A number of hydrogen-bond related quantities-geometries, interaction energies, dipole moments, dipole moment derivatives, and harmonic vibrational frequencies-were calculated at the Hartree-Fock, MP2, and different DFT levels for the HCN dimer and the pe
The influence of uniform and non-uniform electric fields on one-dimensional proton transfer curves for (H2O)(2), H5O2+ and H3O2- has been examined using quantum-mechanical ab initio calculations. Both liquid-state and solid-state environments are discussed. For the charged complexes the transfer barrier is removed or greatly reduced by a field as small as 0.005 a.u. (2.5 X 10(7) V/cm). Local field fluctuations of this size are easily produced in condensed aqueous systems at room temperature. For the asymmetric single-well potential of an (H2O)(2) complex, a field ten times larger is needed to move the minimum from one side to the other across the O ... O bond. Such local fields can be achieved in ionic aqueous systems. The energy barrier for proton transfer in ice Ih has been computed using a periodic Hartree-Fock approach; the barrier for a fully concerted proton transfer is similar to 60 kJ/mol.
Vibrational spectra for the O-H stretching motion of HDO molecules in different surroundings have been calculated by quantum mechanical ab initio methods and compared with experimental spectra. The free water molecule, water chains, and ion-water clusters are discussed. Solvent effects on OH vibrations in liquid water have been calculated as well as "in-crystal" OH frequencies in some ice and ionic crystalline hydrate structures. The importance of nonadditivity effects, electron correlation (at the MP2 level), and long-range interactions for the total frequency downshift is demonstrated. It is shown that the inclusion of these effects, in conjunction with a variational quantum mechanical treatment of the anharmonic vibrational stretching motion (force constants up to the fourth order), yields vibrational frequencies in quantitative agreement with experiment for a wide range of aqueous systems.
Ab initio studies of the uncoupled, anharmonic OH and OD stretching frequency shifts in the three proton-ordered ice phases known, ice II, ice VIII, and ice IX, are presented. The ice structures are simulated by (H2O)5 supermolecules surrounded by point charges representing the correct crystal potentials. The calculations include electron correlation at the MP2 (DZP) level. For the eight different OH (OD) vibrators studied, the crystal environment leads to a downshift of the anharmonic OD frequency in the range 195-265 cm-1, in good agreement with experimental values (222-281 cm-1) when corrections are made for the limited supermolecular size (approximately - 45 cm-1), and, for ice VIII, also for the effects of the nonhydrogen bonded network (approximately + 75 cm- 1). Also the agreement between absolute experimental and theoretical OD frequencies is good when errors due to basis set limitation (approximately - 75 cm-1 ) are taken into account. The calculations suggest a reassignment of two of the experimental OD bands in ice II and all three experimental OD bands in ice IX. Calculations for charge-embedded (H2O)9 and (H2O)13 ice clusters show that at least a nonamer is needed to avoid boundary effects from the size of the supermolecule. Theoretical correlation curves between H-bond parameters-R (O ... 0), nu(OH), r(e)(OH), and infrared absorption intensity-are presented for the three ice phases and are compared to liquid water computations.
Many-body interaction energies and anharmonic OH stretching frequencies have been calculated for water in chain formations, in a ring structure, and in a tetrahedral arrangement. The calculations were of ab initio type, with the electron correlation energy included by Moller-Plesset perturbation correction to second order (MP2) and the basis-set superposition error corrected by the counterpoise procedure. The maximum chain length was seven water molecules, and the ring was five-membered. The molecules were H-bonded head-to-tail. The two- and many-body energies for the chains and ring are all of the same sign (negative), indicating strong cooperativity. The total nonadditivity contribution to the interaction energy is large, 16% for the longest chain and over 18% for the ring. The interaction energy of ah individual chain member with the rest of the chain shows even larger nonadditivity: over 25% for a molecule in the middle of the chain. This quantity should be of relevance for molecular dynamics simulations of liquid water. The OH stretching frequency downshift increases for all members of the chain with increasing chain length and is larger for molecules in the interior of the chain (-357 cm(-1) for the middle molecule in the 7-chain) than for terminal water molecules. The frequency converges only slowly for water molecules in the interior but faster for terminal water molecules. ''Frequency cooperativity'' was investigated by calculating many-body contributions in a manner analogous to the energy calculations. The chain and ring exhibit strong cooperativity. Infrared absorption intensities and charge transfer were investigated.
