Search

The bond dissociation energies of CH4, C2H2, and their fragments are investigated using basis set extrapolations and high levels of correlation. The computed bond dissociation energies (De) are accurate to within 0.2 kcal/mol. The agreement with the experimental (D0) values is excellent if we assume that the zero-point energy of C2H is 9.18 kcal/mol. The effect of core (carbon 1s) correlation on the bond dissociation energies of C–H bonds is shown to vary from 0.2 to 0.7 kcal/mol and that for C–C bonds it varies from 0.4 to 2.2 kcal/mol. © 1995 American Institute of Physics. [ABSTRACT FROM AUTHOR]

The dissociation energies of CN and C2 are determined with multireference configuration- interaction calculations utilizing very large Gaussian basis sets. Valence level one-particle errors, such as basis set incompleteness and basis set superposition error, are essentially eliminated by extrapolation. The errors in the correlation, or n particle, treatment are considered in detail. The effect of 1s correlation is computed using the averaged coupled-pair functional approach. Our best estimate for C2 is in very good agreement with the recent experimental result by Urdahl et al. Our best D0 value for CN is 7.72±0.04 eV. This is larger than previous theoretical values and in good agreement with the recent experimental result (7.738±0.02 eV) of Huang et al. The origin of the differences between the current D0 and our previous theoretical value is discussed. [ABSTRACT FROM AUTHOR]

The effect of N 1s correlation on the spectroscopic constants of N2 is studied using the coupled-cluster singles and doubles approach with a perturbational estimate of the connected triples [CCSD(T)] and internally contracted multireference configuration interaction (ICMRCI) techniques. At the ICMRCI level, we obtain a 1s effect of +1.35 kcal/mol on the dissociation energy. However, the effect is found to be smaller when size-extensive methods are used. The 1s effects computed at the CCSD(T) and internally contracted averaged coupled-pair-functional (ICACPF) levels are in excellent agreement. Our best estimate for the effect of 1s correlation is +0.8 kcal/mol on De, +9 cm-1 on ωe, and -0.002 Å on the bond length. Including our estimate for the effect of 1s correlation, we obtain a De of 227.1 using a correlation-consistent polarized-valence sextuple zeta basis set at the ICMRCI+Q level (where the +Q indicates that the Davidson correction has been included). Basis set incompleteness, which is estimated to be 0.7±0.2 kcal/mol, is still the major source of error. The CCSD(T) ten-electron results are found to be in excellent agreement with those obtained at the ICACPF or ICMRCI+Q levels of theory. [ABSTRACT FROM AUTHOR]

Ab initio results are presented for the low-lying bound states of O2 that dissociate to ground-state atoms. Spectroscopic constants and dissociation energies (De) are reported for the X 3Σ-g, a 1Δg, b 1Σ+g, c 1Σ-u, A’ 3Δu, A 3Σ+u, 3Πu, 5Σ-u, 5Πu, and 5Πg states. For the first six states, which have been experimentally characterized, we obtain accurate results at the multireference configuration-interaction level after correction for higher excitations. For example, our computed De value for the X 3Σ-g state corrected for basis-set superposition error is only 0.7 kcal/mol less than the experimental value. The 5Πg state is estimated to have a De of 0.16±0.03 eV, suggesting that the importance of this state in the nightglow should perhaps be reconsidered. [ABSTRACT FROM AUTHOR]

The bonding in the X 1Σ+g state of Mg2 is investigated using near-complete valence one-particle Slater and Gaussian basis sets containing up to h functions. Full configuration interaction (FCI) calculations are used to calibrate four-electron correlation treatments. We show that the four-electron complete CI limit can be approached using a sequence of either second-order CI (SOCI) or interacting correlated fragment (ICF) calculations. At the valence level, our best estimate of the dissociation energy (De) is 464 cm-1. We show that this is a lower limit and probably within 5 cm-1 of the complete basis value. The inclusion of core–valence correlation using a model operator approach decreases De by about 35 cm-1 and increases the bond length by 0.03 a0, thereby yielding spectroscopic constants in good agreement with experiment. Attempts to compute the core–valence effect accurately by expanding the CI treatment were unsuccessful. [ABSTRACT FROM AUTHOR]

Theoretical spectroscopic constants are reported for the A’ 5Σ+g and C‘ 5Πu states of N2 based on CASSCF/MRCI calculations employing large ANO Gaussian basis sets. Our calculated A’ 5Σ+g potential differs qualitatively from previous calculations in that the inner well is significantly deeper (De ≊3450 cm-1). The potential also has a substantial barrier (≊500 cm-1) to dissociation with a maximum near 4.3 a0, and then a shallow (≊47 cm-1) van der Waals minimum near 6.3 a0. We suggest that it is the v=4 and v=5 levels that are involved in predissociating the a 1Πg and B 3Πg states, as opposed to v=0 and v=1 proposed by Verma. The deeper well in the A’ 5Σ+g state provides considerable support for the theory of Berkowitz, Chupka, and Kistiakowsky who proposed that A’ 5Σ+g is the primary precursor state involved in the yellow Lewis–Rayleigh afterglow of nitrogen. The theoretical De for the C‘ 5Πu state is also much larger than previous estimates. The previously unassigned Hermann infrared system (HIR) band positions correspond exceptionally well to our theoretical values for the C‘ 5Πu →A’ 5Σ+g transition. This and other considerations lead to a rather convincing assignment of this transition as the HIR system. Einstein coefficients and radiative lifetimes are presented for this transition. Since the natural radiative lifetime of this state is much longer than experimental estimates, the C‘ 5Πu state may derive its actual lifetime through interaction with the C 3Πu state. [ABSTRACT FROM AUTHOR]

