We exploit gas-phase cluster ion techniques to provide insight into the local interactions underlying divalent metal ion-driven changes in the spectra of carboxylic acids at the air–water interface. This information clarifies the experimental findings that the CO stretching bands of long-chain acids appear at very similar energies when the head group is deprotonated by high subphase pH or exposed to relatively high concentrations of Ca2+ metal ions. To this end, we report the evolution of the vibrational spectra of size-selected [Ca2+·RCO2−]+·(H2O)n=0to12 and RCO2−·(H2O)n=0to14 cluster ions toward the features observed at the air–water interface. Surprisingly, not only does stepwise hydration of the RCO2− anion and the [Ca2+·RCO2−]+ contact ion pair yield solvatochromic responses in opposite directions, but in both cases, the responses of the 2 (symmetric and asymmetric stretching) CO bands to hydration are opposite to each other. The result is that both CO bands evolve toward their interfacial asymptotes from opposite directions. Simulations of the [Ca2+·RCO2−]+·(H2O)n clusters indicate that the metal ion remains directly bound to the head group in a contact ion pair motif as the asymmetric CO stretch converges at the interfacial value by n = 12. This establishes that direct metal complexation or deprotonation can account for the interfacial behavior. We discuss these effects in the context of a model that invokes the water network-dependent local electric field along the C–C bond that connects the head group to the hydrocarbon tail as the key microscopic parameter that is correlated with the observed trends.

The self-consistent charge density functional tight binding (SCC-DFTB) method has been applied to hydroxide water clusters and a hydroxide ion in bulk water. To determine the impact of various implementations of SCC-DFTB on the energetics and dynamics of a hydroxide ion in gas phase and condensed phase, the DFTB2, DFTB2-γ(h), DFTB2-γ(h)+gaus, DFTB3-diag, DFTB3-diag+gaus, DFTB3-Full+gaus, and DFTB3-3OB implementations have been tested. Energetic stabilities for small hydroxide clusters, OH(-)(H2O)n, where n = 4-7, are inconsistent with the results calculated with the B3LYP and second order Møller-Plesset (MP2) levels of ab initio theory. The condensed phase simulations, OH(-)(H2O)127, using the DFTB2, DFTB2-γ(h), DFTB2-γ(h)+gaus, DFTB3-diag, DFTB3-diag+gaus, DFTB3-Full+gaus and DFTB3-3OB methods are compared to Car-Parrinello molecular dynamics (CPMD) simulations using the BLYP functional. The SCC-DFTB method including a modified O-H repulsive potential and the third order correction (DFTB3-diag/Full+gaus) is shown to poorly reproduce the CPMD computational results, while the DFTB2 and DFTB2-γ(h) method somewhat more closely describe the structural and dynamical nature of the hydroxide ion in condensed phase. The DFTB3-3OB outperforms the MIO parameter set but is no more accurate than DFTB2. It is also shown that the overcoordinated water molecules lead to an incorrect bulk water density and result in unphysical water void formation. The results presented in this paper point to serious drawbacks for various DFTB extensions and corrections for a hydroxide ion in aqueous environments.

The low-lying potential energy minima of the H(+)(H(2)O)(n), n = 6, 21, and 22, protonated water clusters have been investigated using two versions of the self-consistent-charge density-functional tight-binding plus dispersion (SCC-DFTB+D) electronic structure methods. The relative energies of different isomers calculated using the SCC-DFTB+D methods are compared with the results of DFT and MP2 calculations. This comparison reveals that for H(+)(H(2)O)(6) the SCC-DFTB+D method with H-bonding and third-order corrections more closely reproduces the results of the MP2 calculations, whereas for the n = 21 and 22 clusters, the uncorrected SCC-DFTB+D method performs better. Both versions of the SCC-DFTB+D method are found to be biased toward Zundel structures.

The pKa values of HF, HCOOH, CH3COOH, CH3CH2COOH, H2CO3, HOCl, NH4(+), CH3NH3(+), H2O2, and CH3CH2OH in aqueous solution were predicted by QM/MM-MD in combination with umbrella samplings adopting the flexible asymmetric coordinate (FAC). This unique combination yielded remarkably accurate values with the maximum and root-mean-square errors of 0.45 and 0.22 in pKa units, respectively, without any numerical or experimental adjustments. The stability of the initially formed Coulomb pair rather than the proton transfer stage turned out to be the rate-determining step, implying that the stabilizations of the created ions require a large free energy increase. A remarkable correlation between DWR (degree of water rearrangements) and pKa was observed. As such, the large pKa of ethanol can be, in part, attributed to the large water rearrangement, strongly suggesting that proper samplings of water dynamics at dissociated regions are critical for accurate predictions of pKa. Current results exhibit a promising protocol for direct and accurate predictions of pKa. The significant variations in the gas phase deprotonation energies with level of theory appear to be mostly canceled by the similar changes in the averaged solute-solvent interactions, yielding accurate results.

With the help of QM/EFP-MD with modern correlated quantum theories, distinctly different proton transport dynamics for hydronium and hydroxide ions was revealed. The efficiency of proton transfer for hydronium was found to be significantly higher than that for hydroxide, and the difference in efficiency increased as the temperature was lowered. This difference in dynamics suggests that molecular Brownian diffusion may play an important role in hydroxide transport. Our theoretical findings are consistent with recent experimental observations of proton transfer in amorphous solid water.