Improved Potential Energy Surface of Ozone Constructed Using the Fitting by Permutationally Invariant Polynomial Function

New global potential energy surface for the ground electronic state of ozone is constructed at the complete basis set level of the multireference configuration interaction theory. A method of fitting the data points by analytical permutationally invariant polynomial function is adopted. A small set of 500 points is preoptimized using the old surface of ozone. In this procedure the positions of points in the configuration space are chosen such that the RMS deviation of the fit is minimized. New ab initio calculations are carried out at these points and are used to build new surface. Additional points are added to the vicinity of the minimum energy path in order to improve accuracy of the fit, particularly in the region where the surface of ozone exhibits a shallow van der Waals well. New surface can be used to study formation of ozone at thermal energies and its spectroscopy near the dissociation threshold. 1. Introduction The global potential energy surface (PES) for the ground electronic state of ozone, is now 10 years old. The old PES was based on the ab initio calculations of the electronic structure at the icMRCI+Q/cc-pVQZ level of theory using CASSCF(12,9) active space and was represented by a 3D-spline interpolation of the data on a structured rectangular grid [1, 2]. The ab initio data obtained at that level of theory contained two serious deficiencies. First, the computed dissociation energy was too low—about lower than the experimental value available at that time (see below). Second, the surface exhibited an artificial barrier on its way to dissociation—about above the threshold. In the improved, most popular version of that surface [3] an analytic correction function was added to the spline in order to (i) eliminate the barrier by taking into account the results of more accurate ab initio calculations [4, 5] carried out along the one-dimensional minimum energy path to dissociation (MEP) and (ii) make the surface deeper in order to reproduce experimental value of the dissociation energy available at that time [6], . More accurate ab initio calculations of the electronic structure of O3 were impossible 10 years ago but have become quite feasible nowadays. Furthermore, according to new experimental information [7, 8], the surface must be somewhat deeper, . Clearly, there is an opportunity and a need of improving the existing PES of ozone. During the last decade several attempts have been made [8–10] to assess the level of theory needed to construct an accurate PES of O3 and obtain correct value of the dissociation energy without any empirical