Index of content:
Volume 104, Issue 9, 01 March 1996

Theoretical calculations of zero‐temperature absorption spectra of Li in solid H_{2}
View Description Hide DescriptionWe employ variational and diffusionMonte Carlo methods to calculate the ground stateproperties of both pure and Li dopedsolid H_{2}. The absorption spectra of Li in the H_{2}solid are calculated from the sampled ground state configurations and a pseudothermal semiclassical spectral line theory, based on the Franck–Condon principle for a condensed phase system. The T=0 numerical results of various thermodynamic properties of the pure and dopedsolid agree well with those from previous theoretical calculations. However, the Li spectra differ significantly from results of a finite temperature path integral Monte Carlo calculation [D. Scharf et al., J. Chem. Phys. 99, 9013 (1993)], which employed a different Li^{*}–H_{2} potential. The differences in two calculations are extensively discussed here, and comparisons are also made with available experimental data. We demonstrate that accurate measurements of the Li absorptionspectrum provide a powerful means to probe the local structures of the Li trapping sites.

Observation of the β‐methylallyl radical by resonance Raman spectroscopy
View Description Hide DescriptionThe resonanceRaman spectrum of the β‐methylallyl (2‐methyl‐allyl) radical is detected with excitation from 223 nm to 243 nm. The first observation of four fundamentals, the symmetric CH_{2} rock, the symmetric CCC stretch, the C–CH_{3} stretch, and the CCC bend, as well as an overtone, the symmetric CH_{2} twist of the β‐methylallyl radical are reported. The assignments of the resonanceRaman spectra are based upon the previous force field calculation, and comparisons with molecules having similar characteristic vibrations. Results from the resonanceRaman spectra are compared with frequencies obtained by infrared spectroscopy using matrix isolation and our previous work on the allyl radical.

Spin dynamics under the Hamiltonian varying with time in discrete steps: Molecular dynamics‐based simulation of electron and nuclear spin relaxation in aqueous nickel(II)
View Description Hide DescriptionA method of calculating the time correlation functions for electron spin is proposed, based on solving the time‐dependent Schrödinger equation for a spin Hamiltonian that contains a term varying randomly in discrete time steps. It is applied to the study of electron spin relaxation in aqueous solution of nickel(II) ions with S=1. The random term in the spin Hamiltonian in this case is the zero‐field splitting (ZFS) interaction. The method is evaluated by an application to a model system (the pseudorotation model) for which an analytical solution to the electron spin relaxation problem is known. The same method is then employed to study the electron and nuclear spindynamics in a system where the time variation of the zero‐field splitting is obtained by a combination of ab initio quantum chemistry and molecular dynamics simulations.

Intracavity laser absorption spectroscopy of HOCl overtones. I. The 3v _{1}+2v _{2} band and numbers of vibrational states
View Description Hide DescriptionThe near infrared high‐resolution spectra of the a‐type transitions of the weak 3v _{1}+2v _{2} combination band of transient HO^{35(37)}Cl at 12 600 cm^{−1} has been recorded in an ultrasensitive titanium:sapphire intracavity laser absorption spectrometer (ICLAS). We report line assignments, new and refined anharmonicity parameters, and the spectroscopic constants for the excited rovibrational states of 3v _{1}+2v _{2}. The Fermi resonance perturbations in this five quanta region, where the internal energy of the molecule is already more than 2/3 of the dissociation energyE _{0}, remain localized and they are the exception, while the extent of intermode mixing and thus intramolecular vibrational energy distribution (IVR) seems to be still small. A Dunham expansion is used for band origin predictions and representations of vibrational statesN(E) of HO^{35}Cl up to the dissociation threshold. The results are compared with harmonic and anharmonic numbers of states from a recently proposed stretch–bend coupling model.

