Calculations of the structures, stabilities, Raman spectra, and NMR spectra of CdCl_n(OH₂)_a^2-n, CdBr_n(OH₂)_a^2-n, and ZnCl_n(OH₂)_a^2-n species in aqueous solution

Ab initio self-consistent-field MO methods have been used to calculate minimum-energy geometries, Raman stretching frequencies, and Cd NMR shieldings for Cd complexes which may exist in chloride-containing aqueous solutions, e.g., CdCl⁺, CdCl₂, CdCl₃^-, CdCl₄^2-, CdCl₅^3-, CdCl(OH₂)₅⁺, CdCl₂(OH₂)₂, CdCl₂(OH₂)₄, CdCl₃(OH₂)^-, CdCl₃(OH₂)₂^-, and CdCl₃(OH₂)₃^-. The optimum geometries calculated using effective core potentials and double-ζ-quality bases show Cd-O bond distances which are in good agreement with experiment and Cd-Cl distances which are somewhat too long. Cd NMR shieldings have been calculated for all these species using coupled Hartree-Fock theory and including all electrons. An analogous set of geometry and Raman spectra calculations were performed for the Cd bromides but the NMR shieldings have been evaluated only for the unhydrated species. Comparisons between theory and experiment are consistent with the following: (i) for the CdCl₄^2- species, hydration is very weak or nonexistent, (ii) CdCl⁺ and CdCl₂ are strongly hydrated as pseudooctahedral complexes, and (iii) CdCl₃^- is weakly hydrated, probably as a pseudotetrahedral complex CdCl₃(OH₂)^-. Results for the bromides are similar to those for the chlorides, suggesting that CdBr₄^2- is unhydrated and CdBr₃^- only weakly hydrated. Results for some additional structures for the previously studied ZnCl_n(OH₂)_a^2-n series show that Zn and Cd exhibit essentially the same trends in NMR shielding as a function of the number of halide and water ligands.