Determination of experimental charge density in model nickel macrocycle: [3,11-bis(methoxycarbonyl)-1,5,9,13-tetraazacyclohexadeca-1,3,9,11-tetraenato-(2-)-kappa(4)N]nickel(II).

Experimental and theoretical atomic charges and d-orbital populations were obtained for [3,11-bis(methoxycarbonyl)-1,5,9,13-tetraazacyclohexadeca-1,3,9,11-tetraenato(2-)-kappa(4)N]nickel(II) monocrystal (16Ni) using the Hansen and Coppens formalism. Several models of this structure were tested as a function of quality of the fit, convergence of the refinement, value of residual peaks and holes, and Hirshfeld's rigid-bond test. The models with Ni-atomic scattering factors applied for the metal center are significantly better than those with the Ni-ionic scattering factors. The properties of the final electron density distributions are very consistent and similar for all of the models tested. The values of the d-orbital populations for Ni roughly agree with the occupation order of the d-electron levels for square-planar complexes derived from crystal field theory and are comparable with those obtained by the natural population analysis method. Experimental atoms-in-molecules charges calculated for four different models agree quite well for models with different symmetry restrictions and the same scattering factor for the Ni. Both experimental and theoretical methods predict relatively high negative values of charges for the nitrogen and oxygen atoms and also a quite high positive charge for the C(7) atom. The values of rho(r(BCP)) for the Ni-N bonds are in the range from 0.60 up to 0.75e A(-3), with positive laplacian values indicating noncovalent bonding. For the laplacian of the electron density evaluated in the plane of the macrocyclic ring, a typical map for a square-planar complex was obtained with four charge concentrations--3d(xy) orbitals--pointing toward the regions between the M-N bonds and charge depletions directed toward the (3,-1) critical points of the negative laplacian for the nitrogen atoms.