Files in this item



application/pdfXU-DISSERTATION-2015.pdf (80MB)
(no description provided)PDF


Title:Theoretical and computational investigations of the electron and spin correlation in chemical bonds, building a predictive bonding model with generalized valence bond theory
Author(s):Xu, Lu
Director of Research:Dunning, Thom H., Jr.
Doctoral Committee Chair(s):Dunning, Thom H., Jr.
Doctoral Committee Member(s):Hirata, So; Makri, Nancy; Moore, Jeffrey S.
Department / Program:Chemistry
Degree Granting Institution:University of Illinois at Urbana-Champaign
Subject(s):chemical bonding
Generalized Valence Bond (GVB) Theory
Abstract:One of the goals of modern quantum chemistry is to understand the nature of chemical bonds. By understanding the underlying physical and chemical principles of chemical bonding in various molecules, chemists can make predictions about the structures, energetics, reactivities and other properties of chemical species. The understanding of bonding is also critical to achieving one of the ultimate goals of chemists—the design and control of molecules and molecular processes. Therefore, an intuitive, rigorous and predictive bonding model is needed—one that is applicable to a wide range of elements. Various models have been used to rationalize bonding in molecules. Hartree-Fock (HF) and Molecular orbital (MO) theory are widely used. However, HF theory fails in cases where multireference character is important. HF theory also does not describe bond dissociation correctly in most cases; thus, it is of limited use in understanding the details of chemical reactions. Further, molecular orbitals are often delocalized over the entire molecule, making their interpretations in terms consistent with chemical language (bonds, lone pairs, etc.) difficult. Valence bond (VB) theory uses a language much closer to that of chemistry and, early on, helped chemists adapt quantum mechanics to understanding molecular structures and reactions. For example, Pauling’s hybridization model is often used to rationalize bonding in traditional hypervalent molecules such as SF6, as well as the bonding in carbon species such as CH4 and C2H2. However, these concepts are, by and large, empirical with significant limitations, e.g., detailed calculations in the 1980s clearly showed that spd hybrids were not involved in the bonding in hypervalent molecules. The more recent 3 center-4 electron model of hypervalency is similar to the hybridization model in its limitations, and is subject to the constraints of MO theory. In this dissertation, the nature of the chemical bonding is investigated in a number of different molecules using generalized valence bond (GVB) theory, with the molecules being chosen to better understand aspects of the bonding. GVB strikes an attractive balance between accuracy and interpretability by including the most important non-dynamical correlations, yet retaining a wavefunction that is a single product of spatial orbitals, although with a more general spin function than that used in HF theory. Because of GVB’s ability to describe the making and breaking of chemical bonds and describe the bonding in molecules with or without multirefernce character, a bonding model can be constructed based on GVB theory that is applicable to a broad range of molecules. In the course of this and earlier work, the types of bonds that can be formed were expanded beyond traditional chemical bonds, such as covalent bonds, to include a new type of bond—the recoupled pair bond. The recoupled pair bond not only rationalizes the formation of traditional hypervalent molecules such as SF6 and PF5, but it also describes the bonding in radicals such as SF3 and SF5 as well as beryllium, boron and carbon compounds in a consistent and systematic fashion. The differences between sulfur and carbon species are the differences between recoupled pair bonding using 3p lone pairs and 2s lone pairs. The fundamental aspects of the s- and p-recoupled pair bonds and recoupled pair bond dyads with monovalent ligands using CF, SF, CF2 and SF2 as examples, and the fundamentals of the s-recoupled pair bonds with trivalent ligands using BeN, BN and CN as examples will be described in detail in this dissertation (divalent ligands were considered by others). As noted earlier, the GVB wavefunction has a compact functional form consisting of a spatial wavefunction, a product of orbitals, and a spin wavefunction, a product of αβ’s for the doubly occupied orbitals and a general spin function for the electrons in the singly occupied active orbitals. A single set of spatial orbitals, whose functional form is determined by the application of the variational principle, makes interpretation of the GVB orbitals unambiguous and straightforward, and the freedom in the spin wavefunction enables one to probe various spin couplings other than the traditional perfect pairing coupling, which is assumed in the HF wavefunction. In a number of molecules, the perfect pairing spin coupling is not the dominant spin coupling at the equilibrium geometry of the molecule—the simplest carbon compound, C2, is a case in point. The spin couplings and their implications in the homonuclear diatomic molecule C2, N2, P2 and As2 will be presented, and the nature of the multiple bonds in these molecules will be discussed. The use of GVB theory has allowed us to study a wide range of molecules and radicals systematically. One recurring theme in our research is characterizing and understanding the “first-row anomaly”—the long recognized differences in the chemistry of the first row elements versus those lower in the Periodic Table. Examples described in this dissertation include: CHn versus SiHn (n=0–4), N2 versus P2 and As2, and the inversion transition states of FnNH(3-n) versus those of FnPH(3-n) (n=0–3). A comparison of the first and second row compounds will be presented throughout the dissertation.
Issue Date:2015-04-17
Rights Information:Copyright 2015 Lu Xu
Date Available in IDEALS:2015-07-22
Date Deposited:May 2015

This item appears in the following Collection(s)

Item Statistics