How may valence bond theory be used to predict the shape of complexes?? Using examples.?

Question

Any help would be much appreciated

Answer 1

How sophisticated an answer do you want, and who asked it of you, in what context?

Most first-year textbooks will claim or imply that you can use VBT to "predict" molecular shape using a quick back-of-the-envelope procedure based on electron promotion/hybridization.

Example, SF4.  Ridiculously esoteric, but illustrates the principles.

S has a configuration of [Ne](2s)2(2p)4.  You've formed four bonds to fluorine groups, so you need 4 unparied electrons, right now you only have two.  So you promote one electron out of the first p orbital, giving you (2s)2 (p)1(p)1(p)1(d)1, throwing one electron up into the 3d subshell.  How you hybridize those five orbitals, forming an sp3d trigonal bipyramidal arrangement, four hybrids have one electron to form a bond, one hybrid has a lone pair, the result is a see-saw shape for SF4, which is indeed the observed structure.

Basically: take ground state electron config of central atom, promote electrons needed to create a number of unpaired electrons equal to the number of bonds you want to form, hybridize the LPs and single electron orbitals to establish the correct arrangement of bond pairs and lone pairs, make your bonds and you're done.  Usually you already have an idea of the shape of the thing from Lewis/VSEPR, so you know what you're "predicting".

But here's the thing.  How do you know when to hybridize?  You don't always.  PH3 and SH2 are best described without hybridization, because their bond angles are close to 90º, not the a tetrahedral angle.  How do you know you have the right structure in mind?  If you repeat my example with TeF4, you'll predict the same structure, but you'll be wrong.  What's the structure of Me2Sn, or of SnCl2?  There are different ways to get an octet at Sn with those compounds, and they have very different structures, which you have no way to predict using the style of VBT presented in a first year text.

It is possible to use VBT mathematically, to compute the lowest energy arrangement of atoms in different structural forms and make REAL "predictions", but nobody does that anymore, we use MOs if you're going to that amount of trouble.  But here's the real problem.

Bonding theories are rationalizations after the fact.  They are not meant to be used, and are not used by practicing chemists, to make predictions of chemical structure.  They are used to rationalize and explain molecular structure AFTER than structure has been determined EXPERIMENTALLY.

Why use VBT to predict the structure of NH3 or SF4 or SnCl2.  They're known.  They're well established.  You don't need to predict grass is green, you can see that with your own eyes -- science's job is to offer a reason.

Far too many text book authors don't get that.  You don't use VBT ro make predictions about known facts.  You look up or determine the facts, you use VBT and other theories to rationalize those facts, to explain them.  Because blind predictions with something as simple as Lewis or VBT are pedagogically misleading, scientifically unsound, and very often wrong, leading to nonsense conclusions like "PF3 is sp3 hybridized at P, with an angle a little less than 109º", when the angle is really 98.