Simulation of the charge transfer absorption of the H_{2}O/O_{2} van der Waals complex using high level ab initio calculations
View Description Hide DescriptionModel calculations of the UVcharge transfer(CT)absorptionspectrum of a water–oxygen collision pair are reported. The motivation is an assessment of this process as the initial stage of a new nucleation mechanism for atmospheric water vapor. Ab initio MP2(full)/6‐311++G(2d,p) geometries of ground state van der Waals dimers and (complete active space self‐consistent field) CAS‐SCF(10,7)/6‐311++G(2d,p) geometries of the CT state are detailed. MR‐ACPF/ 6‐311++G(2d,p) potential surfaces have been calculated along the intermolecular coordinate. The bound and continuum nuclear wave functions have been calculated numerically on these potential curves and, with the MR‐CISD/6‐311++G(2d,p) electronic transition dipole moment integral, used to calculate the unary and binary absorption coefficients. The implications of the results for explaining water photonucleation in the presence of oxygen are discussed.

Experimental studies of the vapor phase nucleation of refractory compounds. IV. The condensation of magnesium
View Description Hide DescriptionMagnesiumnucleation was studied over the range of approximately 700 to 950 K in a gas evaporation apparatus. Measured supersaturation ratios ranged from approximately 37 to 4.2 over this temperature range, respectively. A comparison of these data and Classical Nucleation Theory shows that the two are not consistent. Although there is a good correlation between the supersaturation and the temperature data when plotted in accordance with Scaled Nucleation Theory, some of the derived parameters are slightly below the limits predicted by the theory.

Fine‐structure transitions in metastable Ne*(^{3} P _{0,2}) colliding with H_{2}, HD, D_{2}, O_{2}, and H_{2}O at thermal energies
View Description Hide DescriptionThe XUV photon emission following collision‐induced fine‐structure transitions Ne*(^{3} P _{0,2})+M →Ne*(^{1,3} P _{1})+M→Ne(^{1} S)+M+hν(74 nm) has been measured for M=H_{2}, HD, D_{2}, O_{2}, and H_{2}O in a beam‐cell experiment. The metastable Ne* velocity ranges from 500 to 1500 m/s. No emission was found for Ne*(^{3} P _{0})+M. An appropriate detector system enables the determination of absolute ^{3} P _{2}→^{3} P _{1} transition cross sections. The hierarchy of the cross sections measured parallels that of collision‐induced ^{2} P _{1/2}→^{2} P _{3/2} transition cross sections in Rb(5 ^{2} P) and Cs(6 ^{2} P) which have comparable fine‐structure energy splittings.

Dynamic curve crossing of an atom interacting with a set of atoms in three dimensions
View Description Hide DescriptionA general method is presented for computing the probability of an atomic curve crossing in three dimensions (3D). This method enables one to compute this probability totally dynamically, under conditions of random scattering with a cluster of atoms. A 3D Landau–Zener diabatic curve crossing probability calculation is shown how to be built into 3D atom‐group of atoms trajectory dynamics calculations. The objective of this method is to mirror real dynamical systems as closely as possible by theory. The atom is assumed to be confined to the cluster of atoms, so that it makes a sequence of attempted crossovers until successfully crossing over, or until curve crossing is no longer physically possible. A set of reference trajectories covering the allowed curve crossing energy range of interest must be obtained. Subsequently, this reference set is used in a runoff procedure yielding the net bonding probability to a subset of the atoms of the cluster. The method is applied to the specific case where the atomic cluster is a solid. More specifically, the loss function of graphite due to 6.0 eV incident oxygen atoms is obtained.

Auger electron–ion coincidence experiment on nitric oxide molecule excited by electron impact
View Description Hide DescriptionThe fragmentation of nitric oxide by electron impact is studied via Auger electron–ion coincidence experiments. The kinetic energy releases of the different fragments have been measured in the 39–70 eV binding energy region of the dication. The experimental data confirm that the three lower lying states of NO^{2+} are bound states, while all the other states in the studied region are repulsive ones. Experimental evidence is provided that all the repulsive states up to 54 eV binding energy dissociate to the lowest dissociation limit and that dissociative channels leading to the formation of N^{2+} and O^{2+} are populated at 62 eV binding energy. The experimental results are compared with previous experimental data obtained with different techniques and with the more recent theoretical predictions.

