Comment on ‘‘Isotope and Temperature Effects in Liquid Water Probed by X-Ray Absorption and Resonant X-Ray Emission Spectroscopy’’ Fuchs et al. [1] interpret the split lone-pair orbital (1b 1 ) in high-resolution x-ray emission spectra (XES) of liquid water based on core-hole dynamics leaving either intact molecular water (d 1 peak) or causing dissociation to an OH radical (d 2 peak), which they furthermore assume can be represented by OH  anion. We find this picture, postulat- ing emission either around the initial or the final structure, but not along the path connecting them, unphysical and unsubstantiated [2]. Figure 1 shows simulated XES spectra at t  0 and integrated along core-hole induced trajectories, 10 fs long, with forces from Density Functional Theory [3]; random directions were assigned to the maximum veloc- ities in the O-H stretch zero-point vibration. The initial geometries sample a molecular dynamics simulation rep- resenting predominantly tetrahedral structures, which is not necessarily representative of the liquid [4]. We indeed find dissociation to various degrees, but towards both O and OH. The varying directions of the initial velocities and dissociative dynamics smear out the bonding 1b 2 and 3a 1 features but give no split in the 1b 1 peak. Resonant excitation into the strongly antibonding OH   pre-edge state of water gives faster dissociation [3]; the experimental 1b 2 and 3a 1 spectral features in fact become much more smeared out. Inexplicably then, according to the model in Ref. [1], spectra a and b in their Fig. 2 instead indicate a decreased d 2 intensity (dissociated OH accord- ing to Ref. [1]) going towards the resonance (most clearly seen for D 2 O). The d 1 , 1b 2 , and 3a 1 spectral features show a Raman shift in energy position with excitation energy so that spectra c show only one lone-pair peak which, at the particular selected excitation energy, happens to coincide with the d 2 position. At slightly lower and higher excitation energies, the d 1 peak no longer overlaps the d 2 position revealing that the d 2 peak has all but disappeared [5] and that the nearly coincident resonant OH  and nonresonant water d 2 positions are purely incidental and not causal. The presence of molecular water 1b 1 and O-H bonding 1b 2 and 3a 1 features in resonantly excited spectra c thus excludes full dissociation giving the d 2 peak. For OH  solutions, only a pre-pre-edge XAS peak at 533 eV [1] is unique to OH  ; at higher excitation energy the spectrum is dominated by water and the nonresonant pure OH  spectrum can thus not be obtained. However, since the   OH   has similar properties as the pre-edge XAS peak in water, the nonresonant OH  spectrum would be expected to be at 0:5–1 eV higher energy based on the observed shift of the water spectra (Fig. 2 of Ref. [1]). The OH  1b 1 peak would then instead be found near the water d 1 peak, which completely invalidates the argument in Ref. [1]. L. G. M. Pettersson, 1 T. Tokushima, 2 Y. Harada, 2 O. Takahashi, 3 S. Shin, 2,4 and A. Nilsson 1,5 1 FYSIKUM, AlbaNova Stockholm University S-10691 Stockholm, Sweden 2 RIKEN/SPring-8 Sayo-cho, Sayo, Hyogo 679-5148, Japan 3 Department of Chemistry Hiroshima University Higashi-Hiroshima 739-8526, Japan 4 Inst. for Solid State Physics University of Tokyo Kashiwanoha, Kashiwa, Chiba 277-8581, Japan 5 Stanford Synchrotron Radiation Laboratory P.O.B. 20450, Stanford, California 94309, USA Received 13 March 2008; published 20 June 2008 DOI: 10.1103/PhysRevLett.100.249801 PACS numbers: 61.25.Em, 78.70.En, 82.30.Rs [1] O. Fuchs et al., Phys. Rev. Lett. 100, 027801 (2008). [2] O. Bjo ¨rneholm et al., Phys. Rev. Lett. 79, 3150 (1997). [3] M. Odelius et al., Phys. Rev. Lett. 94, 227401 (2005). [4] Ph. Wernet et al., Science 304, 995 (2004). [5] T. Tokushima et al., doi:10.1016/j.cplett.2008.04.077. FIG. 1 (color online). Computed XES spectra with (CHD) or without (t  0) core-hole dynamics. Spectra are weighted with the appropriate exponential lifetime decay probability. PRL 100, 249801 (2008) PHYSICAL REVIEW LETTERS week ending 20 JUNE 2008 0031-9007= 08=100(24)=249801(1) 249801-1  2008 The American Physical Society