ON THE N2-HE POTENTIAL-ENERGY SURFACE

A detailed comparison of the predictive powers of two recently determined empirical and two recently proposed theoretical potential energy surfaces for the N2-He interaction has been carried out. In particular, the following properties have been tested: at the microscopic level, total and state-to-state differential cross sections and absolute total integral cross sections, while at the macroscopic level, interaction second virial, diffusion, viscosity, and thermal conductivity coefficients, as well as the rotational relaxation time, depolarized Rayleigh spectral collision broadening, and shear viscosity and thermal conductivity field-effect data in N2-He mixtures. Exact calculations have been employed, from the close-coupling method for treating scattering data at low energies to the classical trajectory method with second-order corrections to compute the effective cross sections that determine the bulk transport and relaxation phenomena. The empirical exponential-spline-Morse-spline-van der Waals surface [J. Chem. Phys. 85, 7011 (1986)], closely followed by the model Bowers-Tang-Toennies surface [J. Chem. Phys. 88, 5465 (1988)], gives better simultaneous agreement with the scattering data, the second virial coefficient data, the bulk transport data, and the depolarized Rayleigh collision-broadening data, which are properties sensitive to the spherical component of the interaction and to the anisotropy of the low repulsive wall. None of the potential surfaces examined here includes a dependence upon the vibrational stretching coordinate of the N2 molecule, since none of the data employed in the fitting is sensitive to this coordinate. The two theoretical model potentials, especially that based upon an earlier Hartree-Fock plus damped dispersion model surface [J. Phys. Chem. 88, 2036 (1984)], gives better agreement with the rotational relaxation and field-effect data, which are properties sensitive to the anisotropy of the high-repulsive wall. It is established that the exponential-spline-Morse-spline-van der Waals and Bowers-Tang-Toennies surfaces are on the whole the more reliable of the empirical and model surfaces examined, respectively. It is concluded that the optimum N2-He potential energy surface should be a blend of the empirical exponential-spline-Morse-spline-van der Waals and of the two model surfaces.