14N Nuclear Quadrupole Coupling — Correlation of Experimental and Theoretical Values with Structure*

In our first two papers on 14N nuclear quadrupole coupling constants (NQCC, Xij) [U 2], we showed that double zeta (DZ) ab initio molecular orbital calculations using a Huzinaga/Dunning 9s5p/4s basis [3] gave good linear correlations between the experimental microwave (MW) NQCC and the cal­ culated electric field gradients (EFG, <?,,) for a range of heterocyclic and other organic molecules. Fundamental to these studies is a value for the nitrogen atom quadrupole coupling constant {Qu)\ this is not readily obtained by electronic or micro­ wave spectroscopy, except as the product qaQu, however, two relatively recent values have been ob­ tained from scattering experiments; electron scatter­ ing yields the value 17.4 ± 0.2 mbarn [4], which is in excellent agreement with both a least squares fit of the data above [2], which yields 17.2 mbarn, and with other theoretical estimates based on small numbers of compounds [5]. However, fast-ion bom­ bardment has yielded 19.3 ± 0.8 mbarn [6 ] (closer to [5 c] than other calculated values), and the error bounds from [5, 6] do not overlap, and clearly do not lead to an agreed experimental value. This dis­ crepancy could be resolved by a marked change in the Sternheimer correction applied in [6 ]; for the moment we adopt the value 17.2 mbarn and treat it as a scaling factor. Strictly the value will be both