Abstract
The proton's Gibbs solvation energy in various organic solvents was determined by experimental and theoretical means to serve as anchor points for the solvent-independent Unified Acidity Scale (UAS) and the Protoelectric Potential Map (PPM). Experimentally, potential differences between two half-cells connected by an "ideal" ionic liquid salt bridge (ILSB) were measured. Here, our established setup with the "ideal" Ionic Liquid [N(2225)][NTf(2)] in the ILSB ([N(2225)] = [N(C(2)H(5))(3)(C(5)H(11))]; Tf = SO(2)CF(3)) was refined to enable measurements of the Gibbs energies of proton transfer Δ(tr)G°(H(+), S(1)→ S(2)) between different solvents S(1) and S(2), such as water, methanol, ethanol, acetonitrile and methyl formate. These results were further evaluated by extensive theoretical studies using the double-hybrid functional DSD-BLYP for structure optimization and highly accurate coupled cluster DLPNO-CCSD(T) calculations that were extrapolated to the basis set limit (CBS) for the gas phase contributions. Solvation effects were added implicitly with the Conductor-like Polarizable continuum Model (CPCM) in a Cluster Continuum approach, including a scheme to use the CPCM with coupled cluster calculations in ORCA's newest version. A monomer and a cluster thermodynamic cycle were used to calculate the proton's solvation energy from the calculated solvent clusters. The cluster cycle performed well with consistent improvement at higher levels of theory, while the monomer cycle suffered from inadequate error compensation. Yet, the monomer cycle results can be significantly improved by the inclusion of experimental data, such as Gibbs energies of evaporation. However, it remains clear that even when using the most sophisticated of all currently available methods, the error bars of the calculations may at best reach about 10-15 kJ mol(-1) and, given low polarity solvents like methyl formate with ε(r) = 8.838, may even substantially exceed this error bar. Analysis shows that one of the largest problems arising for the calculations is the lack of reliable experimental structural data in solution to know how the dissolved particles governing the protochemical potential exactly prevail in solution. The delicate balance of subtle and relatively weak hydrogen bonds or dispersive interactions may lead to large structural differences between the structures known in the solid state (e.g., from single crystal structure determinations) and those actually being thermodynamically relevant in (dynamic) equilibrium in solution. We present some problems encountered during our work on this subject. At least for the more polar "classical" solvents, experimental ILSB- as well as computationally obtained results are in good agreement with the known values from the TATB assumption, which is currently one of the most widely used assumptions for determining single-ion transfer energies. However, with none of the described in part rather sophisticated calculations, it is possible to achieve "chemical accuracy" in solution, as one can with the used "gold standard" coupled cluster methods in the gas phase (i.e., ±4 kJ mol(-1)).