High resolution X-ray emission spectroscopy of water and its assignment based on two structural motifs

We evaluate proposed interpretations of previous X-ray emission spectroscopy (XES) data on liquid water. The split peak in the lone-pair orbital region has been interpreted in terms of either two different structural motifs, tetrahedral and distorted, or as due to core-hole-induced dissociative dynamics; here we add new data on a 1:1 H2O/D2O isotopic mixture and additional spectrum simulations including the core-hole-induced dynamics. The XES spectrum of HDO is quite nicely reproduced as the sum of spectra of H2O and D2O, which we interpret as that core-hole-induced dynamics contribute only to the peak shape and do not affect the intensity ratio between tetrahedrally coordinated and distorted. We find the simulation-based interpretation of the two lone-pair peaks as being of completely different symmetries, molecular 1b(1) and dissociated 3a(1), difficult to reconcile with the experimental intensities in the 1b(2) and 3a(1) spectral regions. We report extensive theoretical simulations of spectra probing both the distance and velocity quantum distributions of the internal OH stretch; sharp features not associated with the lone-pair, that are seen when the OH stretch is treated as a classical oscillator, become smeared out when the zero-point Franck-Condon profile and momentum distribution in the v=0 level of the OH stretch are taken into account. This demonstrates that neglecting zero-point motion in simulating XES spectra of water generates artificially sharp structures. XES spectra of 1 M and 4 M hydrochloric acid (HCl) and sodium hydroxide (NaOH) are reported. These spectra indicate that dissociated species most likely can be excluded as the origin of the double 1b(1) peak structure. We thus argue that the experimental observation of two distinct peaks in the lone-pair region is less likely to be explained by an unstructured continuum model of the liquid, but is easily explained within a two-component fluctuating model. (C) 2010 Elsevier B.V. All rights reserved.