The "in-crystal" frequency of the anharmonic and uncoupled OH stretching vibration of HDO molecules in LiClO4.3H2O(s) has been calculated by quantum-mechanical ab initio and model potential methods and compared with the experimental infrared frequency from isotope-isolated HDO molecules. The effects of the nearest neighbours as well as of the crystalline environment have been investigated by the two computational techniques. In both cases, the one-dimensional potential for an anharmonic OH oscillator was constructed from point-wise energy calculations and the Schrodinger equation for the protonic motion in this potential well was solved by a variational procedure. In the ab initio calculations, vibrational potentials were constructed from RHF and MP2 type calculations of point-charge embedded ClO4-.HDO and (Li+)2.(ClO4-)2.HDO clusters using DZP and TZP basis sets. For the LiClO4.3H2O(s) crystal, the ab initio OH frequency is in close quantitative agreement with experiment when electron correlation by MP2 and the crystal field are included: 3537 cm-1 (MP2(TZP)) versus the experimental value of 3556 cm-1. Inclusion of the crystal field is essential and can in this crystal be satisfactorily represented by Ewald field-consistent point charges outside the hydrogen-bonded ClO4-...HDO cluster. In the model potential calculations, analytical intermolecular pair potential functions from the literature were used in conjunction with an experimental intramolecular potential function for the OH stretching motion. The particular intermolecular model chosen here yields an absolute OH frequency 160 cm-1 below experiment. These calculations exemplify some of the difficulties encountered when employing analytical model potentials in vibrational studies.
Ab initio quantum-mechanical calculations of anharmonic frequencies for the water O-H vibrations have been performed for a series of crystalline hydrates. In each case, the potential-energy curve for the uncoupled water O-H stretch was derived at the Moller-Plesset MP2 level. Nearest neighbors of the water molecule were explicitly included in the supermolecule and the rest of the surroundings were mimicked by point charges to reproduce the crystal field out to infinity. The time-independent Schrodinger equation for the motion of the proton in this potential well was solved variationally and the frequency was obtained from the energy difference between the 0 and 1 eigenstates. Computed frequencies can be directly compared with existing infrared data for isotope-isolated water molecules in these hydrates. The compounds selected (LiClO4.3H2O, LiHCOO.H2O, LiOH.H2O) exhibit experimental O-H frequency shifts in a wide range, from - 150 down to - 930 cm-1. Good agreement is found between experimental and theoretical frequencies (experimental values in parentheses): 3596 (3556) for LiClO4.3H2O, 3129 (3112) and 3488 (3390) for LiHCOO.H2O, and 2817 (2775) cm-1 for LiOH.H2O. Correlation curves of typical H-bond parameters such as nu(O-H) vs R(H...O), r(e)(O-H) and nu(OH)/nu(OD) have been computed and compared with experiment. The vibrational intensities are also discussed.
The proton-ordered ice VIII structure has been investigated by ab initio periodic Hartree-Fock calculations in the pressure interval from 0 to 30 GPa using a 6-31G** basis set, The structure was optimized by energy-minimization at different volumes, and from the resulting energy vs volume relationship; the equation of state of ice VIII was derived; The variation of the structure,intramolecular geometry, Mulliken charges, electron density, Raman spectrum, and infrared stretching vibrations with varying pressure were investigated. The agreement with existing experimental data is generally good. Nearest-neighbor hydrogen-bonded O...O distances decrease from 2.88 to 2.57 Angstrom as the pressure is increased from 0 to 30 GPa. For the same: pressure range, the intramolecular OH bond increases from 0.951 to 0.955 Angstrom (giving a dr(OH)/dP value of 0.000 14 Angstrom/GPA), the Mulliken charge on H increases from +0.386 to +0.452, the calculated bulk modulus increases from similar to 25 to similar to 160 GPa), (corresponding experimental values ire similar to 25 at 2.4 GPa and similar to 135 at 30 GPa), and the electron density redistribution is considerably enhanced. The frequency downshift of the OH stretching vibration varies from -200 cm(-1) at 2.4 GPa to -500 cm(-1) at 20 GPa; the corresponding experimental values are -300 and -650 cm(-1). Electronic density-of-states diagrams are presented.
