Library

Notes: The potential-energy surfaces of the ArClF and ArCl2 complexes are determined by the Hartree-Fock (HF) and Møller–Plesset calculations (up to MP4) in an efficient basis set of 6-31+G(2df ) for the intermolecular energy. The interaction energies are calculated by the supermolecular approach with the full counterpoise corrections for the basis-set superposition error. Three local potential minima are found for ArClF corresponding to the linear Ar–Cl–F and Ar–F–Cl and the asymmetric T-shaped structures. For these the well depths and the distances are D(Ar–Cl–F)=233.5 (MP2) or 219.7 cm−1 (MP4), RArCl=3.38 A(ring); D(Ar–F–Cl)=119.2 (MP2) or 127.2 cm−1 (MP4), RArF=3.3 A(ring); and D(T-shaped)=130.4 (MP2) or 132.6 cm−1 (MP4), RArCl=3.83 A(ring). The results are in accord with the linear ArClF structure as the most-stable structure determined by experiment with the estimate of De=228 cm−1 at RArCl=3.33 A(ring). For the ArCl2 complex, minima are found corresponding to the linear and the T-shaped structures. At the MP2 level the well depths and distances are D(linear)=220.1 cm−1, RArCl=3.5 A(ring); D(T-shaped)=183.6 cm−1, RArCl=3.9 A(ring). Only a small change results at the MP4 level D(linear)=195.3 cm−1, D(T-shaped)=165.2 cm−1. The results for the T-shaped ArCl2 are in good agreement with the experimental results of De=185±1 cm−1 and RArCl=3.8±0.1 A(ring). Estimates for the effects of differences in zero-point energy show the two structures may be of similar stablity.

Notes: The Møller–Plesset perturbation theory up to the complete fourth order is applied to calculate the ab initio van der Waals energy potential of He2. A scheme of constructing a very efficient basis set is proposed based on theoretical considerations. Such a basis set contains a set of Gaussian functions (usually of low angular momentum) centered at the midbond of the van der Waals molecule, which have been proven very efficient to reproduce the intersystem correlation interaction energy normally to be achieved by use of the nucleus-centered high angular momentum polarization functions. A nuclear-centered moderate set [5s4p2d] augmented by a set of midbond functions such as {3s3p2d} produces a value of 10.04 K for the well depth of He2 at the complete fourth-order (MP4) theory. Results from a series of other basis sets suggest that this value is nearly saturated with the further increase of basis, which is in agreement with the estimate of the complete basis-set limit. Calculations of the potential at the minimum and the other distances (3.0–10.0 a0) confirm that the MP4 method, together with the use of midbond functions, is a reliable and convenient choice for the accurate calculations.

Notes: The equilibrium structure of the NH3 dimer is investigated using large, efficient basis sets, 6–311+G(3d,2p) and [7s5p3d,4s1p] extended with bond functions, at the second-order Møller–Plesset perturbation approximation (MP2) and higher levels. Intermolecular energies and optimized dimer structures are obtained with the full counterpoise correction for the basis set superposition error. The stabilities of two possible equilibrium structures, one containing a nearly linear hydrogen bond with Cs symmetry and the other a cyclic configuration with C2h symmetry, are examined. In a basis without bond functions, the Cs structure is found more stable. As bond functions are added, however, the C2h structure becomes more stable. This establishes the importance of the dispersion energy which is disproportionally underestimated for the C2h structure in a purely nucleus-centered basis. The stability of the C2h structure relative to the Cs is retained at the higher levels up to the complete forth order (MP4SDTQ). The minimum energy path connecting the two equivalent Cs structures via the C2h structure is calculated. The resulting potential curves are extraordinarily flat in a broad region around the C2h structure but rise steeply upon approaching the Cs structure containing a nearly linear hydrogen bond, indicating that the donor–acceptor interchange barrier is absent in the NH3 dimer. The equilibrium structure for the NH3 dimer found in the present study probably has the cyclic form with C2h symmetry.

Notes: The vibrational dependence of the intermolecular potential of Ar–HF is investigated through the spectra of levels correlating with HF(v=3). We have previously reported measurements of the (vbKn)=(3000), (3100), and (3110) levels of Ar–HF using intracavity laser-induced fluorescence in a slit supersonic jet [J. Chem. Phys. 98, 2497 (1993)]. These levels are found to be well reproduced (within 0.1 cm−1) by the Ar–HF H6(4,3,2) potential [J. Chem. Phys. 96, 6752 (1992)]. The second overtone experiments are extended to include the (3002) state which is coupled to (3110) through Coriolis interaction, and the (3210) state which is more sensitive to higher-order anisotropic terms in the potential. The observations establish that the level (3002) lies 0.229 cm−1 below (3110), with upper state rotational constant B=0.085 89 cm−1. This is in good accord with the predictions of the H6(4,3,2) potential. The (3210) state lies at 11 484.745 cm−1 with B=0.099 79 cm−1. The band origin is 1.7 cm−1 higher than predicted, and thus contains important new information on the vibrational dependence of the potential. Several detailed features of the spectra can be explained using the H6(4,3,2) potential. The Q-branch lines of the (3210)←(0000) band show evidence of a weak perturbation, which can be explained in terms of mixing with the (3112) state. The (3210) spectrum exhibits parity-dependent rotational predissociation and the widths of the P- and R-branch lines and the magnitude of the l-type doubling can be explained in terms of coupling to the (3200) state, which is estimated to lie 4 cm−1 below the (3210) state.The Q-branch lines show a predissociation cutoff above Q(16); this is in reasonable agreement with the predictions of the H6(4,3,2) potential, but suggests that the binding energy calculated for the potential may be about 1 cm−1 too large. To examine the potential further, high-level ab initio calculations are performed, with an efficient basis set incorporating bond functions. The calculations give a well depth of 92%–95% of that of the H6(4,3,2) potential at θ=0° for v=0 and v=3, respectively; this is in line with earlier results on rare gas pairs. The calculations also reproduce the anisotropy of the H6(4,3,2) potential and its vibrational dependence. The dependence of the intermolecular potential on HF bond length is found explicitly.

