So far we've presented a straightforward view of covalent bonding as the sharing of electrons between two atoms. However, we have yet to answer questions such as these: How are electrons shared? What orbitals do shared electrons reside in? Can we say anything about the energies of these shared electrons? Our task now is to extend the orbital scheme that we've developed for atoms to describe bonding in molecules.
Valence bond theory (VB) is a straightforward extension of Lewis structures. Valence bond theory says that electrons in a covalent bond reside in a region that is the overlap of individual atomic orbitals. For example, the covalent bond in molecular hydrogen can be thought of as result of the overlap of two hydrogen 1 s orbitals.
In order to understand the limitations of valence bond theory, first we must digress to discuss molecular geometry, which is the spatial arrangement of covalent bonds around an atom. A very simple and intuitive approach, the Valence Shell Electron Pair Repulsion (VSEPR) model, is used to explain molecular geometry. VSEPR states that electron pairs tend to arrange themselves around an atom in such a way that repulsions between pairs are minimized. /PARGRAPH For instance, VSEPR predicts that carbon, which has a valence of four, should have a tetrahedral geometry. This is the observed geometry of methane ( CH 4 ). In such an arrangement, each bond about carbon points to the vertices of an imaginary tetrahedron, with bond angles of 109.5 degrees, which is the largest bond angle that can be attained between all four bonding pairs at once. Similarly, the best arrangement for three electron pairs is a trigonal planar geometry with bond angles of 120 degrees. The best arrangement for two pairs is a linear geometry with a bond angle of 180 degrees.
The VSEPR scheme includes lone pairs as well as bonded pairs. Since lone pairs are closer to the atom, they actually take up slightly more space then bonded pairs. However, lone pairs are "invisible" as far as the geometry of the atom is concerned. For instance, ammonia ( CH 3 ) has three C-H bonded pairs and one lone pair. These four electrons will, like methane, occupy a tetrahedral arrangement. Since lone pairs take more space, the H-N-H bond angle is reduced from 109.5 degrees to about 107 degrees. The geometry of ammonia is trigonal pyramidal rather than tetrahedral since the lone pair is not included. By similar reasoning, water has a bent geometry with a bond angle of about 105 degrees.
Note that multiple bonds don't affect the VSEPR scheme. A double or triple bond is considered no more repulsive than a single bond.
The Valence Bond model runs into problems as soon as we try to take molecular geometries into account. The tetrahedral geometry of methane is clearly impossible if carbon uses its 2s and 2p orbitals to form the C-H bonds, which should yield bond angles of 90 degrees.