
bonds: one donating and one accepting; this could be indicative of chain-
or ring-like structures in the liquid but is not reproduced by present
simulation techniques [1]. This result has caused a heated debate in the
literature [e.g., 2-11]. Here I will discuss the interpretation of the
spectra on the basis of theoretical modeling and additional experimental
data. Neutron and x-ray diffraction have been of particular importance in
determining the structure of the liquid. A recent reevaluation of the
neutron and x-ray diffraction data by A. Soper [12] shows equally good
fit of the data using an asymmetric H-bonding model according to [1] as
for the
standard tetrahedral model. Using the structures obtained by Soper we,
however, show that none of the thus obtained models fit the available XA
or IR/Raman data and furthermore have to question standard approaches to
connect water structure with IR/Raman spectra. To resolve the situation we
are using Reverse Monte Carlo (RMC) [13] modeling to fit a range of water
models (hexamer rings, tetrahedral, mixed) to new, more extended x-ray and
neutron diffraction data as well as data from Compton scattering
experiments.