Notes: The effect of bond functions and the related problems are studied by performing the fourth-order Møller–Plesset perturbation calculations for the dissociation energy of the ground state F2 molecule, using a series of basis sets systematically extended with polarization functions and bond functions. The results show the usefulness of bond functions if the basis set superposition errors (BSSE) are corrected by the counterpoise method. A new interpretation of BSSE effect and the counterpoise method is given.

Notes: An intermolecular potential energy surface for the dimer of hydrogen chloride in the ground state is calculated at the levels of the second-order (MP2) and fourth-order (MP4) Møller–Plesset approximations using a large basis set containing bond functions. The surface is characterized by the minimum energy pathway through two equivalent hydrogen-bonded structures. The hydrogen-bonded equilibrium geometry has the centers of mass distance Rm=3.78 A(ring) and polar angles θ1=8.0° and θ2=90.0° (at MP2 level). The well depth at the hydrogen-bonded minimum is Vm=−710.9 cm−1 at MP2 and Vm=−643.9 cm−1 at MP4 level. The interchange barrier between the two equivalent minima occurs at R=3.68 A(ring), θ1=θ2=46.0°, with the barrier height of 58.6 cm−1 at MP2 and 45.9 cm−1 at MP4 level (with the MP2 geometries). These results are in good agreement with a new empirical potential of Elrod and Saykally. Our calculations show that the bonding in the HCl dimer is dominated by the dispersion forces, which is different from the bonding in other classical hydrogen-bonded systems such as the hydrogen fluoride dimer and the water dimer. © 1995 American Institute of Physics.

Notes: Two forms for the HeCl2 potential are compared to the available experimental data. First, an atom–atom form that incorporates the recently measured anisotropic He–Cl potential is used. The anisotropy of this potential is slight, and its strengths and weaknesses are similar to previous potentials in which the He–Cl interaction was treated as isotropic. In particular, the fit to the scattering data is poor. Second, a fit to ab initio points calculated using Møller–Plesset perturbation theory to fourth order was performed. The resulting potential is much more anisotropic than any potential previously proposed and tested for HeCl2. This potential fits the rotationally resolved excitation spectra as well as do previous empirical potentials, and is consistent with certain features of the total differential scattering data with which previous potentials were not. Although the ab initio potential has a global minimum in the linear configuration, the probability distribution of the ground vibrational level still maximizes in the perpendicular configuration, accounting for the good fit to the rotationally resolved spectrum. We conclude that noble gas–halogen potentials are much more anisotropic than previously believed, and we suggest several experiments that could help to confirm this anisotropy. © 1995 American Institute of Physics.

Notes: The method of the bond function basis set combined with the counterpoise procedure is studied in detail by the complete fourth-order Møller–Plesset perturbation (MP4) theory, following from a recent communication report [J. Chem. Phys. 98, 2481 (1993)]. This method is applied to calculate molecular dissociation energies De as well as equilibrium bond distances re and harmonic frequencies ωe of a number of diatomic molecules (N2, O2, F2, Cl2, HF, HCl, and CO) and the results are compared with those from other methods, without either counterpoise procedure or bond functions or both. The usefulness of the method is shown by the results for all the molecules using a moderately polarized basis set (2p1d for H atom and 2d1f for heavy atoms) augmented with the universal bond functions 3s3p2d. The method has consistently recovered 98%–99% of the experimental values for De, compared to as low as 90% without bond functions. The effect of bond functions is less significant on the predictions of re and ωe, due primarily to the inadequacy of the MP4 theory, but our method is still shown to be favored over the other methods. The electric dipole moments of the polar molecules (HF, HCl, and CO) are also examined and it is found that the use of bond functions results in a significant improvement of the dipole values. Detailed discussions are given to explain the need for bond functions and the counterpoise procedure. The high linear independence with nucleus-centered basis functions is explained to be responsible for the efficiency of bond functions. The counterpoise procedure is logically justified from the conventional noncounterpoise procedure. Potential problems and limitations associated with the proposed method are also discussed.