Building a potencial energy surface for the ground singlet state of the hydrogen peroxide

Coelho,  Daniela Veloso

http://hdl.handle.net/10400.1/10788

Use this identifier to reference this record.

Name:	Description:	Size:	Format:
these.pdf		1.39 MB	Adobe PDF	Download

Send Feedback

Authors

Coelho, Daniela Veloso

Advisor(s)

Brandão, João Carlos Pereira Pers

Abstract(s)

The hydrogen peroxide system has attracted widespread attention due to its overwhelming importance in many chemical processes such as combustion, atmospheric and interstellar chemistry. The main goal of the present PhD thesis is the construction of a six-dimensional Potential Energy Surface (PES) for the ground state of the H2O2 system with chemical accuracy in order to achieve trustworthy kinetic data. The analytical representation of the PES is based on a double many-body expansion (DMBE) formalism, where a 3 3 matrix is used to accurately reproduce all the dissociation channels. The rst part of this thesis concerns the concept of PES, the coordinates used to represent it as well as its major topological features. A survey of the theoretical framework of the ab initio calculations and the general strategy to obtain the analytical representation of the PES are also reported. The second part focuses on the description of the ab initio electronic calculations performed for mapping the most important regions of the con guration space. An extrapolation/scaling scheme is proposed to accomplish high-quality ab initio energies. The parameters used in this procedure are obtained by interpolation among several reference geometries. This new DMBE hydrogen peroxide PES also accounts for the electrostatic dipoledipole interaction between two OH(2 ) fragments. The functional form used to represent the short-range interactions is based on a sum of polynomial functions of the fourth degree multiplied by a range factor, both built with intrinsic permutation symmetry and centred at speci c reference geometries, to which the ab initio points computed are assigned based on a k-means algorithm. Finally, important features of the model function are characterized and the major conclusions are summarized. Further improvements and possible applications for the present potential are also outlined.