HALOGEN NUCLEAR-QUADRUPOLE COUPLING-CONSTANTS IN NON-AXIALLY SYMMETRICAL MOLECULES - AB-INITIO CALCULATIONS, WHICH INCLUDE CORRELATION, COMPARED WITH EXPERIMENT

Ab initio determination of the electric field gradient (EFG) tensors a t halogen and other centres enabled determination of the nuclear quadr upole coupling constants (NQCC) for a diverse set of C-2 nu, C-3 nu an d other symmetry molecules of general formula MH2X2 and MHX3, where th e halogen atoms (X) are Cl, Br and I, and the heavy central atoms (M) are C and Si. The study presents results at a standardised level of ca lculation, triple zeta in the valence space plus polarisation function s (TZVP) for the equilibrium geometry stage; all-electron MP2 correlat ion is included in all these studies. For the bromo and iodo compounds , especially the latter, it is essential to allow core polarisation, b y decontraction of the p,d-functions. This is conveniently done by ini tial optimization of the structure with a partly contracted basis, fol lowed by reestablishment of the equilibrium structure with the decontr acted basis. The NQCCs, derived from the EFGs, using the 'best' values for the atomic quadruple moments Cl, Br and I, lead to good agreement with the inertial axis (IA) data obtained from microwave spectroscopy . When the data from the present study is plotted against the values d erived from the IA data, obtained by whatever approximations chosen by the MW authors, we obtain a linear regression for the data (85 points ) with the slope 1.0365 and intercept -0.1737, with standard errors of 0.0042 and 0.2042, respectively; these are statistically identical re sults irrespective of whether the data is restricted to IA or EFG prin cipal axis (PA) data. Since as in the C-3 nu MH3X compounds studied pr eviously, a close correlation of the microwave spectral data with the calculations was observed using the 'best' current values for Q(Z), th ere seems no need to postulate that the values of Q(Br) for both Br-79 and Br-81 are seriously in error. A scaling downwards of Q(Z) by abou t 5% for Br and I increases the agreement with experiment, but the con tributions of relativistic effects are unknown, and could lead to furt her reassessment. Of the two common assumptions used in MW spectroscop y, to convert from IA to EFG-PA data, either (a) cylindrical symmetry of the NQCC along the bond direction, or (b) coincidence of the tensor principal element with the bond axis, the latter is found to be a muc h more realistic approximation.