Computational Investigation of Condensed Phase Properties of Ionic Systems

Abstract

Four computational studies have been performed with various computational techniques including density functional theory, post-HF theory, and implicit solvation modeling to investigate the dehydrogenation mechanism of LiNH2BH3(s), acid dissociation constants (pKa1, pKa2, and pKa3) of polyprotic acids, the dissolution free energy of alkali-metal dianion salts (M2X1), and redox potentials of polyboranes. Theoretical background and computational methodology in these studies are introduced in chapter 1. Chapter 2-5 present four different studies exploring condensed-phase properties of inorganic materials including reaction mechanisms. Chapter 2 reports on the formation of LiNH2BH3 from (LiH)4 and NH3BH3 and their subsequent dehydrogenation. The free energy of activation for loss of H2 is reduced from 37.2 kcal/mol in NH3BH3 to 11.0 kcal/mol in (LiH)4 + NH3BH3. Further, H2 elimination from the (LiNH2BH3)2 dimer is predicted to be much easier than from the monomer which may suggest that a cooperative H2-loss mechanism is possible in solid LiNH2BH3. Chapter 3 reports a systematic study of ΔGaq/pKa for monoprotic, diprotic, and triprotic acids based on DFT/aug-cc-pVTZ combined with CPCM and SMD solvation modeling. All DFT/cavity set combinations considered showed similar accuracy for ΔGaq1/pKa1 (70% within ±2.5 kcal/mol of experiment) while only the M05-2X/Pauling cavity combination gave reasonable results for ΔGaq2/pKa2 when both pKa values are separated by more than three units (70% within ±5.0 kcal/mol of experiment). Chapter 4 reports on the dissolution Gibbs free energies (ΔGodiss) of salts (M2X1) using the Conductor-like Polarizable Continuum Model (CPCM) solvation modeling. The absolute solvation free energies of the alkali metal cations (ΔGsolv(M+)) come from the literature, which coincide well with half reduction potential versus SHE data. Lattice free energies (ΔGlatt) of salts were determined by three different approaches: (1) volumetric, (2) a cohesive Gibbs free energy (ΔGcoh) plus gaseous dissociation free energy (ΔGgas), and (3) the Born-Haber cycle. Only the M05-2X/Pauling combination with the three different methods for estimating ΔGlatt yields the expected negative dissolution free energies (ΔGodiss) of M2SO4. Chapter 5 reports the reduction potentials (EoRed versus SHE) of hypercloso boron hydrides BnHn (n=6-13) and B12X12 (X=F, Cl, OH, and CH3) with CPCM and SMD solvation modeling. The EoRed of BnHn-/2- (n=6-12) with the G4/M06-2X/Pauling (energy/solvation/cavity) combination agrees within 0.2 V of experimental values. The experimental oxidative stability (E1/2) of BnXn2- (X=F, Cl, OH, and CH3) is usually located between the values predicted using the B3LYP and M06-2X functionals. The disproportionation free energies (ΔGdpro) of 2BnHn- → BnHn + BnHn2- reveal that the stabilities of BnHn- (n=6-13) to disproportionation decrease in the order B8H8- > B9H9- > B11H11- > B10H10-. The spin densities in B12X12- (X=F, Cl, OH, and CH3) tend to delocalize on the boron atoms rather than on the exterior functional groups. The partitioning of ΔGsolv(BnHn2-) over spheres allows a rationalization of the nonlinear correlation between ΔGE.A. and EoRed for B6H6 /2-, B11H11-/2-, and B13H13-/2-.