D3.1 Localized Bonds

Lewis structures are easy-to-draw planar representations of bonding in molecules. They help us figure out and to think about which atoms are bonded to which and whether bonds are single or multiple. However, they do not directly correspond to the molecular orbitals that determine actual electron-density distributions in a molecule. We need an additional tool that will connect Lewis structures to modern quantum theory (molecular orbital theory) that would provide more information on molecular properties, for example, 3D geometry of a molecule, from Lewis structures.

Valence bond theory is a model that focuses on the formation of individual chemical bonds, such as the formation of a σ bond between two atoms within a polyatomic molecule. Like molecular orbital theory, valence bond theory deals with how atomic orbitals change and combine when a molecule forms. But unlike molecular orbital theory, instead of forming molecular orbitals that span the entire molecule, valence bond theory models the combination of valence orbitals of each atom individually, such that the resulting combination in the model gives stronger bonding in specific directions. Hence, valence bond theory allows us to derive idealized 3D geometries for molecules based only on their Lewis structures. Moreover, the localized bonding picture presented by Lewis structures and valence bond theory are often much more intuitive for understanding molecular properties (the sprawling forms of molecular orbitals for large molecules can even overwhelm expert chemists).

In a previous course, you may have used VSEPR (valence shell electron pair repulsion theory) to predict the 3D shapes of molecules. VSEPR involves counting electron regions (pairs) around a central atom, assuming that electron regions repel and stay as far apart as possible, and bonding terminal atoms to electron regions. VSEPR is often good at predicting the arrangement of bonds around an atom, and it is OK to use it to predict idealized linear, trigonal planar, and tetrahedral arrangements of bonds that you will encounter in this course, but VSEPR has significant limitations:

• VSEPR has little or no basis in modern quantum theory; you have just spent significant time studying quantum theory and we want you to be able to use that experience.
• It is often difficult to apply VSEPR to molecules described by two or more resonance structures (we will discuss this topic in more detail in a later section). Thus VSEPR makes it more difficult to understand many molecular structures—for example, structures of protein molecules.
• VSEPR assumes that lone pairs occupy more space than bond pairs, but there is no evidence, experimental or theoretical, to support that assumption; in fact, there is some evidence to the contrary.
• VSEPR assumes that all lone pairs are equivalent, but there is experimental evidence that they are not. For example, the two lone pairs in a water molecule do not have the same ionization energy and do not have equivalent probability distributions (Journal of Chemical Education 1987, Vol. 64, pp 124-128 link).
• VSEPR often cannot explain relative bond angles. For example, why is the H-P-H angle in PH₃ 93.5° while the H-N-H angle in NH₃ is 107.5°? (If the decrease in bond angle from the tetrahedral angle of 109.5° to 107.5° for NH₃ is due to a “fatter” lone pair, why does the angle decrease so much more for the larger P atom? A “fatter” lone pair should be less likely to repel the other bonds because they are farther apart.)

For these reasons, VSEPR is a model that has limited applicability. In this course, we will use a better model—valence bond theory—which is consistent with modern quantum theory, makes more accurate and more comprehensive predictions than VSEPR, and is a better basis for understanding more advanced bonding topics. If you want to, it is OK to use VSEPR to predict idealized shapes, but applying the ideas presented in this and the next few sections will allow you to describe molecular structures better and understand bond properties more accurately.