Dynamic aspects of electronic predissociation
View Description Hide DescriptionWe consider electronic excitation induced with a continuous wave laser to an excited bound state which can predissociate due to a radiationless transition to a dissociative state. The conditions for a separation of the process into the preparation of a vibrational eigenstate which subsequently dissociates due to a radiationless transition are established. We point out that the probability of the radiationless transition can be calculated from a time‐dependent nuclear autocorrelation function, an expression which nicely reflects the pictorial aspect of the Franck–Condon principle.

A localized orbitals based embedded cluster procedure for modeling chemisorption on large finite clusters and infinitely extended surfaces
View Description Hide DescriptionA new embedded cluster procedure for modeling chemisorption on metal surfaces is developed. The procedure is similar in philosophy to the approach used by Whitten and co‐workers in that energy calculations are performed in a cluster region basis consisting of localized occupied and virtual orbitals. However, we present a new localization procedure to generate the cluster region functions which is based on orbital occupation numbers determined from the density matrix obtained in a calculation on the extended substrate. Our localization procedure avoids having to perform separate unitary transformations on the canonical occupied and virtual orbitals and as a consequence has the attractive feature of enabling the embedded cluster calculations to be applied to both large finite clusters and infinitely extended systems in essentially the same manner. We illustrate the embedded cluster procedure by performing partial SCF calculations in the cluster region basis for H adsorption at an on‐top site of a Li(100) monolayer.
When the extended surface is modeled by large finite clusters, the localized orbitals in the cluster region rapidly converge to being completely occupied or completely empty, and we find partial SCF calculations to readily reproduce the full SCF results of the large finite cluster. For the infinitely extended surface, the occupation numbers for the localized functions in cluster regions converge much more slowly than in the finite case, but even with less than perfect occupation numbers we still obtain good H adsorption properties in the partial SCF calculations. Unlike the finite cluster case where charges are automatically balanced, we found in order to achieve good results in the partial SCF calculations on the infinitely extended systems it was necessary to carefully balance the charges used in the long range electron and nuclear interactions. All of the calculations involving clusters are performed with the GAMESS program and the calculations on the infinite extended surface are performed with the periodic Hartree–Fock CRYSTAL program.

Ab initio study of cis‐butadiene valence and Rydberg states using the effective valence shell Hamiltonian method
View Description Hide DescriptionLow‐lying σ‐ and π‐electron vertical excitation energies of s‐cis‐1,3‐butadiene are calculated using the ab initio effective valence shell Hamiltonian (H^{ v }) method. The only experimentally known vertical excitation energy is that to the 1 ^{1} B _{2} state at 5.49 eV, while the H^{ v } computation in the π‐valence space yields 5.62 eV. Calculated excitation energies to various valence and Rydberg states are in good agreement with theoretical multiconfigurational single reference state second‐order perturbation theory calculations by Roos and co‐workers and with values from other highly correlated computations. The H^{ v } calculations for cis‐butadiene further investigate the dependence of the computations on the nature and the choice of molecular orbitals and provide the first comprehensive study of the convergence with respect to the enlargement of the valence space for π‐electron systems. The present computations also represent the first H^{ v } treatment of the σ→π* and π→σ* excited states in conjugated π‐electron systems, along with an analysis of the required degree of σ–π correlation within the valence (or reference) space. Vertical π‐and σ‐ionization potentials are also produced as a byproduct of the H^{ v } calculations for neutral cis‐butadiene, providing the first predictions of these ionization energies. The computations conclusively reconfirm the high accuracy of the H^{ v } method.

The solvation reaction field for a hydrogen atom in a dielectric continuum
View Description Hide DescriptionA reaction field exists even for a nonpolar solute embedded in a spherical cavity within a surrounding homogeneous dielectric continuum. This arises from the tail of the electronic wave function that penetrates beyond the cavity boundary into the dielectric region. This effect, which is neglected or treated only in cursory fashion in most reaction field implementations, is examined in detail for the simple case of a ground state hydrogen atom, where very accurate solutions of the relevant equations can be obtained. Properties considered include the penetration of the electron outside the cavity, the electronic density at the nucleus, the electron binding energy, the electrostaticfree energy of solvation, the polarizability, and the vertical 1s→2pexcitation energy. Also, the effect of the common approximation of neglecting the volume polarization and treating only the surface polarization contribution to the reaction field is critically evaluated.

