Hanspeter Huber

Deuterium quadrupole coupling constants. A theoretical investigation

Deuterium quadrupole coupling constants were obtained for 39 sites from ab initio SCF calculations. An accuracy comparable to the experimental one is reached with a moderately large basis set of very high local quality. Comparison with experiment shows that electron correlation does not contribute substantially to this property. The experimental values for DNCO and DFCO should be reexamined, the value for CD3Cl in the literature was not properly transformed to the bond axes system, and the experimental value for ND3 agrees with the calculated one, thus showing that a dynamical model is not necessary. Several empirical relations are presented, and a simple classical model is proposed which relates the deuterium quadrupole coupling constant to the bond length with high accuracy. A theoretical study of the HCOOH dimer gives some insight into the importance of different mechanisms which lower the quadrupole coupling constant in the solid.

Ab initio calculation of the deuterium quadrupole coupling in liquid water

The quadrupole coupling constant and asymmetry parameter for the deuteron in liquid heavy water was determined using purely theoretical methods. Molecular‐dynamics simulations with the ab initio potential‐energy surface of Lie and Clementi were used to generate snapshots of the liquid. The electric‐field gradient at the deuteron was then calculated for these configurations and averaged to obtain the liquid quadrupole coupling constant. At 300 K a quadrupole coupling constant of 256±5 kHz and an asymmetry parameter of 0.164±0.003 were obtained. The temperature dependence of the quadrupole coupling constant was investigated.

Structural and thermodynamic properties of fluid carbon dioxide from a new ab initio potential energy surface

An intermolecular potential energy surface for the carbon dioxide dimer is calculated fully ab initio using a large basis set and including electron correlation. From this potential the dimer structure and the second virial coefficients are determined. In addition, it is applied in molecular dynamics simulations to obtain the fluid structure, the pressure, the internal energy, the thermal pressure coefficient, and the molar heat at constant volume. The results are compared with those from simulations with a previous ab initio potential. In this way we gain information regarding the sensitivity of each property to the quality of the quantum chemically obtained potential. Equilibration of carbon dioxide simulations must be done with great care due to the very slow energy transfer between the intramolecular vibrations and the other degrees of freedom. This point is addressed in some detail.

Melting curves for neon calculated from pure theory

The melting curve of neon is determined from nonequilibrium molecular dynamics simulations performed at constant pressure, using ab initio pair potentials. The effects of various approximations on the predicted melting points are investigated through the use of pair potentials calculated at different levels of accuracy, and the inclusion of quantum effects on the motion via a Wigner–Kirkwood quantum effective potential. To avoid superheating of the model crystal, nucleation sites for melting are provided by creating clusters of void defects in the crystal prior to heating. The calculated melting curves are shown to be in good agreement with experimental measurements. Comparisons are made with similar calculations previously carried out for argon.

The carbon dioxide dimer

Ab initio calculations with a large basis set (5s 4p 2d), including correlation by second-order Moller-Plesset perburbation theory (MP2), are performed for the dimer of carbon dioxide. The energies, structures and IR frequencies are all calculated at this level. It is found that the T-shaped conformation is a transition state between two slipped parallel dimers of such low energy that even at relatively low temperatures a planar geared motion is possible. (Similar behaviour has been found previously for the acetylene dimer.) The experimental data available are in excellent agreement with our results.

A new ab initio potential for the neon dimer and its application in molecular dynamics simulations of the condensed phase

An ab initio potential for the neon dimer is derived yielding about 92% of the experimentally estimated interaction energy. In addition to large basis sets on the nuclei, bond functions are utilized for polarization in Moller-Plesset MP4(SDTQ) calculations. The analytical potential was used to solve the nuclear Schrodinger equation for comparison with and prediction of spectroscopic properties and to perform classical molecular dynamics simulations in the condensed phase at high pressures. In the supercritical state at room temperature and pressures up to 1000 MPa the error in the predicted pressure is always less than 4%.

Ab initio calculation of the intermolecular potential energy surface of (CO2)2 and first applications in simulations of fluid CO2

Abstract A four-dimensional potential energy surface for the carbon dioxide dimer composed from rigid monomers is determined. 226 energy points were calculated ab initio with a large basis set on the MP2 level including full counterpoise correction. An analytical site-site potential is fitted to these points. The stationary points of the analytical surface are determined, harmonic vibrational frequencies and second virial coefficients are calculated and compared to experimental values. The potential energy surface is applied in molecular dynamics simulations to obtain pair distribution functions of supercritical and liquid carbon dioxide, which are compared to experiment. In addition some state points of the phase diagram are calculated.

The use of molecular dynamics simulations with ab initio SCF calculations for the determination of the oxygen-17 quadrupole coupling constant in liquid water

Molecular dynamics simulations for H2 17O were performed with two different potentials from the literature. After equilibration, snapshots were taken at intervals and the configurations thereby obtained used to randomly select clusters of molecules. For the central oxygen of these clusters the electric field gradient was then calculated with standard ab initio self-consistent field (SCF) calculations. The field gradients were converted to quadrupole couplings with a factor obtained previously in gas phase calculations and averaged to obtain the 17O coupling in liquid water. At 300 K a best estimate of 8·9 ± 0·3 MHz for the coupling constant and 0·72 ± 0·04 for η were found. The coupling constant at 360 K is about 0·1 MHz larger but the temperature dependence is not significant.

Ab initio calculation of the second virial coefficient of neon and the potential energy curve of Ne2

Abstract For the ab initio calculation of properties of dense gases and liquids accurate potential energy curves are needed. Here we describe a method to obtain an “optimal” curve for a given number of ab initio calculations. Strictly speaking, we give a recipe, with which interatomic distances calculations should be performed to yield the second virial coefficient with a minimal loss in accuracy due to the limited number of points. With this technique we obtain the virial coefficient from about 15 points essentially without loss in accuracy. Experimental data for a direct test of the quality of the potential energy curve are scarce. For the neon dimer we have calculated the potential well depth, the equilibrium distance and the vibrational frequency for such a comparison with experiment. An indirect indication of the quality of the curve is given by the calculation of the second virial coefficient at different temperatures. The best calculations presented here use extended basis sets, correct for the basis set superposition error by the counterpoise method, and include electron correlation using Moller-Plesset perturbation theory on the MP4 (SDTQ) level.

Near-Hartree-Fock calculation of the electric field gradients and their first and second derivatives with respect to the bond-length at the location of the deuterium nucleus

The electric field gradients and their first and second derivatives with respect to the bond length at the location of the deuterium nucleus are calculated for D2, DF, DCN, CD4, CD3F, CD3Cl and CD3NO2, respectively. These parameters are required for vibration and vibration-rotation coupling corrections to N.M.R. measurements in liquid crystal solvents. A special basis set with local near-Hartree-Fock quality has been constructed which yields very accurate results for the examples where experimental results are available. The results show that at least in some cases the effect of electron correlation has been overestimated.