The spectrum of water vapor is of fundamental importance for a variety of processes, including the absorption and retention of sunlight in Earth's atmosphere. Therefore, there has long been an urgent need for a robust and accurate predictive model for this spectrum. In our work on the high-resolution spectrum of water, we report first-principles calculations that approach experimental accuracy. To achieve this, we performed exceptionally large electronic structure calculations and considered a variety of effects, including quantum electrodynamics, which have routinely been neglected in studies of small many-electron molecules. The high accuracy of the resulting ab initio procedure is demonstrated for the main isotopomers of water.
The high accuracy ab initio adiabatic potential energy surfaces (PESs) of the ground electronic state of the water molecule, determined originally by Polyansky et al. [Science 299, 539 (2003)] and called CVRQD, are extended and carefully characterized and analyzed. The CVRQD potential energy surfaces are obtained from extrapolation to the complete basis set of nearly full configuration interaction valence-only electronic structure computations, augmented by core, relativistic, quantum electrodynamics, and diagonal Born-Oppenheimer corrections. We also report ab initio calculations of several quantities characterizing the CVRQD PESs, including equilibrium and vibrationally averaged (0 K) structures, harmonic and anharmonic force fields, harmonic vibrational frequencies, vibrational fundamentals, and zero-point energies. They can be considered as the best ab initio estimates of these quantities available today. Results of first-principles computations on the rovibrational energy levels of several isotopologues of the water molecule are also presented, based on the CVRQD PESs and the use of variational nuclear motion calculations employing an exact kinetic energy operator given in orthogonal internal coordinates. The variational nuclear motion calculations also include a simplified treatment of nonadiabatic effects. This sophisticated procedure to compute rovibrational energy levels reproduces all the known rovibrational levels of the water isotopologues considered, H(2) (16)O, H(2) (17)O, H(2) (18)O, and D(2) (16)O, to better than 1 cm(-1) on average. Finally, prospects for further improvement of the ground-state adiabatic ab initio PESs of water are discussed.
A valence-only ͑V͒ dipole moment surface ͑DMS͒ has been computed for water at the internally contracted multireference configuration interaction level using the extended atom-centered correlation-consistent Gaussian basis set aug-cc-pV6Z. Small corrections to these dipole values, resulting from core correlation ͑C͒ and relativistic ͑R͒ effects, have also been computed and added to the V surface. The resulting DMS surface is hence called CVR. Interestingly, the C and R corrections cancel out each other almost completely over the whole grid of points investigated. The ground-state CVR dipole of H 2 16 O is 1.8676 D. This value compares well with the best ab initio one determined in this study, 1.8539Ϯ 0.0013 D, which in turn agrees well with the measured ground-state dipole moment of water, 1.8546͑6͒ D. Line intensities computed with the help of the CVR DMS shows that the present DMS is highly similar to though slightly more accurate than the best previous DMS of water determined by Schwenke and Partridge ͓J. Chem. Phys. 113, 16 ͑2000͔͒. The influence of the precision of the rovibrational wave functions computed using different potential energy surfaces ͑PESs͒ has been investigated and proved to be small, due mostly to the small discrepancies between the best ab initio and empirical PESs of water. Several different measures to test the DMS of water are advanced. The seemingly most sensitive measure is the comparison between the ab initio line intensities and those measured by ultralong pathlength methods which are sensitive to very weak transitions.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.