Electronic structure and geometries of heteroatomic clusters

Rational design of carbonitrile-carboxaldehyde cation receptor models: probing the nature of the heteroatom-metal interaction

In this work, hybrid functional and G4 methods were employed in the rational design of carbonitrile-carboxaldehyde receptor models for cation recognition. Electron-sharing and ionic interactions between the models and the cations were analyzed utilizing the concepts of overlap population, atomic valence, electrostatic potential, and CHELPG charge in order to elucidate the nature of the heteroatom-metal interaction, the N versus O disparity, and the effect of pH. Receptor fragment models from ionomycin were employed to rationalize the selection of receptor models for discriminating group I cations and enhancing the selectivity for Mg(II) rather than Ca(II), and to examine the effects of keto-enol forms and negatively charged sites. The changes in geometries, overlap population, metal valence, and CHELPG charge upon solvation in heptane medium as compared to the gas phase were negligible. The optimized geometries reveal that the interaction between group II cations and the keto, enol, and enolate forms of 2-cyanoethanal causes 12 % bending of the C-C-N angle from linearity. Overlap populations show that the electron-sharing interaction favors group II cations but that the same mechanism allows Li(I) to compete. The total spin of Li(I) is 17 % greater than that of Ca(II), but the G4 binding energies of the two are separated by more than 50 kcal/mol, favoring group II cations, which may eliminate interference from Li(I). 1,2-Dicyanoethylene, which has only one form, shows similar characteristics. CHELPG analysis shows that Mg(II) transfers 25 and 18 % of its positive charge to 2-cyanoethanal enolate and 1,2-dicyanoethylene, respectively. Hydrogen atoms receive most of the positive charge in both receptors, but the N-termini exhibit strikingly different characteristics. Electrostatic potential contour profiles were found to be in good agreement with the atomic charge distributions. The application of uncharged 1,3-dicarbonyl and 2-cyanocarbonyl receptors and a judicious choice of polymeric membrane that suppresses the Hofmeister effect should lead to high selectivity for magnesium, whereas the utilization of multiple negatively charged ionophores should result in selectivity for calcium.

Ionization potentials of small lithium clusters (Lin) and hydrogenated lithium clusters (LinH)

We present accurate ionization potentials (IPs) for small lithium clusters and hydrogenated lithium clusters (n=1-4), computed using coupled-cluster singles and doubles theory augmented with a perturbative correction for connected triple excitations [CCSD(T)] with the correlation-consistent weighted core-valence quadruple-zeta basis set (cc-pwCVQZ). In some cases the full CCSDT method has been used. Comparison of computed binding energies with experiment for the pure cationic lithium clusters reveals excellent agreement, demonstrating that previous discrepancies between computed and experimentally derived atomization energies for the corresponding neutral clusters are due to the use of an inaccurate experimental IP for Li(4). The experimental IP for Li(4) falls 0.43 eV below our theoretical adiabatic value of 4.74 eV, which should be a lower bound to the measured IP. Our recommended zero-point corrected adiabatic IPs for Li, Li(2), Li(3), Li(4), LiH, Li(2)H, Li(3)H, and Li(4)H are 5.39, 5.14, 4.11, 4.74, 7.69, 3.98, 4.69, and 4.05 eV, respectively. Zero-point vibrationally corrected CCSD(T) atomization energies per atom for Li(2) (+), Li(3) (+), Li(4) (+), LiH(+), Li(2)H(+), Li(3)H(+), and Li(4)H(+) are 0.64, 0.96, 0.90, 0.056, 1.62, 1.40, and 1.40 eV, respectively.

Binding energies of small lithium clusters (Lin) and hydrogenated lithium clusters (LinH)

Large coupled cluster computations utilizing the Dunning weighted correlation-consistent polarized core-valence (cc-pwCVXZ) hierarchy of basis sets have been conducted, resulting in a panoply of internally consistent geometries and atomization energies for small Li(n) and Li(n)H (n=1-4) clusters. In contrast to previous ab initio results, we predict a monotonic increase in atomization energies per atom with increasing cluster size for lithium clusters, in accordance with the historical Knudsen-effusion measurements of Wu. For hydrogenated lithium clusters, our results support previous theoretical work concerning the relatively low atomization energy per atom for Li(2)H compared to LiH and Li(3)H. The CCSD(T)/cc-pwCVQZ atomization energies for LiH, Li(2)H, Li(3)H, and the most stable isomer of Li(4)H, including zero-point energy corrections, are 55.7, 79.6, 113.0, and 130.6 kcal/mol, respectively. The latter results are not consistent with the most recent experiments of Wu.