Electron correlation effects on the theoretical calculation of nuclear magnetic resonance spin–spin coupling constants
View Description Hide DescriptionThe equation‐of‐motion coupled cluster singles and doubles (EOM‐CCSD) method for general second‐order properties is derived providing a quadratic, CI‐like approximation and its linked form from coupled cluster (CC) energy derivative theory. The effects of the quadratic contribution, of the atomic basis set employed, and of electron correlation on NMR spin–spin coupling constant calculations using EOM‐CCSD methods are investigated for a selected set of difficult molecules, notably CH_{3}F, B_{2}H_{6}, CH_{3}CN, C_{2}H_{4}, and CH_{3}NH_{2}. We find that the quadratic contribution is insignificant for the couplings in the molecules considered in this study and in addition the quadratic contribution only slightly depends on the basis set used. Therefore it seems well justified to use the less expensive CI‐like approximation or its linked‐diagram form to evaluate spin–spin coupling constants. The Fermi‐contact contribution shows the largest variation with the change of basis sets. The diamagnetic spin–orbit (DSO) and the spin–dipole (SD) contribution vary little, seemingly being converged at the DZP level while the paramagnetic spin–orbit (PSO) term shows moderate variations. Except for very few cases, the FC contribution is dominant in all the couplings in the selected set of molecules and it is also most sensitive to the inclusion of electron correlation. The other contributions are less affected by electron correlation. Although of lesser importance, the significance of the noncontact contributions and electron correlation effects on accurate calculation of coupling constants such as ^{1} J(^{13}C^{19}F) in CH_{3}F and ^{1} J(^{13}C^{15}N) in CH_{3}CN is clearly demonstrated.

Basis set superposition problem in interaction energy calculations with explicitly correlated bases: Saturated second‐ and third‐order energies for He_{2}
View Description Hide DescriptionExplicitly correlated basis set of Gaussian‐type geminals has been employed in supermolecular calculations of the interaction energy of two helium atoms using the second‐ and third‐order of the many‐body perturbation theory and the Mo/ller–Plesset partitioning of the Hamiltonian. A geminal extension of the counterpoise procedure of Boys and Bernardi has been proposed to correct for the basis set superposition error. Performance of the proposed correction scheme has been analyzed at the second‐order level using a sequence of geminal bases varying in the degree of completeness in representing the intra‐ and intermonomer correlation effects. The nonlinear parameters of these bases were optimized by minimizing the second‐order energy of the helium atom and the second‐order dispersion energy of the He dimer. The best upper bounds to date have been obtained for both quantities. The numerical results show that the counterpoise procedure should be used at all levels of basis set completeness. By employing the union of the largest of the obtained bases and reoptimizing some of the nonlinear parameters using the complete second‐order energy functional for the dimer, the best estimates to date of the second‐ and third‐order supermolecular interaction energies for He_{2} have been computed. At the minimum interatomic separation these energies are estimated to be accurate to 0.01 K or better. Adding higher‐order terms computed using orbital bases, leads to a helium dimer interaction potential with the depth of 11.00 K, somewhat larger than current experimental results.

Damping of perturbation corrections in quasidegenerate situations
View Description Hide DescriptionShifting the pole of the energy denominators from the real axis to the imaginary results in a damping of the equation in degenerate cases. The term‐by‐term absolute value of the expression provides a simple and useful formula to treat quasidegenerate problems. The power of the new expression is illustrated on several examples calculating the correlation energy of some molecules. The Mo/ller–Plesset results are scarcely affected by the proposed modification in normal cases but significant improvement is observed in quasidegenerate situations. The proposed formula is size consistent and its evaluation does not need extra computational effort.

