Sterically crowded molecules

Sterically crowded molecules, like neopentane, are too unstable. 

AMI. While MNDO was widely accepted and extensively used, there were still some deficiencies in the model. In particular, excessive repulsions were observed in MNDO potential energy surfaces just outside chemical bonding distances. This deficiency manifested itself (5,7) in the inability of MNDO to model hydrogen bonding, as well as in large positive errors in the AHf of sterically crowded molecules and in heats of activation. Again Dewar set off to correct this deficiency. 

Limitations of MNDO. From its inception, some important limitations of MNDO were apparent. Sterically crowded molecules were calculated too unstable for example, the AHf of neopentane is predicted by MNDO to be —24.6 kcal/mol, compared with the observed -40.3 kcal/mol. On the other hand, four—membered rings were predicted to be too stable, this reaching a limit in cubane, which was predicted to be 49.6 kcal/mol too stable. Later on, other limitations were discovered, the most important from a biochemical standpoint being the virtually complete lack of a hydrogen bond. Other deficiencies included the extreme instability of hypervalent molecules. This effectivdy precluded the application of MNDO to organophosphorus compounds of biologic interest. Finally, activation barriers were predicted to be too high. 

Characteristic of f-olefin and cyclopropene complex formation are the upfield shifts observed in the NMR resonances of the olefinic protons (<5 =3.30-5.29 ppm) and carbons (<5 = 59.9-72.4 ppm). For the phosphite complexes, these upfield shift resonances decrease as the steric bulk of the imido ligand increases, corresponding to weaker binding of the cyclopropene in the more sterically crowded molecule. Difference NOE spectroscopy... 

Several directly measured values of AH° for homolytic dissociation of a metal-metal-bonded carbonyl in solution have been obtained (9). This was for the complexes [(n3-C3H5)Fe(CO)2 )2 where L = CO or a number of different P-donor ligands. The low value AH = 56.5 kJ mol-1 when L = CO was not unexpected for such a sterically crowded molecule. The P-donor substituents increased the steric crowding and displaced the equilibria in favor of the monomers but the effect seemed to be controlled more by AS° than AH°. In general metal-metal bond energies, however they may have been estimated, are too large to allow for direct measurement of equilibrium constants in solution in this way. 

Table II shows some literature examples of the breakdown in additivity which can occur in flexible molecules (Compounds 1-4) or in sterically crowded molecules such as the diphenylmethanes (Compounds 5 and 6). In the first four cases, a dipole is probably interacting with polarizable 7r-electrons of the aromatic ring. In Compound 5, the overall shape probably prevents water molecules from forming a solvate iceberg between the two rings—a situation which can be considered an intramolecular hydrophobic bond. In Compound 6 a combination of intramolecular hydrophobic effects and possibly interaction of the side chain dipole with one or both aromatic rings leads to a wide discrepancy between experimentally determined and calculated log F values.

Table 2 presents an abstract of Table 1 in ref. 12. In general, all three methods MNDO, AMI, and PM3 give rather remarkable results. In many cases, however, PM3 is more accurate than AMI, and both are more accurate than MNDO. MNDO predicts sterically crowded molecules to be too unstable and favors, in contrast, small rings, a shortcoming in most ZDO methods largely corrected in AMI and PM3. Of special interest is the observation that PM3 successfully reproduces the heats of formation of hypervalent compounds without the use of d orbitals. 

MNDO predicts sterically crowded molecules to be too unstable this effect is most marked in neopentane. In contrast, four-membered rings are predicted to be too stable the predicted AHf for cubane, for example, has an error of -49.6 kcal/mol. These trends are largely corrected in both AMI and PM3. 

---