The hydrate crystal lithium hydroxide monohydrate LiOH.H2O has been studied by ab initio periodic Hartree-Fock calculations. The influence of the crystalline environment on the local molecular properties (molecular geometry, atomic charges, electron density, molecular vibrations and deuterium quadrupole coupling constants) of the water molecule, the lithium and hydroxide ions has been calculated. A number of crystalline bulk properties are also presented, optimized crystalline structure, lattice energy and electronic band structure. The optimized cell parameters from calculations with a large basis set of triple-zeta quality differ by only 1-3% from the experimental neutron-determined cell, whereas the STO-3g basis set performs poorly (differences of 5-10%). With the triple-zeta basis also the atomic positions and intermolecular distances agree very well with the experiment. The lattice energy differs by approximately 8% from the experimental value, and by at most 3% when a density-functional electron correlation correction is applied. Large electron-density rearrangements occur in the water molecule and in the hydrogen bond and are in qualitative and quantitative agreement with experimental X-ray diffraction results. The quadrupole-coupling constants of the water and hydroxide deuterium atoms are found to be very sensitive to the O-H bond length and are in good agreement with experimental values when the calculation is based on the experimental structure. The anharmonic O-H stretching vibrations in the crystal are presented and found to be very close to results from calculations on molecular clusters. The electronic band and density-of-states spectra are discussed. Model calculations on a hydrogen fluoride chain were used to rationalize the results.
OH stretching frequencies of HDO molecules in liquid water have been calculated by molecular dynamics simulation and compared to quantum-corrected OH stretching frequencies. In the MD simulation the MCY intermolecular water-water potential was used together with an experimental intramolecular free water potential. The frequencies calculated classically by Fourier transformation of the velocity autocorrelation function are found to be almost-equal-to 300 cm-1 too high compared with experiment. Quantum corrections show that the classical error contribution to this discrepancy is almost-equal-to 140 cm-1. To reach full agreement with experiment also the potential model needs to be improved. It is suggested that in constructing flexible water potentials the goal should not be set for an MD-derived OH frequency in absolute agreement with experiment (at 3400 cm-1) but instead some 200 cm-1 higher.
Band shapes of liquid water OH vibrational spectra oblained from molecular dynamics (MD) simulation and from a quantum-mechanical method are investigated. The so-called "frozen-field approximation" applied to the calculation of quantum-mechanical high-frequency vibrational spectra is critically examined. It is demonstrated that the band width of the OH stretching spectrum is seriously overestimated through the neglect of the dynamics of the environment in the frozen-field approximation. We show that the proper inclusion of the dynamics in this quantum-mechanical method leads not only to a correct absolute frequency for the model potential used, but also to the correct description of the band width. The basic steps in this method are: (1) an MD simulation yielding an ensemble of liquid water configurations, (2) a quantum-mechanical uncoupled local-mode calculation of the OH frequency for each molecule, using model potentials for the inter- and intra-molecular interactions, (3) inclusion of the influence from the dynamics of the surroundings by filtering out rapid frequency fluctuations. The remaining discrepancy between experimental and computed OH spectra is attributed to shortcomings in the potential model used.
Uncoupled OH and OD stretching bands of HDO molecules have been calculated for an ionic aqueous solution, based on the trajectories from a classical statistical-mechanical computer simulation and subsequent quantum-mechanical calculations of the vibrational energy levels. Each V(r(OH)) potential function has been constructed as a sum of intra- and intermolecular energies, where different intermolecular water-water potential functions from the literature (MCY, TIPS2, RWK2 and CF2) have been tested in conjunction with the experimentally derived HMS intramolecular potential. In this way, vibrational densities-of-states as well as infrared absorption bands have been calculated for HDO molecules in the bulk and in the ionic hydration shells (Li+, HCOO-). Calculated frequencies and band widths for the TIPS2 and MCY potentials are fairly close to experimental values. The calculated OH shift between the gas and liquid water phases is - 303 cm-1 with the TIPS2 potential, as compared to the experimental value of - 307 cm-1. The MCY potential gives - 260 cm-1, while RWK2 as well as the CF2 potentials give rise to a non-negligible number of spurious frequencies. Water molecules in the first hydration shell of Li+ exhibit only slightly lower stretching frequencies than bulk water. The frequencies of the OH and OD groups of HDO molecules bonded to the formate oxygen atoms are lower than in bulk water, while the frequency of the OH/OD group pointing away from the formate ion is higher compared to bulk water.