IS THE MILLS-NIXON EFFECT REAL

The long debated issue of the Mills-Nixon effect is reexamined. If benzene's hydrogens are bent as if to form small rings, pure strain is imposed on the aromatic system. The result of this strain is fixation of benzene's bonds according to the Mills-Nixon postulation. It was found (eq 1) that DELTA-R is proportional to sin2 alpha, where DELTA-R is the difference between the long and the short bonds in the alternated benzene and alpha is the deviation from the natural bond angle (120-degrees). However, the 3-21G-optimized geometries for 2 and 3 show that these systems are much more delocalized (i.e., DELTA-R is much smaller) than [GRAPHICS] is expected from the bent benzene model. The "effective" bond angles (i.e., the angles that should have caused the observed DELTA-R) of these molecules were calculated by using the DELTA-R values from the 3-21G-optimized geometries and the relationship found between the bending angle and DELTA-R (eq 1). Thus, 3-21G angles are 86.2 and 93.6-degrees for 2 and 3, respectively, whereas the "effective" angles (eq 1) are 92.0 and 108.1-degrees, respectively. This difference was attributed to the formation of "banana bonds" in these systems. Comparison of the 3-21G results and the effective bond angles with the literature X-ray structures and electron-density-deformation analyses for benzenes annealated to three- and four-membered rings shows excellent agreement. It is concluded that when strain is imposed on a benzene the system responds by localizing its bonds in the Mills-Nixon manner. However, when small rings are annealed, the Mills-Nixon effect diminishes due to the formation of "banana bonds". Thus, special conditions must be fulfilled in order to observe the effect. Some predictions as to the geometries and electron densities of yet unstudied systems are given.