The metal–carbon bond dissociation energies (D0) and geometries for the first- and second-row transition-metal methyl neutrals and positive ions are determined. The computed D0 values for the positive ions compare favorably with experiment, except for RuCH+3, RhCH+3, and PdCH+3 where the experimental values are 10–15 kcal/mol larger. The computed D0 values for the hydride and methyl positive ions are similar for all metals in both transition rows except for Cu and Ag. However, for the neutral systems the D0 values for the methyls are smaller, especially on the right-hand side of both transition rows where the differences approach 15 kcal/mol. In general, the dissociation energies do not follow simple trends, as the individual D0 values are significantly affected by the relative spacings between the atomic states of the metal. The study of all of the methyl neutral and ions of both transition rows presented here provides a consistent set of data for the dissociation energies, thereby allowing a critical assessment of the experimental data for these molecular species, and an enhanced understanding of the different bonding mechanisms. [ABSTRACT FROM AUTHOR]

Theoretical dissociation energies for the first- and second-row transition metal hydride positive ions are critically compared against recent experimental values obtained from ion beam reactive scattering methods. Theoretical spectroscopic parameters (re,ωe) and dipole moments are presented for the ground and several low-lying excited states. The calculations employ large Gaussian basis sets and account for electron correlation using the single-reference single- and double-excitation configuration interaction and coupled-pair-functional methods. The Darwin and mass-velocity contributions to the relativistic energy are included in the all-electron calculations on the first-row systems using first-order perturbation theory, and in the second-row systems using the Hay and Wadt relativistic effective core potentials. Owing to the smaller correlation effects in the second-row systems, the theoretical D0 values are probably at least as accurate as for the first row. Therefore, the theorectical D0 values for the second-row transition metal hydride positive ions should provide a critical measure of the experimental values, which are not as refined as many of those in the first transition row. [ABSTRACT FROM AUTHOR]

Ab initio calculations at the self-consistent field and singles plus doubles configuration-interaction level are used to determine accurate spectroscopic parameters (De,re,ωe) for the 2Π and 2∑+ states of the alkali oxides and the a 3Π (or A 1Π) states of the alkaline-earth oxides. Numerical Hartree–Fock (NHF) calculations performed on KO demonstrate that the extended Slater basis sets employed are near the Hartree–Fock limit. When the dissociation is referenced to the ionic limits, the differential correlation contribution is found to increase the D0 by a relatively small amount (0.0–0.2 eV). An accurate description of the alkali and alkaline-earth oxides requires correlating both the oxygen and metal electrons. The theoretical dissociation energies (D0) permit a critical assessment of the experimental literature and allow us to recommend D0 values that are accurate to 0.1 eV for all systems considered. There is a strong correlation between the dissociation energy (to ions) and re, because the bonding is predominantly electrostatic in origin. Theoretical 2Π–2∑+ energy separations are presented for the alkali oxides. An extensive study of the 2Π–2∑+ energy separation in KO reveals a 2∑+ ground state at all levels of theory. Basis set studies in combination with NHF calculations indicate different basis set requirements for the 2Π and 2∑+ states. In the NHF limit the 2∑+ state of KO is lower by about 250 cm-1. The separation is almost unaffected when the 15 valence electrons are correlated at the singles plus doubles level using an extended Slater basis. [ABSTRACT FROM AUTHOR]

Ab initio calculations using self-consistent-field and correlated wave functions are used to determine accurate spectroscopic parameters (re, ωe, D0) for the alkali and alkaline-earth monofluorides and monochlorides. Numerical Hartree–Fock (NHF) calculations are performed on selected systems to ensure that the extended Slater basis sets employed are near the Hartree–Fock limit. Because the bonding is predominantly electrostatic in origin, there is a strong correlation between the dissociation energy (to ions) and re. By dissociating to the ionic limits, most of the differential correlation effects can be embedded in the accurate experimental electron affinities and ionization potentials. With this model, correlation effects are relatively small (0.0–0.4 eV), but invariably increase D0. The importance of correlating the electrons on both the anion and the metal is discussed. For the heavier alkali chlorides, RbCl and CsCl, we show that a core-valence treatment, which excludes the double excitations out of the metal (n-1) shell, gives significantly better re and D0 parameters than does the full singles plus doubles configuration-interaction CI(SD) calculation. The theoretical studies combined with the experimental literature allow us to recommend D0 values that are accurate to 0.1 eV for all systems considered. [ABSTRACT FROM AUTHOR]

The binding energy for the X 2Π ground state of MgO+ has been determined using the coupled-cluster singles and doubles approach with a perturbational estimate of the triples and the modified coupled-pair functional approach in conjunction with very large basis sets. Our best value of 2.32±0.06 eV is in excellent agreement with our previous result and with the photodissociation result of Freiser and co-workers, but is smaller than the recent value determined using the guided-ion beam method. [ABSTRACT FROM AUTHOR]