On the relation between the Wertheim’s two‐density integral equation theory for associating fluids and Chandler–Silbey–Ladanyi integral equation theory for site–site molecular fluids
View Description Hide DescriptionIt is demonstrated that Chandler–Silbey–Ladanyi integral equation theory for the site–site molecular fluids is the limiting case of complete association of more general two‐density integral equation theory for associating fluids developed by Wertheim. The analysis is presented for a site–site molecular system with any number and geometrical arrangement of the sites in the molecule and arbitrary type of the site–site pair interaction.

Microscopic model with temperature‐dependent interactions for the free molecule and for the trigonal phase of benzil
View Description Hide DescriptionThe molecule of benzil (diphenylethanedione, C_{14}H_{10}O_{2}) has been approximated by a system of rigid segments to model the lowest‐frequency part of its vibrational spectrum. The interactions of internal degrees of freedom have been described with the use of phenomenological force constants. The structure of the trigonal (P3_{1}21) phase has then been modelled by means of a temperature‐dependent atom–atom potential based on thermal motions of atoms. The potential gives the correct account of the softening of an E‐symmetry, zone‐center mode which underlies the phase transition to the low‐temperature monoclinic phase (P2_{1}). The low‐frequency modes at the zone center, supposed until now to be difference overtones, have been shown to result from a coupling between internal and external degrees of freedom. A low‐frequency soft mode at the point M of the zone border has been found, which explains the behavior of observed peaks in diffuse x‐ray scattering experiments. The values and the temperature evolution of the effective elastic constants calculated within the model are in a very good agreement with the results of ultrasonic and Brillouin scattering data. The model has been shown insufficient in the description of dielectric and piezoelectricproperties of benzil.

Kinetic laws at the collapse transition of a homopolymer
View Description Hide DescriptionWe present results from numerical analysis of the equations derived in the Gaussian self‐consistent method for kinetics at the collapse transition of a homopolymer in dilute solution. The kinetic laws are obtained with and without hydrodynamics for different quench depths and viscosities of the solvent. Some of our earlier analytical estimates are confirmed, and new ones generated. Thus the first kinetic stage for small quenches is described by a power law decrease in time of the squared radius of gyration with the universal exponent α_{ i }=9/11 (7/11) with (without) hydrodynamics. We find the scaling laws of the characteristic time of the coarsening stage, τ_{ m }∼N ^{γm }, and the final relaxation time, τ_{ f }∼N ^{γf }, as a function of the degree of polymerizationN. These exponents are equal to γ_{ m }=3/2, γ_{ f }=1 in the regime of strong hydrodynamicinteraction, and γ_{ m }=2, γ_{ f }=5/3 without hydrodynamics. We regard this paper as the completion of our work on the collapse kinetics of a bead and spring model of a homopolymer, but discuss the possibility of studying more complex systems.

Ab initio study of a CO monolayer adsorbed on the (101̄0) surface of ZnO
View Description Hide DescriptionPeriodic Hartree–Fock total energy calculations on two‐dimensionally periodic slabs have been used to predict the equilibrium geometry of a monolayer of carbon monoxide molecules adsorbed on the nonpolar (101̄0) surface of ZnO. Two physisorbed (or weakly chemisorbed) minimum energy configurations are found. In one the CO molecules adsorb with their oxygen atoms coordinated to surface Zn atoms, while in the other the carbon atoms are coordinated to surface Zn atoms. The two calculated minima are very close in energy. In the second geometry, the C–Zn ‘‘bond’’ and the C–O bond make angles of 32.5° and 39.5° with the surface normal, and the intramolecular bond shortens slightly from its free value in reasonable agreement with experimental results. No binding of CO to the surface oxygen atoms is predicted. Surface‐related changes in the vibrational frequencies for the adsorbed molecules agree reasonably well with infrared spectroscopic data, and the ‘‘carbon‐down’’ binding energy of the molecule with the surface is in good agreement with thermal desorption data (though electron correlation effects have to be included in the calculation to obtain acceptable results for low surface coverage).